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 an issue where some non-default TLS server configurations can cause
177 unbounded memory growth when processing TLSv1.3 sessions. An attacker may
178 exploit certain server configurations to trigger unbounded memory growth that
179 would lead to a Denial of Service
181 This problem can occur in TLSv1.3 if the non-default SSL_OP_NO_TICKET option
182 is being used (but not if early_data is also configured and the default
183 anti-replay protection is in use). In this case, under certain conditions,
184 the session cache can get into an incorrect state and it will fail to flush
185 properly as it fills. The session cache will continue to grow in an unbounded
186 manner. A malicious client could deliberately create the scenario for this
187 failure to force a Denial of Service. It may also happen by accident in
194 * Fixed bug where SSL_export_keying_material() could not be used with QUIC
195 connections. (#23560)
199 ### Changes between 3.2.0 and 3.2.1 [30 Jan 2024]
201 * A file in PKCS12 format can contain certificates and keys and may come from
202 an untrusted source. The PKCS12 specification allows certain fields to be
203 NULL, but OpenSSL did not correctly check for this case. A fix has been
204 applied to prevent a NULL pointer dereference that results in OpenSSL
205 crashing. If an application processes PKCS12 files from an untrusted source
206 using the OpenSSL APIs then that application will be vulnerable to this
207 issue prior to this fix.
209 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
210 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
211 and PKCS12_newpass().
213 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
214 function is related to writing data we do not consider it security
221 * When function EVP_PKEY_public_check() is called on RSA public keys,
222 a computation is done to confirm that the RSA modulus, n, is composite.
223 For valid RSA keys, n is a product of two or more large primes and this
224 computation completes quickly. However, if n is an overly large prime,
225 then this computation would take a long time.
227 An application that calls EVP_PKEY_public_check() and supplies an RSA key
228 obtained from an untrusted source could be vulnerable to a Denial of Service
231 The function EVP_PKEY_public_check() is not called from other OpenSSL
232 functions however it is called from the OpenSSL pkey command line
233 application. For that reason that application is also vulnerable if used
234 with the "-pubin" and "-check" options on untrusted data.
236 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
237 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
243 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
244 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
249 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
250 for PowerPC CPUs saves the contents of vector registers in different
251 order than they are restored. Thus the contents of some of these vector
252 registers is corrupted when returning to the caller. The vulnerable code is
253 used only on newer PowerPC processors supporting the PowerISA 2.07
256 The consequences of this kind of internal application state corruption can
257 be various - from no consequences, if the calling application does not
258 depend on the contents of non-volatile XMM registers at all, to the worst
259 consequences, where the attacker could get complete control of the
260 application process. However unless the compiler uses the vector registers
261 for storing pointers, the most likely consequence, if any, would be an
262 incorrect result of some application dependent calculations or a crash
263 leading to a denial of service.
269 * Disable building QUIC server utility when OpenSSL is configured with
274 ### Changes between 3.1 and 3.2.0 [23 Nov 2023]
276 * Fix excessive time spent in DH check / generation with large Q parameter
279 Applications that use the functions DH_generate_key() to generate an
280 X9.42 DH key may experience long delays. Likewise, applications that use
281 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
282 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
283 Where the key or parameters that are being checked have been obtained from
284 an untrusted source this may lead to a Denial of Service.
290 * The BLAKE2b hash algorithm supports a configurable output length
291 by setting the "size" parameter.
293 *Čestmír Kalina and Tomáš Mráz*
295 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
299 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
300 and the corresponding provider-storemgmt API function
301 OSSL_FUNC_store_delete().
305 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
306 a passphrase callback when opening a store.
310 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
311 from 8 bytes to 16 bytes.
312 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
313 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
314 requires a salt length of 128 bits. This affects OpenSSL command line
315 applications such as "genrsa" and "pkcs8" and API's such as
316 PEM_write_bio_PrivateKey() that are reliant on the default value.
317 The additional commandline option 'saltlen' has been added to the
318 OpenSSL command line applications for "pkcs8" and "enc" to allow the
319 salt length to be set to a non default value.
323 * Changed the default value of the `ess_cert_id_alg` configuration
324 option which is used to calculate the TSA's public key certificate
325 identifier. The default algorithm is updated to be sha256 instead
328 *Małgorzata Olszówka*
330 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
331 table for point multiplication of the base point, which increases the size of
332 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
333 been added to disable the precomputed table.
337 * Added client side support for QUIC
339 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
341 * Added multiple tutorials on the OpenSSL library and in particular
342 on writing various clients (using TLS and QUIC protocols) with libssl.
346 * Added secp384r1 implementation using Solinas' reduction to improve
347 speed of the NIST P-384 elliptic curve. To enable the implementation
348 the build option `enable-ec_nistp_64_gcc_128` must be used.
352 * Improved RFC7468 compliance of the asn1parse command.
354 *Matthias St. Pierre*
356 * Added SHA256/192 algorithm support.
360 * Added support for securely getting root CA certificate update in
365 * Improved contention on global write locks by using more read locks where
370 * Improved performance of OSSL_PARAM lookups in performance critical
375 * Added the SSL_get0_group_name() function to provide access to the
376 name of the group used for the TLS key exchange.
380 * Provide a new configure option `no-http` that can be used to disable the
381 HTTP support. Provide new configure options `no-apps` and `no-docs` to
382 disable building the openssl command line application and the documentation.
386 * Provide a new configure option `no-ecx` that can be used to disable the
387 X25519, X448, and EdDSA support.
391 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
392 the EVP_KDF_CTX_set_params() function they are now concatenated not just
393 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
397 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
398 the provider context as a parameter.
402 * TLS round-trip time calculation was added by a Brigham Young University
403 Capstone team partnering with Sandia National Laboratories. A new function
404 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
409 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
410 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
411 option. Use of the "advanced" s_client command command via the "-adv" option
416 * Added an "advanced" command mode to s_client. Use this with the "-adv"
417 option. The old "basic" command mode recognises certain letters that must
418 always appear at the start of a line and cannot be escaped. The advanced
419 command mode enables commands to be entered anywhere and there is an
420 escaping mechanism. After starting s_client with "-adv" type "{help}"
421 to show a list of available commands.
425 * Add Raw Public Key (RFC7250) support. Authentication is supported
426 by matching keys against either local policy (TLSA records synthesised
427 from the expected keys) or DANE (TLSA records obtained by the
428 application from DNS). TLSA records will also match the same key in
429 the server certificate, should RPK use not happen to be negotiated.
433 * Added support for modular exponentiation and CRT offloading for the
438 * Added further assembler code for the RISC-V architecture.
442 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
443 from a given EC_GROUP.
447 * Improved support for non-default library contexts and property queries
448 when parsing PKCS#12 files.
452 * Implemented support for all five instances of EdDSA from RFC8032:
453 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
454 The streaming is not yet supported for the HashEdDSA variants
455 (Ed25519ph and Ed448ph).
459 * Added SM4 optimization for ARM processors using ASIMD and AES HW
464 * Implemented SM4-XTS support.
468 * Added platform-agnostic OSSL_sleep() function.
472 * Implemented deterministic ECDSA signatures (RFC6979) support.
476 * Implemented AES-GCM-SIV (RFC8452) support.
480 * Added support for pluggable (provider-based) TLS signature algorithms.
481 This enables TLS 1.3 authentication operations with algorithms embedded
482 in providers not included by default in OpenSSL. In combination with
483 the already available pluggable KEM and X.509 support, this enables
484 for example suitable providers to deliver post-quantum or quantum-safe
485 cryptography to OpenSSL users.
489 * Added support for pluggable (provider-based) CMS signature algorithms.
490 This enables CMS sign and verify operations with algorithms embedded
491 in providers not included by default in OpenSSL.
495 * Added support for Hybrid Public Key Encryption (HPKE) as defined
496 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
497 Message Layer Security (MLS) and other IETF specifications.
498 HPKE can also be used by other applications that require
499 encrypting "to" an ECDH public key. External APIs are defined in
500 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
504 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
509 * Add support for certificate compression (RFC8879), including
510 library support for Brotli and Zstandard compression.
514 * Add the ability to add custom attributes to PKCS12 files. Add a new API
515 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
516 for a user specified callback and optional argument.
517 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
518 added to the existing STACK_OF attrs.
522 * Major refactor of the libssl record layer.
526 * Add a mac salt length option for the pkcs12 command.
530 * Add more SRTP protection profiles from RFC8723 and RFC8269.
534 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
536 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
538 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
539 supported and enabled.
543 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
544 to the list of ciphersuites providing Perfect Forward Secrecy as
545 required by SECLEVEL >= 3.
547 *Dmitry Belyavskiy, Nicola Tuveri*
549 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
550 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
551 SSL_get0_iana_groups() function-like macro, retrieves the list of
552 supported groups sent by the peer.
553 The function SSL_client_hello_get_extension_order() populates
554 a caller-supplied array with the list of extension types present in the
555 ClientHello, in order of appearance.
559 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
560 to make it possible to use empty passphrase strings.
564 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
568 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
569 to change functions used for allocating the memory of asynchronous call stack.
573 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
577 * A failure exit code is returned when using the openssl x509 command to check
578 certificate attributes and the checks fail.
582 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
583 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
584 of 160 bits and above and less than 224 bits were previously accepted by
585 default but are now no longer allowed. By default TLS compression was
586 already disabled in previous OpenSSL versions. At security level 2 it cannot
591 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
596 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
597 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
598 will need to load the legacy crypto provider.
602 * CCM8 cipher suites in TLS have been downgraded to security level zero
603 because they use a short authentication tag which lowers their strength.
607 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
608 by default. Also spaces surrounding `=` in DN output are removed.
612 * Add X.509 certificate codeSigning purpose and related checks on key usage and
613 extended key usage of the leaf certificate according to the CA/Browser Forum.
617 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
618 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
619 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
620 X.509 version 3 if the certificate information includes X.509 extensions.
624 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
625 such as adding a trace facility for debugging certificate chain building.
629 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
630 in particular supporting requests for central key generation, generalized
631 polling, and various types of genm/genp exchanges defined in CMP Updates.
635 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
636 like correcting the TLS and proxy support and adding tracing for debugging.
640 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
644 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
645 a certificate to be added is already present. `CMS_sign_ex()` and
646 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
647 and no longer throw an error for them.
651 * Fixed and extended `util/check-format.pl` for checking adherence to the
652 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
653 The checks are meanwhile more complete and yield fewer false positives.
657 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
658 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
659 calls. They can be used as the transport BIOs for QUIC.
661 *Hugo Landau, Matt Caswell and Tomáš Mráz*
663 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
664 sending and receiving multiple messages in a single call. An implementation
665 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
669 * Support for loading root certificates from the Windows certificate store
670 has been added. The support is in the form of a store which recognises the
671 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
672 arguments. This store is built by default and can be disabled using the new
673 compile-time option `no-winstore`. This store is not currently used by
674 default and must be loaded explicitly using the above store URI. It is
675 expected to be loaded by default in the future.
679 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
680 kernel versions that support KTLS have a known bug in CCM processing. That
681 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
682 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
687 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
692 * Zerocopy KTLS sendfile() support on Linux.
696 * The OBJ_ calls are now thread safe using a global lock.
700 * New parameter `-digest` for openssl cms command allowing signing
701 pre-computed digests and new CMS API functions supporting that
706 * OPENSSL_malloc() and other allocation functions now raise errors on
707 allocation failures. The callers do not need to explicitly raise errors
708 unless they want to for tracing purposes.
712 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
713 decryption as a protection against Bleichenbacher-like attacks.
714 The RSA decryption API will now return a randomly generated deterministic
715 message instead of an error in case it detects an error when checking
716 padding during PKCS#1 v1.5 decryption. This is a general protection against
717 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
719 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
720 on the RSA decryption context.
724 * Added support for Brainpool curves in TLS-1.3.
726 *Bernd Edlinger and Matt Caswell*
728 * Added OpenBSD specific build targets.
732 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
733 a basic thread pool implementation for select platforms.
740 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
742 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
743 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
744 that alter the key or IV length ([CVE-2023-5363]).
748 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
750 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
752 The POLY1305 MAC (message authentication code) implementation in OpenSSL
753 does not save the contents of non-volatile XMM registers on Windows 64
754 platform when calculating the MAC of data larger than 64 bytes. Before
755 returning to the caller all the XMM registers are set to zero rather than
756 restoring their previous content. The vulnerable code is used only on newer
757 x86_64 processors supporting the AVX512-IFMA instructions.
759 The consequences of this kind of internal application state corruption can
760 be various - from no consequences, if the calling application does not
761 depend on the contents of non-volatile XMM registers at all, to the worst
762 consequences, where the attacker could get complete control of the
763 application process. However given the contents of the registers are just
764 zeroized so the attacker cannot put arbitrary values inside, the most likely
765 consequence, if any, would be an incorrect result of some application
766 dependent calculations or a crash leading to a denial of service.
772 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
774 * Fix excessive time spent checking DH q parameter value.
776 The function DH_check() performs various checks on DH parameters. After
777 fixing CVE-2023-3446 it was discovered that a large q parameter value can
778 also trigger an overly long computation during some of these checks.
779 A correct q value, if present, cannot be larger than the modulus p
780 parameter, thus it is unnecessary to perform these checks if q is larger
783 If DH_check() is called with such q parameter value,
784 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
785 intensive checks are skipped.
791 * Fix DH_check() excessive time with over sized modulus.
793 The function DH_check() performs various checks on DH parameters. One of
794 those checks confirms that the modulus ("p" parameter) is not too large.
795 Trying to use a very large modulus is slow and OpenSSL will not normally use
796 a modulus which is over 10,000 bits in length.
798 However the DH_check() function checks numerous aspects of the key or
799 parameters that have been supplied. Some of those checks use the supplied
800 modulus value even if it has already been found to be too large.
802 A new limit has been added to DH_check of 32,768 bits. Supplying a
803 key/parameters with a modulus over this size will simply cause DH_check() to
810 * Do not ignore empty associated data entries with AES-SIV.
812 The AES-SIV algorithm allows for authentication of multiple associated
813 data entries along with the encryption. To authenticate empty data the
814 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
815 with NULL pointer as the output buffer and 0 as the input buffer length.
816 The AES-SIV implementation in OpenSSL just returns success for such call
817 instead of performing the associated data authentication operation.
818 The empty data thus will not be authenticated. ([CVE-2023-2975])
820 Thanks to Juerg Wullschleger (Google) for discovering the issue.
822 The fix changes the authentication tag value and the ciphertext for
823 applications that use empty associated data entries with AES-SIV.
824 To decrypt data encrypted with previous versions of OpenSSL the application
825 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
830 * When building with the `enable-fips` option and using the resulting
831 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
832 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
833 not operate with truncated digests (FIPS 140-3 IG G.R).
837 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
839 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
840 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
842 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
843 numeric text form. For gigantic sub-identifiers, this would take a very
844 long time, the time complexity being O(n^2) where n is the size of that
845 sub-identifier. ([CVE-2023-2650])
847 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
848 IDENTIFIER to canonical numeric text form if the size of that OBJECT
849 IDENTIFIER is 586 bytes or less, and fail otherwise.
851 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
852 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
853 most 128 sub-identifiers, and that the maximum value that each sub-
854 identifier may have is 2^32-1 (4294967295 decimal).
856 For each byte of every sub-identifier, only the 7 lower bits are part of
857 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
858 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
863 * Multiple algorithm implementation fixes for ARM BE platforms.
867 * Added a -pedantic option to fipsinstall that adjusts the various
868 settings to ensure strict FIPS compliance rather than backwards
873 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
874 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
875 trigger a crash of an application using AES-XTS decryption if the memory
876 just after the buffer being decrypted is not mapped.
877 Thanks to Anton Romanov (Amazon) for discovering the issue.
882 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
883 The previous fix for this timing side channel turned out to cause
884 a severe 2-3x performance regression in the typical use case
885 compared to 3.0.7. The new fix uses existing constant time
886 code paths, and restores the previous performance level while
887 fully eliminating all existing timing side channels.
888 The fix was developed by Bernd Edlinger with testing support
893 * Add FIPS provider configuration option to disallow the use of
894 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
895 The option '-no_drbg_truncated_digests' can optionally be
896 supplied to 'openssl fipsinstall'.
900 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
901 that it does not enable policy checking. Thanks to David Benjamin for
902 discovering this issue.
907 * Fixed an issue where invalid certificate policies in leaf certificates are
908 silently ignored by OpenSSL and other certificate policy checks are skipped
909 for that certificate. A malicious CA could use this to deliberately assert
910 invalid certificate policies in order to circumvent policy checking on the
911 certificate altogether.
916 * Limited the number of nodes created in a policy tree to mitigate
917 against CVE-2023-0464. The default limit is set to 1000 nodes, which
918 should be sufficient for most installations. If required, the limit
919 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
920 time define to a desired maximum number of nodes or zero to allow
926 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
928 * Add FIPS provider configuration option to enforce the
929 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
930 The option '-ems-check' can optionally be supplied to
931 'openssl fipsinstall'.
935 * The FIPS provider includes a few non-approved algorithms for
936 backward compatibility purposes and the "fips=yes" property query
937 must be used for all algorithm fetches to ensure FIPS compliance.
939 The algorithms that are included but not approved are Triple DES ECB,
940 Triple DES CBC and EdDSA.
944 * Added support for KMAC in KBKDF.
948 * RNDR and RNDRRS support in provider functions to provide
949 random number generation for Arm CPUs (aarch64).
953 * s_client and s_server apps now explicitly say when the TLS version
954 does not include the renegotiation mechanism. This avoids confusion
955 between that scenario versus when the TLS version includes secure
956 renegotiation but the peer lacks support for it.
960 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
962 *Tomasz Kantecki, Andrey Matyukov*
964 * The various OBJ_* functions have been made thread safe.
968 * Parallel dual-prime 1536/2048-bit modular exponentiation for
969 AVX512_IFMA capable processors.
971 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
973 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
974 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
975 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
976 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
977 `OPENSSL_NO_DEPRECATED_3_1`.
979 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
980 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
981 definitions for these functions regardless of whether
982 `OPENSSL_NO_DEPRECATED_3_1` is defined.
984 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
985 functions regardless of whether they are using them. It is recommended that
986 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
990 * When generating safe-prime DH parameters set the recommended private key
991 length equivalent to minimum key lengths as in RFC 7919.
995 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
996 maximum size that is smaller or equal to the digest length to comply with
997 FIPS 186-4 section 5. This is implemented by a new option
998 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
999 `rsa_pss_saltlen` parameter, which is now the default. Signature
1000 verification is not affected by this change and continues to work as before.
1007 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
1008 listed here are only a brief description.
1009 The migration guide contains more detailed information related to new features,
1010 breaking changes, and mappings for the large list of deprecated functions.
1012 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
1014 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
1016 * Fixed NULL dereference during PKCS7 data verification.
1018 A NULL pointer can be dereferenced when signatures are being
1019 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
1020 algorithm used for the signature is known to the OpenSSL library but
1021 the implementation of the hash algorithm is not available the digest
1022 initialization will fail. There is a missing check for the return
1023 value from the initialization function which later leads to invalid
1024 usage of the digest API most likely leading to a crash.
1027 PKCS7 data is processed by the SMIME library calls and also by the
1028 time stamp (TS) library calls. The TLS implementation in OpenSSL does
1029 not call these functions however third party applications would be
1030 affected if they call these functions to verify signatures on untrusted
1035 * Fixed X.400 address type confusion in X.509 GeneralName.
1037 There is a type confusion vulnerability relating to X.400 address processing
1038 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
1039 but the public structure definition for GENERAL_NAME incorrectly specified
1040 the type of the x400Address field as ASN1_TYPE. This field is subsequently
1041 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
1042 than an ASN1_STRING.
1044 When CRL checking is enabled (i.e. the application sets the
1045 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
1046 pass arbitrary pointers to a memcmp call, enabling them to read memory
1047 contents or enact a denial of service.
1052 * Fixed NULL dereference validating DSA public key.
1054 An invalid pointer dereference on read can be triggered when an
1055 application tries to check a malformed DSA public key by the
1056 EVP_PKEY_public_check() function. This will most likely lead
1057 to an application crash. This function can be called on public
1058 keys supplied from untrusted sources which could allow an attacker
1059 to cause a denial of service attack.
1061 The TLS implementation in OpenSSL does not call this function
1062 but applications might call the function if there are additional
1063 security requirements imposed by standards such as FIPS 140-3.
1066 *Shane Lontis, Tomáš Mráz*
1068 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
1070 An invalid pointer dereference on read can be triggered when an
1071 application tries to load malformed PKCS7 data with the
1072 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
1074 The result of the dereference is an application crash which could
1075 lead to a denial of service attack. The TLS implementation in OpenSSL
1076 does not call this function however third party applications might
1077 call these functions on untrusted data.
1082 * Fixed Use-after-free following BIO_new_NDEF.
1084 The public API function BIO_new_NDEF is a helper function used for
1085 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
1086 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
1087 be called directly by end user applications.
1089 The function receives a BIO from the caller, prepends a new BIO_f_asn1
1090 filter BIO onto the front of it to form a BIO chain, and then returns
1091 the new head of the BIO chain to the caller. Under certain conditions,
1092 for example if a CMS recipient public key is invalid, the new filter BIO
1093 is freed and the function returns a NULL result indicating a failure.
1094 However, in this case, the BIO chain is not properly cleaned up and the
1095 BIO passed by the caller still retains internal pointers to the previously
1096 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
1097 then a use-after-free will occur. This will most likely result in a crash.
1100 *Viktor Dukhovni, Matt Caswell*
1102 * Fixed Double free after calling PEM_read_bio_ex.
1104 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
1105 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
1106 data. If the function succeeds then the "name_out", "header" and "data"
1107 arguments are populated with pointers to buffers containing the relevant
1108 decoded data. The caller is responsible for freeing those buffers. It is
1109 possible to construct a PEM file that results in 0 bytes of payload data.
1110 In this case PEM_read_bio_ex() will return a failure code but will populate
1111 the header argument with a pointer to a buffer that has already been freed.
1112 If the caller also frees this buffer then a double free will occur. This
1113 will most likely lead to a crash.
1115 The functions PEM_read_bio() and PEM_read() are simple wrappers around
1116 PEM_read_bio_ex() and therefore these functions are also directly affected.
1118 These functions are also called indirectly by a number of other OpenSSL
1119 functions including PEM_X509_INFO_read_bio_ex() and
1120 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
1121 internal uses of these functions are not vulnerable because the caller does
1122 not free the header argument if PEM_read_bio_ex() returns a failure code.
1125 *Kurt Roeckx, Matt Caswell*
1127 * Fixed Timing Oracle in RSA Decryption.
1129 A timing based side channel exists in the OpenSSL RSA Decryption
1130 implementation which could be sufficient to recover a plaintext across
1131 a network in a Bleichenbacher style attack. To achieve a successful
1132 decryption an attacker would have to be able to send a very large number
1133 of trial messages for decryption. The vulnerability affects all RSA padding
1134 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1137 *Dmitry Belyavsky, Hubert Kario*
1139 * Fixed X.509 Name Constraints Read Buffer Overflow.
1141 A read buffer overrun can be triggered in X.509 certificate verification,
1142 specifically in name constraint checking. The read buffer overrun might
1143 result in a crash which could lead to a denial of service attack.
1144 In a TLS client, this can be triggered by connecting to a malicious
1145 server. In a TLS server, this can be triggered if the server requests
1146 client authentication and a malicious client connects.
1151 * Fixed X.509 Policy Constraints Double Locking security issue.
1153 If an X.509 certificate contains a malformed policy constraint and
1154 policy processing is enabled, then a write lock will be taken twice
1155 recursively. On some operating systems (most widely: Windows) this
1156 results in a denial of service when the affected process hangs. Policy
1157 processing being enabled on a publicly facing server is not considered
1158 to be a common setup.
1163 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1164 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1165 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1166 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1167 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1168 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1169 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1170 for legacy EC and SM2 keys is also changed similarly to honor the
1171 equivalent conversion format flag as specified in the underlying
1172 `EC_KEY` object being exported to a provider, when this function is
1173 called through `EVP_PKEY_export()`.
1177 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1179 * Fixed two buffer overflows in punycode decoding functions.
1181 A buffer overrun can be triggered in X.509 certificate verification,
1182 specifically in name constraint checking. Note that this occurs after
1183 certificate chain signature verification and requires either a CA to
1184 have signed the malicious certificate or for the application to continue
1185 certificate verification despite failure to construct a path to a trusted
1188 In a TLS client, this can be triggered by connecting to a malicious
1189 server. In a TLS server, this can be triggered if the server requests
1190 client authentication and a malicious client connects.
1192 An attacker can craft a malicious email address to overflow
1193 an arbitrary number of bytes containing the `.` character (decimal 46)
1194 on the stack. This buffer overflow could result in a crash (causing a
1198 An attacker can craft a malicious email address to overflow four
1199 attacker-controlled bytes on the stack. This buffer overflow could
1200 result in a crash (causing a denial of service) or potentially remote code
1201 execution depending on stack layout for any given platform/compiler.
1206 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1207 parameters in OpenSSL code.
1208 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1209 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1210 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1211 Using these invalid names may cause algorithms to use slower methods
1212 that ignore the CRT parameters.
1216 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1221 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1222 data to be signed before signing the certificate.
1226 * Added RIPEMD160 to the default provider.
1230 * Ensured that the key share group sent or accepted for the key exchange
1231 is allowed for the protocol version.
1235 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1237 * OpenSSL supports creating a custom cipher via the legacy
1238 EVP_CIPHER_meth_new() function and associated function calls. This function
1239 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1240 to use the new provider mechanism in order to implement custom ciphers.
1242 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1243 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1244 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1245 and decryption initialisation functions). Instead of using the custom cipher
1246 directly it incorrectly tries to fetch an equivalent cipher from the
1247 available providers. An equivalent cipher is found based on the NID passed to
1248 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1249 given cipher. However it is possible for an application to incorrectly pass
1250 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1251 is used in this way the OpenSSL encryption/decryption initialisation function
1252 will match the NULL cipher as being equivalent and will fetch this from the
1253 available providers. This will succeed if the default provider has been
1254 loaded (or if a third party provider has been loaded that offers this
1255 cipher). Using the NULL cipher means that the plaintext is emitted as the
1258 Applications are only affected by this issue if they call
1259 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1260 encryption/decryption initialisation function. Applications that only use
1261 SSL/TLS are not impacted by this issue.
1266 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1271 * Fixed the linux-mips64 Configure target which was missing the
1272 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1277 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1282 * Correctly handle a retransmitted ClientHello in DTLS
1286 * Fixed detection of ktls support in cross-compile environment on Linux
1290 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1295 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1296 report correct results in some cases
1300 * Fix UWP builds by defining VirtualLock
1304 * For known safe primes use the minimum key length according to RFC 7919.
1305 Longer private key sizes unnecessarily raise the cycles needed to compute the
1306 shared secret without any increase of the real security. This fixes a
1307 regression from 1.1.1 where these shorter keys were generated for the known
1312 * Added the loongarch64 target
1316 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1317 only passed to the FIPS provider and not to the default or legacy provider.
1321 * Fixed reported performance degradation on aarch64. Restored the
1322 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1323 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1324 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1325 The new algorithm is still used for 32 bit targets.
1329 * Added a missing header for memcmp that caused compilation failure on some
1334 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1336 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1337 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1338 This issue makes the RSA implementation with 2048 bit private keys
1339 incorrect on such machines and memory corruption will happen during
1340 the computation. As a consequence of the memory corruption an attacker
1341 may be able to trigger a remote code execution on the machine performing
1344 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1345 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1346 are affected by this issue.
1351 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1352 implementation would not encrypt the entirety of the data under some
1353 circumstances. This could reveal sixteen bytes of data that was
1354 preexisting in the memory that wasn't written. In the special case of
1355 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1357 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1358 they are both unaffected.
1361 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1363 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1365 * In addition to the c_rehash shell command injection identified in
1366 CVE-2022-1292, further bugs where the c_rehash script does not
1367 properly sanitise shell metacharacters to prevent command injection have been
1370 When the CVE-2022-1292 was fixed it was not discovered that there
1371 are other places in the script where the file names of certificates
1372 being hashed were possibly passed to a command executed through the shell.
1374 This script is distributed by some operating systems in a manner where
1375 it is automatically executed. On such operating systems, an attacker
1376 could execute arbitrary commands with the privileges of the script.
1378 Use of the c_rehash script is considered obsolete and should be replaced
1379 by the OpenSSL rehash command line tool.
1382 *Daniel Fiala, Tomáš Mráz*
1384 * Case insensitive string comparison no longer uses locales. It has instead
1385 been directly implemented.
1389 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1391 * Case insensitive string comparison is reimplemented via new locale-agnostic
1392 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1393 comparison. The previous implementation had problems when the Turkish locale
1398 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1399 metacharacters to prevent command injection. This script is distributed by
1400 some operating systems in a manner where it is automatically executed. On
1401 such operating systems, an attacker could execute arbitrary commands with the
1402 privileges of the script.
1404 Use of the c_rehash script is considered obsolete and should be replaced
1405 by the OpenSSL rehash command line tool.
1410 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1411 certificate on an OCSP response. The bug caused the function in the case
1412 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1413 response (meaning a successful verification) even in the case where the
1414 response signing certificate fails to verify.
1416 It is anticipated that most users of `OCSP_basic_verify` will not use the
1417 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1418 a negative value (indicating a fatal error) in the case of a certificate
1419 verification failure. The normal expected return value in this case would be
1422 This issue also impacts the command line OpenSSL "ocsp" application. When
1423 verifying an ocsp response with the "-no_cert_checks" option the command line
1424 application will report that the verification is successful even though it
1425 has in fact failed. In this case the incorrect successful response will also
1426 be accompanied by error messages showing the failure and contradicting the
1427 apparently successful result.
1432 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1433 AAD data as the MAC key. This made the MAC key trivially predictable.
1435 An attacker could exploit this issue by performing a man-in-the-middle attack
1436 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1437 that the modified data would still pass the MAC integrity check.
1439 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1440 endpoint will always be rejected by the recipient and the connection will
1441 fail at that point. Many application protocols require data to be sent from
1442 the client to the server first. Therefore, in such a case, only an OpenSSL
1443 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1445 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1446 sent in both directions. In this case both clients and servers could be
1447 affected, regardless of the application protocol.
1449 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1450 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1451 the handshake when using this ciphersuite.
1453 The confidentiality of data is not impacted by this issue, i.e. an attacker
1454 cannot decrypt data that has been encrypted using this ciphersuite - they can
1457 In order for this attack to work both endpoints must legitimately negotiate
1458 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1459 OpenSSL 3.0, and is not available within the default provider or the default
1460 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1461 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1462 following must have occurred:
1464 1) OpenSSL must have been compiled with the (non-default) compile time option
1465 enable-weak-ssl-ciphers
1467 2) OpenSSL must have had the legacy provider explicitly loaded (either
1468 through application code or via configuration)
1470 3) The ciphersuite must have been explicitly added to the ciphersuite list
1472 4) The libssl security level must have been set to 0 (default is 1)
1474 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1476 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1477 others that both endpoints have in common
1482 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1483 occupied by the removed hash table entries.
1485 This function is used when decoding certificates or keys. If a long lived
1486 process periodically decodes certificates or keys its memory usage will
1487 expand without bounds and the process might be terminated by the operating
1488 system causing a denial of service. Also traversing the empty hash table
1489 entries will take increasingly more time.
1491 Typically such long lived processes might be TLS clients or TLS servers
1492 configured to accept client certificate authentication.
1495 *Hugo Landau, Aliaksei Levin*
1497 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1498 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1499 statistics are no longer supported. For compatibility, these statistics are
1500 still listed in the output but are now always reported as zero.
1504 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1506 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1507 for non-prime moduli.
1509 Internally this function is used when parsing certificates that contain
1510 elliptic curve public keys in compressed form or explicit elliptic curve
1511 parameters with a base point encoded in compressed form.
1513 It is possible to trigger the infinite loop by crafting a certificate that
1514 has invalid explicit curve parameters.
1516 Since certificate parsing happens prior to verification of the certificate
1517 signature, any process that parses an externally supplied certificate may thus
1518 be subject to a denial of service attack. The infinite loop can also be
1519 reached when parsing crafted private keys as they can contain explicit
1520 elliptic curve parameters.
1522 Thus vulnerable situations include:
1524 - TLS clients consuming server certificates
1525 - TLS servers consuming client certificates
1526 - Hosting providers taking certificates or private keys from customers
1527 - Certificate authorities parsing certification requests from subscribers
1528 - Anything else which parses ASN.1 elliptic curve parameters
1530 Also any other applications that use the BN_mod_sqrt() where the attacker
1531 can control the parameter values are vulnerable to this DoS issue.
1536 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1537 to the list of ciphersuites providing Perfect Forward Secrecy as
1538 required by SECLEVEL >= 3.
1540 *Dmitry Belyavskiy, Nicola Tuveri*
1542 * Made the AES constant time code for no-asm configurations
1543 optional due to the resulting 95% performance degradation.
1544 The AES constant time code can be enabled, for no assembly
1545 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1549 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1554 * The negative return value handling of the certificate verification callback
1555 was reverted. The replacement is to set the verification retry state with
1556 the SSL_set_retry_verify() function.
1560 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1562 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1563 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1564 verify a certificate supplied by a server. That function may return a
1565 negative return value to indicate an internal error (for example out of
1566 memory). Such a negative return value is mishandled by OpenSSL and will cause
1567 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1568 success and a subsequent call to SSL_get_error() to return the value
1569 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1570 returned by OpenSSL if the application has previously called
1571 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1572 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1573 totally unexpected and applications may not behave correctly as a result. The
1574 exact behaviour will depend on the application but it could result in
1575 crashes, infinite loops or other similar incorrect responses.
1577 This issue is made more serious in combination with a separate bug in OpenSSL
1578 3.0 that will cause X509_verify_cert() to indicate an internal error when
1579 processing a certificate chain. This will occur where a certificate does not
1580 include the Subject Alternative Name extension but where a Certificate
1581 Authority has enforced name constraints. This issue can occur even with valid
1587 * Corrected a few file name and file reference bugs in the build,
1588 installation and setup scripts, which lead to installation verification
1589 failures. Slightly enhanced the installation verification script.
1593 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1598 * Fixed PVK encoder to properly query for the passphrase.
1602 * Multiple fixes in the OSSL_HTTP API functions.
1606 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1607 OSSL_PARAM_INTEGER data type and return error on negative numbers
1608 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1609 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1613 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1617 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1621 * Multiple threading fixes.
1625 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1629 * Allow fetching an operation from the provider that owns an unexportable key
1630 as a fallback if that is still allowed by the property query.
1634 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1636 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1641 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1642 S390X capability vector to zero. This simplifies testing of different code
1643 paths on S390X architecture.
1647 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1648 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1649 SP 800-38D". The communication will fail at this point.
1653 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1654 confidential in EC_GROUP data.
1658 * The byte order mark (BOM) character is ignored if encountered at the
1659 beginning of a PEM-formatted file.
1663 * Added CMS support for the Russian GOST algorithms.
1667 * Due to move of the implementation of cryptographic operations
1668 to the providers, validation of various operation parameters can
1669 be postponed until the actual operation is executed where previously
1670 it happened immediately when an operation parameter was set.
1672 For example when setting an unsupported curve with
1673 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1674 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1676 *OpenSSL team members and many third party contributors*
1678 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1679 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1680 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1681 instead to retrieve these algorithms from a provider.
1685 * On build targets where the multilib postfix is set in the build
1686 configuration the libdir directory was changing based on whether
1687 the lib directory with the multilib postfix exists on the system
1688 or not. This unpredictable behavior was removed and eventual
1689 multilib postfix is now always added to the default libdir. Use
1690 `--libdir=lib` to override the libdir if adding the postfix is
1695 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1696 no longer interoperable with OpenSSL 1.1.1.
1700 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1701 function codes, this function can only cause problems for calling
1706 * Add a configurable flag to output date formats as ISO 8601. Does not
1707 change the default date format.
1711 * Version of MSVC earlier than 1300 could get link warnings, which could
1712 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1713 Support for this flag has been removed.
1717 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1718 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1719 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1720 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1721 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1725 * The signatures of the functions to get and set options on SSL and
1726 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1727 Some source code changes may be required.
1731 * The public definitions of conf_method_st and conf_st have been
1732 deprecated. They will be made opaque in a future release.
1734 *Rich Salz and Tomáš Mráz*
1736 * Client-initiated renegotiation is disabled by default. To allow it, use
1737 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1738 flag, or the "ClientRenegotiation" config parameter as appropriate.
1742 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1743 or modify relative pathname inclusion.
1747 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1748 validated. Please consult the README-FIPS and
1749 README-PROVIDERS files, as well as the migration guide.
1751 *OpenSSL team members and many third party contributors*
1753 * For the key types DH and DHX the allowed settable parameters are now different.
1757 * The openssl commands that read keys, certificates, and CRLs now
1758 automatically detect the PEM or DER format of the input files.
1760 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1762 * Added enhanced PKCS#12 APIs which accept a library context.
1766 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1770 * Added support for Kernel TLS (KTLS).
1772 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1774 * Support for RFC 5746 secure renegotiation is now required by default for
1775 SSL or TLS connections to succeed.
1779 * The signature of the `copy` functional parameter of the
1780 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1781 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1782 the signature of the `pub_decode` functional parameter of the
1783 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1784 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1788 * The error return values from some control calls (ctrl) have changed.
1792 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1796 * Many functions in the EVP_ namespace that are getters of values from
1797 implementations or contexts were renamed to include get or get0 in their
1798 names. Old names are provided as macro aliases for compatibility and
1803 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1804 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1805 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1810 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1813 * The output from the command line applications may have minor
1818 * The output from numerous "printing" may have minor changes.
1822 * Windows thread synchronization uses read/write primitives (SRWLock) when
1823 supported by the OS, otherwise CriticalSection continues to be used.
1827 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1828 work on read only BIO source/sinks that do not support these functions.
1829 This allows piping or redirection of a file BIO using stdin to be buffered
1830 into memory. This is used internally in OSSL_DECODER_from_bio().
1834 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1835 this function would return one of the values OSSL_STORE_INFO_NAME,
1836 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1837 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1838 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1839 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1840 using this function should be amended to handle the changed return value.
1844 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1845 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1846 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1847 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1848 contains more than one certificate identifier: This means that all
1849 certificates referenced there MUST be part of the validation chain.
1853 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1854 RC5, DESX and DES have been moved to the legacy provider.
1858 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1859 RIPEMD-160 have been moved to the legacy provider.
1863 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1868 * The deprecated functions EVP_PKEY_get0_RSA(),
1869 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1870 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1871 well as the similarly named "get1" functions behave differently in
1876 * A number of functions handling low-level keys or engines were deprecated
1877 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1878 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1879 EVP_PKEY_get0_siphash().
1883 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1884 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1885 will need to load the legacy crypto provider. This includes these PBE
1886 algorithms which use this KDF:
1887 - NID_pbeWithMD2AndDES_CBC
1888 - NID_pbeWithMD5AndDES_CBC
1889 - NID_pbeWithSHA1AndRC2_CBC
1890 - NID_pbeWithMD2AndRC2_CBC
1891 - NID_pbeWithMD5AndRC2_CBC
1892 - NID_pbeWithSHA1AndDES_CBC
1896 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1897 BIO_debug_callback() functions.
1901 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1902 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1906 * The RAND_METHOD APIs have been deprecated.
1910 * The SRP APIs have been deprecated.
1914 * Add a compile time option to prevent the caching of provider fetched
1915 algorithms. This is enabled by including the no-cached-fetch option
1916 at configuration time.
1920 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1921 count of PKCS12_DEFAULT_ITER.
1923 *Tomáš Mráz and Sahana Prasad*
1925 * The openssl speed command does not use low-level API calls anymore.
1929 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1932 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1934 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1938 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1939 providers may supply their own group implementations (using either the "key
1940 exchange" or the "key encapsulation" methods) which will automatically be
1941 detected and used by libssl.
1943 *Matt Caswell, Nicola Tuveri*
1945 * The undocumented function X509_certificate_type() has been deprecated;
1949 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1953 * Removed RSA padding mode for SSLv23 (which was only used for
1954 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1955 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1960 * Deprecated the obsolete X9.31 RSA key generation related functions.
1962 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1963 is not allowed to return a value > 1, this is no more taken as failure.
1965 *Viktor Dukhovni and David von Oheimb*
1967 * Deprecated the obsolete X9.31 RSA key generation related functions
1968 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1969 BN_X931_generate_prime_ex().
1973 * The default key generation method for the regular 2-prime RSA keys was
1974 changed to the FIPS 186-4 B.3.6 method.
1978 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1982 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1986 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1987 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1989 *Rich Salz, Richard Levitte, and David von Oheimb*
1991 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1995 * Deprecated `OCSP_parse_url()`.
1999 * Validation of SM2 keys has been separated from the validation of regular EC
2004 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
2005 switches: a validation failure triggers an early exit, returning a failure
2006 exit status to the parent process.
2010 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
2011 to ignore unknown ciphers.
2015 * The `-cipher-commands` and `-digest-commands` options
2016 of the command line utility `list` have been deprecated.
2017 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
2021 * Added convenience functions for generating asymmetric key pairs:
2022 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
2023 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
2027 * All of the low-level EC_KEY functions have been deprecated.
2029 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
2031 * Deprecated all the libcrypto and libssl error string loading
2036 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
2037 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
2042 * The `-crypt` option to the `passwd` command line tool has been removed.
2046 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
2051 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
2055 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
2056 EVP_PKEY_get1_tls_encodedpoint().
2060 * The security callback, which can be customised by application code, supports
2061 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
2062 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
2066 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
2067 interface. Their functionality remains unchanged.
2071 * Added new option for 'openssl list', '-providers', which will display the
2072 list of loaded providers, their names, version and status. It optionally
2073 displays their gettable parameters.
2077 * Removed EVP_PKEY_set_alias_type().
2081 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
2082 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
2086 * Changed all "STACK" functions to be macros instead of inline functions. Macro
2087 parameters are still checked for type safety at compile time via helper
2092 * Remove the RAND_DRBG API
2094 *Paul Dale and Matthias St. Pierre*
2096 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
2097 as well as actual hostnames.
2101 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2102 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2103 conversely, silently ignore DTLS protocol version bounds when configuring
2104 TLS-based contexts. The commands can be repeated to set bounds of both
2105 types. The same applies with the corresponding "min_protocol" and
2106 "max_protocol" command-line switches, in case some application uses both TLS
2109 SSL_CTX instances that are created for a fixed protocol version (e.g.
2110 `TLSv1_server_method()`) also silently ignore version bounds. Previously
2111 attempts to apply bounds to these protocol versions would result in an
2112 error. Now only the "version-flexible" SSL_CTX instances are subject to
2113 limits in configuration files in command-line options.
2117 * Deprecated the `ENGINE` API. Engines should be replaced with providers
2122 * Reworked the recorded ERR codes to make better space for system errors.
2123 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
2124 given code is a system error (true) or an OpenSSL error (false).
2128 * Reworked the test perl framework to better allow parallel testing.
2130 *Nicola Tuveri and David von Oheimb*
2132 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2133 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2137 * 'Configure' has been changed to figure out the configuration target if
2138 none is given on the command line. Consequently, the 'config' script is
2139 now only a mere wrapper. All documentation is changed to only mention
2142 *Rich Salz and Richard Levitte*
2144 * Added a library context `OSSL_LIB_CTX` that applications as well as
2145 other libraries can use to form a separate context within which
2146 libcrypto operations are performed.
2150 * Added various `_ex` functions to the OpenSSL API that support using
2151 a non-default `OSSL_LIB_CTX`.
2155 * Handshake now fails if Extended Master Secret extension is dropped
2160 * Dropped interactive mode from the `openssl` program.
2164 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2166 *David von Oheimb and Shane Lontis*
2168 * Deprecated `EC_METHOD_get_field_type()`.
2172 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2173 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2174 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2178 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2182 * Add CAdES-BES signature verification support, mostly derived
2183 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2185 *Filipe Raimundo da Silva*
2187 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2191 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2192 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2196 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2200 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2201 EC_KEY_precompute_mult().
2205 * Deprecated EC_POINTs_mul().
2209 * Removed FIPS_mode() and FIPS_mode_set().
2213 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2217 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2218 EC_POINT_get_Jprojective_coordinates_GFp().
2222 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2223 arrays to be more easily constructed via a series of utility functions.
2224 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2225 the various push functions and finally convert to a passable OSSL_PARAM
2226 array using OSSL_PARAM_BLD_to_param().
2230 * The security strength of SHA1 and MD5 based signatures in TLS has been
2235 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2236 contain a provider side internal key.
2240 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2244 * Project text documents not yet having a proper file name extension
2245 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2246 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2250 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2251 have been converted to Markdown with the goal to produce documents
2252 which not only look pretty when viewed online in the browser, but
2253 remain well readable inside a plain text editor.
2255 To achieve this goal, a 'minimalistic' Markdown style has been applied
2256 which avoids formatting elements that interfere too much with the
2257 reading flow in the text file. For example, it
2259 * avoids [ATX headings][] and uses [setext headings][] instead
2260 (which works for `<h1>` and `<h2>` headings only).
2261 * avoids [inline links][] and uses [reference links][] instead.
2262 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2264 [ATX headings]: https://github.github.com/gfm/#atx-headings
2265 [setext headings]: https://github.github.com/gfm/#setext-headings
2266 [inline links]: https://github.github.com/gfm/#inline-link
2267 [reference links]: https://github.github.com/gfm/#reference-link
2268 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2269 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2271 *Matthias St. Pierre*
2273 * The test suite is changed to preserve results of each test recipe.
2274 A new directory test-runs/ with subdirectories named like the
2275 test recipes are created in the build tree for this purpose.
2279 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2280 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2281 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2283 *David von Oheimb, Martin Peylo*
2285 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2286 It supports arbitrary request and response content types, GET redirection,
2287 TLS, connections via HTTP(S) proxies, connections and exchange via
2288 user-defined BIOs (allowing implicit connections), persistent connections,
2289 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2290 The legacy OCSP-focused (and only partly documented) API
2291 is retained for backward compatibility, while most of it is deprecated.
2295 * Added `util/check-format.pl`, a tool for checking adherence to the
2296 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2297 The checks performed are incomplete and yield some false positives.
2298 Still the tool should be useful for detecting most typical glitches.
2302 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2303 If domain name resolution yields multiple IP addresses all of them are tried
2304 after `connect()` failures.
2308 * All of the low-level RSA functions have been deprecated.
2312 * X509 certificates signed using SHA1 are no longer allowed at security
2317 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2318 modified to use PKEY APIs. These commands are now in maintenance mode
2319 and no new features will be added to them.
2323 * The command line utility rsautl has been deprecated.
2327 * The command line utilities genrsa and rsa have been modified to use PKEY
2328 APIs. They now write PKCS#8 keys by default. These commands are now in
2329 maintenance mode and no new features will be added to them.
2333 * All of the low-level DH functions have been deprecated.
2335 *Paul Dale and Matt Caswell*
2337 * All of the low-level DSA functions have been deprecated.
2341 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2342 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2346 * Deprecated low-level ECDH and ECDSA functions.
2350 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2354 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2355 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2356 a new formulation to include all the things it can be used for,
2357 as well as words of caution.
2361 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2365 * All of the low-level HMAC functions have been deprecated.
2367 *Paul Dale and David von Oheimb*
2369 * Over two thousand fixes were made to the documentation, including:
2370 - Common options (such as -rand/-writerand, TLS version control, etc)
2371 were refactored and point to newly-enhanced descriptions in openssl.pod.
2372 - Added style conformance for all options (with help from Richard Levitte),
2373 documented all reported missing options, added a CI build to check
2374 that all options are documented and that no unimplemented options
2376 - Documented some internals, such as all use of environment variables.
2377 - Addressed all internal broken L<> references.
2381 * All of the low-level CMAC functions have been deprecated.
2385 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2386 functions have been deprecated.
2388 *Paul Dale and David von Oheimb*
2390 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2391 set of functions. The documentation mentioned negative values for some
2392 errors, but this was never the case, so the mention of negative values
2395 Code that followed the documentation and thereby check with something
2396 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2400 * All of the low-level cipher functions have been deprecated.
2402 *Matt Caswell and Paul Dale*
2404 * Removed include/openssl/opensslconf.h.in and replaced it with
2405 include/openssl/configuration.h.in, which differs in not including
2406 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2407 was added to include both.
2409 This allows internal hacks where one might need to modify the set
2410 of configured macros, for example this if deprecated symbols are
2411 still supposed to be available internally:
2413 #include <openssl/configuration.h>
2415 #undef OPENSSL_NO_DEPRECATED
2416 #define OPENSSL_SUPPRESS_DEPRECATED
2418 #include <openssl/macros.h>
2420 This should not be used by applications that use the exported
2421 symbols, as that will lead to linking errors.
2425 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2426 used in exponentiation with 512-bit moduli. No EC algorithms are
2427 affected. Analysis suggests that attacks against 2-prime RSA1024,
2428 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2429 difficult to perform and are not believed likely. Attacks against DH512
2430 are considered just feasible. However, for an attack the target would
2431 have to reuse the DH512 private key, which is not recommended anyway.
2432 Also applications directly using the low-level API BN_mod_exp may be
2433 affected if they use BN_FLG_CONSTTIME.
2438 * Most memory-debug features have been deprecated, and the functionality
2439 replaced with no-ops.
2443 * Added documentation for the STACK API.
2447 * Introduced a new method type and API, OSSL_ENCODER, to represent
2448 generic encoders. These do the same sort of job that PEM writers
2449 and d2i functions do, but with support for methods supplied by
2450 providers, and the possibility for providers to support other
2455 * Introduced a new method type and API, OSSL_DECODER, to represent
2456 generic decoders. These do the same sort of job that PEM readers
2457 and i2d functions do, but with support for methods supplied by
2458 providers, and the possibility for providers to support other
2463 * Added a .pragma directive to the syntax of configuration files, to
2464 allow varying behavior in a supported and predictable manner.
2465 Currently added pragma:
2469 This allows dollar signs to be a keyword character unless it's
2470 followed by a opening brace or parenthesis. This is useful for
2471 platforms where dollar signs are commonly used in names, such as
2472 volume names and system directory names on VMS.
2476 * Added functionality to create an EVP_PKEY from user data.
2480 * Change the interpretation of the '--api' configuration option to
2481 mean that this is a desired API compatibility level with no
2482 further meaning. The previous interpretation, that this would
2483 also mean to remove all deprecated symbols up to and including
2484 the given version, no requires that 'no-deprecated' is also used
2485 in the configuration.
2487 When building applications, the desired API compatibility level
2488 can be set with the OPENSSL_API_COMPAT macro like before. For
2489 API compatibility version below 3.0, the old style numerical
2490 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2491 For version 3.0 and on, the value is expected to be the decimal
2492 value calculated from the major and minor version like this:
2494 MAJOR * 10000 + MINOR * 100
2498 -DOPENSSL_API_COMPAT=30000 For 3.0
2499 -DOPENSSL_API_COMPAT=30200 For 3.2
2501 To hide declarations that are deprecated up to and including the
2502 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2503 given when building the application as well.
2507 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2508 access to certificate and CRL stores via URIs and OSSL_STORE
2511 This adds the following functions:
2513 - X509_LOOKUP_store()
2514 - X509_STORE_load_file()
2515 - X509_STORE_load_path()
2516 - X509_STORE_load_store()
2517 - SSL_add_store_cert_subjects_to_stack()
2518 - SSL_CTX_set_default_verify_store()
2519 - SSL_CTX_load_verify_file()
2520 - SSL_CTX_load_verify_dir()
2521 - SSL_CTX_load_verify_store()
2525 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2526 The presence of this system service is determined at run-time.
2530 * Added functionality to create an EVP_PKEY context based on data
2531 for methods from providers. This takes an algorithm name and a
2532 property query string and simply stores them, with the intent
2533 that any operation that uses this context will use those strings
2534 to fetch the needed methods implicitly, thereby making the port
2535 of application written for pre-3.0 OpenSSL easier.
2539 * The undocumented function NCONF_WIN32() has been deprecated; for
2540 conversion details see the HISTORY section of doc/man5/config.pod
2544 * Introduced the new functions EVP_DigestSignInit_ex() and
2545 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2546 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2547 pages for further details.
2551 * Over two thousand fixes were made to the documentation, including:
2552 adding missing command flags, better style conformance, documentation
2555 *Rich Salz, Richard Levitte*
2557 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2558 X25519, X448, Ed25519 and Ed448.
2562 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2567 * Deprecated the public definition of `ERR_STATE` as well as the function
2568 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2573 * Added ERR functionality to give callers access to the stored function
2574 names that have replaced the older function code based functions.
2576 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2577 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2578 ERR_peek_error_all() and ERR_peek_last_error_all().
2580 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2581 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2582 ERR_func_error_string().
2586 * Extended testing to be verbose for failing tests only. The make variables
2587 VERBOSE_FAILURE or VF can be used to enable this:
2589 $ make VF=1 test # Unix
2590 $ mms /macro=(VF=1) test ! OpenVMS
2591 $ nmake VF=1 test # Windows
2595 * Added the `-copy_extensions` option to the `x509` command for use with
2596 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2597 all extensions in the request are copied to the certificate or vice versa.
2599 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2601 * Added the `-copy_extensions` option to the `req` command for use with
2602 `-x509`. When given with the `copy` or `copyall` argument,
2603 all extensions in the certification request are copied to the certificate.
2607 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2608 they generate are by default RFC 5280 compliant in the following sense:
2609 There is a subjectKeyIdentifier extension with a hash value of the public key
2610 and for not self-signed certs there is an authorityKeyIdentifier extension
2611 with a keyIdentifier field or issuer information identifying the signing key.
2612 This is done unless some configuration overrides the new default behavior,
2613 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2617 * Added several checks to `X509_verify_cert()` according to requirements in
2618 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2619 (which may be done by using the CLI option `-x509_strict`):
2620 * The basicConstraints of CA certificates must be marked critical.
2621 * CA certificates must explicitly include the keyUsage extension.
2622 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2623 * The issuer name of any certificate must not be empty.
2624 * The subject name of CA certs, certs with keyUsage crlSign,
2625 and certs without subjectAlternativeName must not be empty.
2626 * If a subjectAlternativeName extension is given it must not be empty.
2627 * The signatureAlgorithm field and the cert signature must be consistent.
2628 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2629 must not be marked critical.
2630 * The authorityKeyIdentifier must be given for X.509v3 certs
2631 unless they are self-signed.
2632 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2636 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2637 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2641 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2642 used even when parsing explicit parameters, when loading a encoded key
2643 or calling `EC_GROUP_new_from_ecpkparameters()`/
2644 `EC_GROUP_new_from_ecparameters()`.
2645 This prevents bypass of security hardening and performance gains,
2646 especially for curves with specialized EC_METHODs.
2647 By default, if a key encoded with explicit parameters is loaded and later
2648 encoded, the output is still encoded with explicit parameters, even if
2649 internally a "named" EC_GROUP is used for computation.
2653 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2654 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2655 NULL. After this change, only the cofactor parameter can be NULL. It also
2656 does some minimal sanity checks on the passed order.
2661 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2662 An attack is simple, if the first CMS_recipientInfo is valid but the
2663 second CMS_recipientInfo is chosen ciphertext. If the second
2664 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2665 encryption key will be replaced by garbage, and the message cannot be
2666 decoded, but if the RSA decryption fails, the correct encryption key is
2667 used and the recipient will not notice the attack.
2668 As a work around for this potential attack the length of the decrypted
2669 key must be equal to the cipher default key length, in case the
2670 certificate is not given and all recipientInfo are tried out.
2671 The old behaviour can be re-enabled in the CMS code by setting the
2672 CMS_DEBUG_DECRYPT flag.
2676 * Early start up entropy quality from the DEVRANDOM seed source has been
2677 improved for older Linux systems. The RAND subsystem will wait for
2678 /dev/random to be producing output before seeding from /dev/urandom.
2679 The seeded state is stored for future library initialisations using
2680 a system global shared memory segment. The shared memory identifier
2681 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2682 the desired value. The default identifier is 114.
2686 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2687 when primes for RSA keys are computed.
2688 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2689 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2690 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2691 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2692 This avoids possible fingerprinting of newly generated RSA modules.
2696 * Correct the extended master secret constant on EBCDIC systems. Without this
2697 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2698 negotiate EMS will fail. Unfortunately this also means that TLS connections
2699 between EBCDIC systems with this fix, and EBCDIC systems without this
2700 fix will fail if they negotiate EMS.
2704 * Changed the library initialisation so that the config file is now loaded
2705 by default. This was already the case for libssl. It now occurs for both
2706 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2707 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2711 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2712 where the former acts as a replacement for `ERR_put_error()`, and the
2713 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2714 `ERR_raise_data()` adds more flexibility by taking a format string and
2715 an arbitrary number of arguments following it, to be processed with
2720 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2721 to check if a named provider is loaded and available. When called, it
2722 will also activate all fallback providers if such are still present.
2726 * Enforce a minimum DH modulus size of 512 bits.
2730 * Changed DH parameters to generate the order q subgroup instead of 2q.
2731 Previously generated DH parameters are still accepted by DH_check
2732 but DH_generate_key works around that by clearing bit 0 of the
2733 private key for those. This avoids leaking bit 0 of the private key.
2737 * Significantly reduce secure memory usage by the randomness pools.
2741 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2746 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2747 algorithms. An implementation of a key exchange algorithm can be obtained
2748 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2749 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2750 the older EVP_PKEY_derive_init() function. See the man pages for the new
2751 functions for further details.
2755 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2759 * Removed the function names from error messages and deprecated the
2764 * Removed NextStep support and the macro OPENSSL_UNISTD
2768 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2769 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2770 Also removed "export var as function" capability; we do not export
2771 variables, only functions.
2775 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2776 an error and 1 indicating success. In previous versions of OpenSSL this
2777 was a void type. If a key was set longer than the maximum possible this
2782 * Support SM2 signing and verification schemes with X509 certificate.
2786 * Use SHA256 as the default digest for TS query in the `ts` app.
2790 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2794 * Default cipher lists/suites are now available via a function, the
2795 #defines are deprecated.
2799 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2800 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2801 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2805 * Join the directories crypto/x509 and crypto/x509v3
2809 * Added command 'openssl kdf' that uses the EVP_KDF API.
2813 * Added command 'openssl mac' that uses the EVP_MAC API.
2817 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2818 as default directories. Also added the command 'openssl info'
2819 for scripting purposes.
2823 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2828 * Add prediction resistance to the DRBG reseeding process.
2832 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2833 mandated by IEEE Std 1619-2018.
2837 * Added newline escaping functionality to a filename when using openssl dgst.
2838 This output format is to replicate the output format found in the `*sum`
2839 checksum programs. This aims to preserve backward compatibility.
2841 *Matt Eaton, Richard Levitte, and Paul Dale*
2843 * Removed the heartbeat message in DTLS feature, as it has very
2844 little usage and doesn't seem to fulfill a valuable purpose.
2845 The configuration option is now deprecated.
2849 * Changed the output of 'openssl {digestname} < file' to display the
2850 digest name in its output.
2854 * Added a new generic trace API which provides support for enabling
2855 instrumentation through trace output.
2857 *Richard Levitte & Matthias St. Pierre*
2859 * Added build tests for C++. These are generated files that only do one
2860 thing, to include one public OpenSSL head file each. This tests that
2861 the public header files can be usefully included in a C++ application.
2863 This test isn't enabled by default. It can be enabled with the option
2864 'enable-buildtest-c++'.
2868 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2872 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2876 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2880 * Added KMAC to EVP_MAC.
2884 * Added property based algorithm implementation selection framework to
2889 * Added SCA hardening for modular field inversion in EC_GROUP through
2890 a new dedicated field_inv() pointer in EC_METHOD.
2891 This also addresses a leakage affecting conversions from projective
2892 to affine coordinates.
2894 *Billy Bob Brumley, Nicola Tuveri*
2896 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2897 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2898 those algorithms that were already supported through the EVP_PKEY API
2899 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2900 and scrypt are now wrappers that call EVP_KDF.
2904 * Build devcrypto engine as a dynamic engine.
2906 *Eneas U de Queiroz*
2908 * Add keyed BLAKE2 to EVP_MAC.
2912 * Fix a bug in the computation of the endpoint-pair shared secret used
2913 by DTLS over SCTP. This breaks interoperability with older versions
2914 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2915 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2916 interoperability with such broken implementations. However, enabling
2917 this switch breaks interoperability with correct implementations.
2919 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2920 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2924 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2928 * Changed the license to the Apache License v2.0.
2932 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2934 - Major releases (indicated by incrementing the MAJOR release number)
2935 may introduce incompatible API/ABI changes.
2936 - Minor releases (indicated by incrementing the MINOR release number)
2937 may introduce new features but retain API/ABI compatibility.
2938 - Patch releases (indicated by incrementing the PATCH number)
2939 are intended for bug fixes and other improvements of existing
2940 features only (like improving performance or adding documentation)
2941 and retain API/ABI compatibility.
2945 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2949 * Remove the 'dist' target and add a tarball building script. The
2950 'dist' target has fallen out of use, and it shouldn't be
2951 necessary to configure just to create a source distribution.
2955 * Recreate the OS390-Unix config target. It no longer relies on a
2956 special script like it did for OpenSSL pre-1.1.0.
2960 * Instead of having the source directories listed in Configure, add
2961 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2966 * Add GMAC to EVP_MAC.
2970 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2974 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2975 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2976 to facilitate the continued use of MACs through raw private keys in
2977 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2981 * Deprecate ECDH_KDF_X9_62().
2985 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2986 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2987 are retained for backwards compatibility.
2991 * AES-XTS mode now enforces that its two keys are different to mitigate
2992 the attacked described in "Efficient Instantiations of Tweakable
2993 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2994 Details of this attack can be obtained from:
2995 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2999 * Rename the object files, i.e. give them other names than in previous
3000 versions. Their names now include the name of the final product, as
3001 well as its type mnemonic (bin, lib, shlib).
3005 * Added new option for 'openssl list', '-objects', which will display the
3006 list of built in objects, i.e. OIDs with names.
3010 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
3011 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
3016 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
3017 improves application performance by removing data copies and providing
3018 applications with zero-copy system calls such as sendfile and splice.
3022 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
3026 * `PKCS12_parse` now maintains the order of the parsed certificates
3027 when outputting them via `*ca` (rather than reversing it).
3031 * Deprecated pthread fork support methods.
3035 * Added support for FFDHE key exchange in TLS 1.3.
3039 * Added a new concept for OpenSSL plugability: providers. This
3040 functionality is designed to replace the ENGINE API and ENGINE
3041 implementations, and to be much more dynamic, allowing provider
3042 authors to introduce new algorithms among other things, as long as
3043 there's an API that supports the algorithm type.
3045 With this concept comes a new core API for interaction between
3046 libcrypto and provider implementations. Public libcrypto functions
3047 that want to use providers do so through this core API.
3049 The main documentation for this core API is found in
3050 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
3051 refer to other manuals describing the API specific for supported
3052 algorithm types (also called operations).
3059 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
3061 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
3063 * Avoid loading of a dynamic engine twice.
3067 * Prioritise DANE TLSA issuer certs over peer certs
3071 * Fixed random API for MacOS prior to 10.12
3073 These MacOS versions don't support the CommonCrypto APIs
3077 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
3079 * Fixed an SM2 Decryption Buffer Overflow.
3081 In order to decrypt SM2 encrypted data an application is expected to
3082 call the API function EVP_PKEY_decrypt(). Typically an application will
3083 call this function twice. The first time, on entry, the "out" parameter
3084 can be NULL and, on exit, the "outlen" parameter is populated with the
3085 buffer size required to hold the decrypted plaintext. The application
3086 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
3087 again, but this time passing a non-NULL value for the "out" parameter.
3089 A bug in the implementation of the SM2 decryption code means that the
3090 calculation of the buffer size required to hold the plaintext returned
3091 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
3092 size required by the second call. This can lead to a buffer overflow
3093 when EVP_PKEY_decrypt() is called by the application a second time with
3094 a buffer that is too small.
3096 A malicious attacker who is able present SM2 content for decryption to
3097 an application could cause attacker chosen data to overflow the buffer
3098 by up to a maximum of 62 bytes altering the contents of other data held
3099 after the buffer, possibly changing application behaviour or causing
3100 the application to crash. The location of the buffer is application
3101 dependent but is typically heap allocated.
3106 * Fixed various read buffer overruns processing ASN.1 strings
3108 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
3109 structure which contains a buffer holding the string data and a field
3110 holding the buffer length. This contrasts with normal C strings which
3111 are represented as a buffer for the string data which is terminated
3112 with a NUL (0) byte.
3114 Although not a strict requirement, ASN.1 strings that are parsed using
3115 OpenSSL's own "d2i" functions (and other similar parsing functions) as
3116 well as any string whose value has been set with the ASN1_STRING_set()
3117 function will additionally NUL terminate the byte array in the
3118 ASN1_STRING structure.
3120 However, it is possible for applications to directly construct valid
3121 ASN1_STRING structures which do not NUL terminate the byte array by
3122 directly setting the "data" and "length" fields in the ASN1_STRING
3123 array. This can also happen by using the ASN1_STRING_set0() function.
3125 Numerous OpenSSL functions that print ASN.1 data have been found to
3126 assume that the ASN1_STRING byte array will be NUL terminated, even
3127 though this is not guaranteed for strings that have been directly
3128 constructed. Where an application requests an ASN.1 structure to be
3129 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3130 been directly constructed by the application without NUL terminating
3131 the "data" field, then a read buffer overrun can occur.
3133 The same thing can also occur during name constraints processing
3134 of certificates (for example if a certificate has been directly
3135 constructed by the application instead of loading it via the OpenSSL
3136 parsing functions, and the certificate contains non NUL terminated
3137 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3138 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3140 If a malicious actor can cause an application to directly construct an
3141 ASN1_STRING and then process it through one of the affected OpenSSL
3142 functions then this issue could be hit. This might result in a crash
3143 (causing a Denial of Service attack). It could also result in the
3144 disclosure of private memory contents (such as private keys, or
3145 sensitive plaintext).
3150 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3152 * Fixed a problem with verifying a certificate chain when using the
3153 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3154 the certificates present in a certificate chain. It is not set by default.
3156 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3157 the chain that have explicitly encoded elliptic curve parameters was added
3158 as an additional strict check.
3160 An error in the implementation of this check meant that the result of a
3161 previous check to confirm that certificates in the chain are valid CA
3162 certificates was overwritten. This effectively bypasses the check
3163 that non-CA certificates must not be able to issue other certificates.
3165 If a "purpose" has been configured then there is a subsequent opportunity
3166 for checks that the certificate is a valid CA. All of the named "purpose"
3167 values implemented in libcrypto perform this check. Therefore, where
3168 a purpose is set the certificate chain will still be rejected even when the
3169 strict flag has been used. A purpose is set by default in libssl client and
3170 server certificate verification routines, but it can be overridden or
3171 removed by an application.
3173 In order to be affected, an application must explicitly set the
3174 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3175 for the certificate verification or, in the case of TLS client or server
3176 applications, override the default purpose.
3181 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3182 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3183 renegotiation ClientHello omits the signature_algorithms extension (where it
3184 was present in the initial ClientHello), but includes a
3185 signature_algorithms_cert extension then a NULL pointer dereference will
3186 result, leading to a crash and a denial of service attack.
3188 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3189 (which is the default configuration). OpenSSL TLS clients are not impacted by
3193 *Peter Kästle and Samuel Sapalski*
3195 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3197 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3198 create a unique hash value based on the issuer and serial number data
3199 contained within an X509 certificate. However, it was failing to correctly
3200 handle any errors that may occur while parsing the issuer field (which might
3201 occur if the issuer field is maliciously constructed). This may subsequently
3202 result in a NULL pointer deref and a crash leading to a potential denial of
3208 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3209 padding mode to correctly check for rollback attacks. This is considered a
3210 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3215 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3216 functions. Previously they could overflow the output length argument in some
3217 cases where the input length is close to the maximum permissible length for
3218 an integer on the platform. In such cases the return value from the function
3219 call would be 1 (indicating success), but the output length value would be
3220 negative. This could cause applications to behave incorrectly or crash.
3225 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3226 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3227 could be exploited in a side channel attack to recover the password. Since
3228 the attack is local host only this is outside of the current OpenSSL
3229 threat model and therefore no CVE is assigned.
3231 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3236 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3238 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3239 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3240 If an attacker can control both items being compared then this could lead
3241 to a possible denial of service attack. OpenSSL itself uses the
3242 GENERAL_NAME_cmp function for two purposes:
3243 1) Comparing CRL distribution point names between an available CRL and a
3244 CRL distribution point embedded in an X509 certificate
3245 2) When verifying that a timestamp response token signer matches the
3246 timestamp authority name (exposed via the API functions
3247 TS_RESP_verify_response and TS_RESP_verify_token)
3252 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3254 * Certificates with explicit curve parameters are now disallowed in
3255 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3259 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3260 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3261 conversely, silently ignore DTLS protocol version bounds when configuring
3262 TLS-based contexts. The commands can be repeated to set bounds of both
3263 types. The same applies with the corresponding "min_protocol" and
3264 "max_protocol" command-line switches, in case some application uses both TLS
3267 SSL_CTX instances that are created for a fixed protocol version (e.g.
3268 TLSv1_server_method()) also silently ignore version bounds. Previously
3269 attempts to apply bounds to these protocol versions would result in an
3270 error. Now only the "version-flexible" SSL_CTX instances are subject to
3271 limits in configuration files in command-line options.
3275 * Handshake now fails if Extended Master Secret extension is dropped
3280 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3282 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3284 * Fixed segmentation fault in SSL_check_chain()
3285 Server or client applications that call the SSL_check_chain() function
3286 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3287 dereference as a result of incorrect handling of the
3288 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3289 or unrecognised signature algorithm is received from the peer. This could
3290 be exploited by a malicious peer in a Denial of Service attack.
3295 * Added AES consttime code for no-asm configurations
3296 an optional constant time support for AES was added
3297 when building openssl for no-asm.
3298 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3299 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3300 At this time this feature is by default disabled.
3301 It will be enabled by default in 3.0.
3305 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3307 * Revert the change of EOF detection while reading in libssl to avoid
3308 regressions in applications depending on the current way of reporting
3309 the EOF. As the existing method is not fully accurate the change to
3310 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3311 branch and will be present in the 3.0 release.
3315 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3316 when primes for RSA keys are computed.
3317 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3318 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3319 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3320 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3321 This avoids possible fingerprinting of newly generated RSA modules.
3325 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3327 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3328 while reading in libssl then we would report an error back to the
3329 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3330 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3331 therefore give a hint as to what went wrong.
3335 * Check that ed25519 and ed448 are allowed by the security level. Previously
3336 signature algorithms not using an MD were not being checked that they were
3337 allowed by the security level.
3341 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3342 was not quite right. The behaviour was not consistent between resumption
3343 and normal handshakes, and also not quite consistent with historical
3344 behaviour. The behaviour in various scenarios has been clarified and
3345 it has been updated to make it match historical behaviour as closely as
3350 * *[VMS only]* The header files that the VMS compilers include automatically,
3351 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3352 that the C++ compiler doesn't understand. This is a shortcoming in the
3353 compiler, but can be worked around with `__cplusplus` guards.
3355 C++ applications that use OpenSSL libraries must be compiled using the
3356 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3357 functions. Otherwise, only functions with symbols of less than 31
3358 characters can be used, as the linker will not be able to successfully
3359 resolve symbols with longer names.
3363 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3364 The presence of this system service is determined at run-time.
3368 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3373 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3375 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3376 number generator (RNG). This was intended to include protection in the
3377 event of a fork() system call in order to ensure that the parent and child
3378 processes did not share the same RNG state. However, this protection was not
3379 being used in the default case.
3381 A partial mitigation for this issue is that the output from a high
3382 precision timer is mixed into the RNG state so the likelihood of a parent
3383 and child process sharing state is significantly reduced.
3385 If an application already calls OPENSSL_init_crypto() explicitly using
3386 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3389 *Matthias St. Pierre*
3391 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3392 used even when parsing explicit parameters, when loading a encoded key
3393 or calling `EC_GROUP_new_from_ecpkparameters()`/
3394 `EC_GROUP_new_from_ecparameters()`.
3395 This prevents bypass of security hardening and performance gains,
3396 especially for curves with specialized EC_METHODs.
3397 By default, if a key encoded with explicit parameters is loaded and later
3398 encoded, the output is still encoded with explicit parameters, even if
3399 internally a "named" EC_GROUP is used for computation.
3403 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3404 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3405 NULL. After this change, only the cofactor parameter can be NULL. It also
3406 does some minimal sanity checks on the passed order.
3411 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3412 An attack is simple, if the first CMS_recipientInfo is valid but the
3413 second CMS_recipientInfo is chosen ciphertext. If the second
3414 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3415 encryption key will be replaced by garbage, and the message cannot be
3416 decoded, but if the RSA decryption fails, the correct encryption key is
3417 used and the recipient will not notice the attack.
3418 As a work around for this potential attack the length of the decrypted
3419 key must be equal to the cipher default key length, in case the
3420 certificate is not given and all recipientInfo are tried out.
3421 The old behaviour can be re-enabled in the CMS code by setting the
3422 CMS_DEBUG_DECRYPT flag.
3427 * Early start up entropy quality from the DEVRANDOM seed source has been
3428 improved for older Linux systems. The RAND subsystem will wait for
3429 /dev/random to be producing output before seeding from /dev/urandom.
3430 The seeded state is stored for future library initialisations using
3431 a system global shared memory segment. The shared memory identifier
3432 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3433 the desired value. The default identifier is 114.
3437 * Correct the extended master secret constant on EBCDIC systems. Without this
3438 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3439 negotiate EMS will fail. Unfortunately this also means that TLS connections
3440 between EBCDIC systems with this fix, and EBCDIC systems without this
3441 fix will fail if they negotiate EMS.
3445 * Use Windows installation paths in the mingw builds
3447 Mingw isn't a POSIX environment per se, which means that Windows
3448 paths should be used for installation.
3453 * Changed DH_check to accept parameters with order q and 2q subgroups.
3454 With order 2q subgroups the bit 0 of the private key is not secret
3455 but DH_generate_key works around that by clearing bit 0 of the
3456 private key for those. This avoids leaking bit 0 of the private key.
3460 * Significantly reduce secure memory usage by the randomness pools.
3464 * Revert the DEVRANDOM_WAIT feature for Linux systems
3466 The DEVRANDOM_WAIT feature added a select() call to wait for the
3467 /dev/random device to become readable before reading from the
3468 /dev/urandom device.
3470 It turned out that this change had negative side effects on
3471 performance which were not acceptable. After some discussion it
3472 was decided to revert this feature and leave it up to the OS
3473 resp. the platform maintainer to ensure a proper initialization
3474 during early boot time.
3476 *Matthias St. Pierre*
3478 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3480 * Add build tests for C++. These are generated files that only do one
3481 thing, to include one public OpenSSL head file each. This tests that
3482 the public header files can be usefully included in a C++ application.
3484 This test isn't enabled by default. It can be enabled with the option
3485 'enable-buildtest-c++'.
3489 * Enable SHA3 pre-hashing for ECDSA and DSA.
3493 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3494 This changes the size when using the `genpkey` command when no size is given.
3495 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3496 generation commands to use 2048 bits by default.
3500 * Reorganize the manual pages to consistently have RETURN VALUES,
3501 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3502 util/fix-doc-nits accordingly.
3504 *Paul Yang, Joshua Lock*
3506 * Add the missing accessor EVP_PKEY_get0_engine()
3510 * Have commands like `s_client` and `s_server` output the signature scheme
3511 along with other cipher suite parameters when debugging.
3515 * Make OPENSSL_config() error agnostic again.
3519 * Do the error handling in RSA decryption constant time.
3523 * Prevent over long nonces in ChaCha20-Poly1305.
3525 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3526 for every encryption operation. RFC 7539 specifies that the nonce value
3527 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3528 and front pads the nonce with 0 bytes if it is less than 12
3529 bytes. However it also incorrectly allows a nonce to be set of up to 16
3530 bytes. In this case only the last 12 bytes are significant and any
3531 additional leading bytes are ignored.
3533 It is a requirement of using this cipher that nonce values are
3534 unique. Messages encrypted using a reused nonce value are susceptible to
3535 serious confidentiality and integrity attacks. If an application changes
3536 the default nonce length to be longer than 12 bytes and then makes a
3537 change to the leading bytes of the nonce expecting the new value to be a
3538 new unique nonce then such an application could inadvertently encrypt
3539 messages with a reused nonce.
3541 Additionally the ignored bytes in a long nonce are not covered by the
3542 integrity guarantee of this cipher. Any application that relies on the
3543 integrity of these ignored leading bytes of a long nonce may be further
3544 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3545 is safe because no such use sets such a long nonce value. However user
3546 applications that use this cipher directly and set a non-default nonce
3547 length to be longer than 12 bytes may be vulnerable.
3549 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3555 * Add DEVRANDOM_WAIT feature for Linux systems
3557 On older Linux systems where the getrandom() system call is not available,
3558 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3559 Contrary to getrandom(), the /dev/urandom device will not block during
3560 early boot when the kernel CSPRNG has not been seeded yet.
3562 To mitigate this known weakness, use select() to wait for /dev/random to
3563 become readable before reading from /dev/urandom.
3565 * Ensure that SM2 only uses SM3 as digest algorithm
3569 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3571 * Change the info callback signals for the start and end of a post-handshake
3572 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3573 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3574 confused by this and assume that a TLSv1.2 renegotiation has started. This
3575 can break KeyUpdate handling. Instead we no longer signal the start and end
3576 of a post handshake message exchange (although the messages themselves are
3577 still signalled). This could break some applications that were expecting
3578 the old signals. However without this KeyUpdate is not usable for many
3583 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3585 * Timing vulnerability in DSA signature generation
3587 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3588 timing side channel attack. An attacker could use variations in the signing
3589 algorithm to recover the private key.
3591 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3596 * Timing vulnerability in ECDSA signature generation
3598 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3599 timing side channel attack. An attacker could use variations in the signing
3600 algorithm to recover the private key.
3602 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3607 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3608 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3609 of two gigabytes and the error handling improved.
3611 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3612 categorized as a normal bug, not a security issue, because the DRBG reseeds
3613 automatically and is fully functional even without additional randomness
3614 provided by the application.
3616 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3618 * Add a new ClientHello callback. Provides a callback interface that gives
3619 the application the ability to adjust the nascent SSL object at the
3620 earliest stage of ClientHello processing, immediately after extensions have
3621 been collected but before they have been processed. In particular, this
3622 callback can adjust the supported TLS versions in response to the contents
3627 * Add SM2 base algorithm support.
3631 * s390x assembly pack: add (improved) hardware-support for the following
3632 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3633 aes-cfb/cfb8, aes-ecb.
3637 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3638 parameter is no longer accepted, as it leads to a corrupt table. NULL
3639 pem_str is reserved for alias entries only.
3643 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3644 step for prime curves. The new implementation is based on formulae from
3645 differential addition-and-doubling in homogeneous projective coordinates
3646 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3647 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3648 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3649 to work in projective coordinates.
3651 *Billy Bob Brumley, Nicola Tuveri*
3653 * Change generating and checking of primes so that the error rate of not
3654 being prime depends on the intended use based on the size of the input.
3655 For larger primes this will result in more rounds of Miller-Rabin.
3656 The maximal error rate for primes with more than 1080 bits is lowered
3659 *Kurt Roeckx, Annie Yousar*
3661 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3665 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3666 moving between systems, and to avoid confusion when a Windows build is
3667 done with mingw vs with MSVC. For POSIX installs, there's still a
3668 symlink or copy named 'tsget' to avoid that confusion as well.
3672 * Revert blinding in ECDSA sign and instead make problematic addition
3673 length-invariant. Switch even to fixed-length Montgomery multiplication.
3677 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3678 step for binary curves. The new implementation is based on formulae from
3679 differential addition-and-doubling in mixed Lopez-Dahab projective
3680 coordinates, modified to independently blind the operands.
3682 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3684 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3685 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3686 EC_METHODs to implement their own specialized "ladder step", to take
3687 advantage of more favorable coordinate systems or more efficient
3688 differential addition-and-doubling algorithms.
3690 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3692 * Modified the random device based seed sources to keep the relevant
3693 file descriptors open rather than reopening them on each access.
3694 This allows such sources to operate in a chroot() jail without
3695 the associated device nodes being available. This behaviour can be
3696 controlled using RAND_keep_random_devices_open().
3700 * Numerous side-channel attack mitigations have been applied. This may have
3701 performance impacts for some algorithms for the benefit of improved
3702 security. Specific changes are noted in this change log by their respective
3707 * AIX shared library support overhaul. Switch to AIX "natural" way of
3708 handling shared libraries, which means collecting shared objects of
3709 different versions and bitnesses in one common archive. This allows to
3710 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3711 doesn't affect the way 3rd party applications are linked, only how
3712 multi-version installation is managed.
3716 * Make ec_group_do_inverse_ord() more robust and available to other
3717 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3718 mitigations are applied to the fallback BN_mod_inverse().
3719 When using this function rather than BN_mod_inverse() directly, new
3720 EC cryptosystem implementations are then safer-by-default.
3724 * Add coordinate blinding for EC_POINT and implement projective
3725 coordinate blinding for generic prime curves as a countermeasure to
3726 chosen point SCA attacks.
3728 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3730 * Add blinding to ECDSA and DSA signatures to protect against side channel
3731 attacks discovered by Keegan Ryan (NCC Group).
3735 * Enforce checking in the `pkeyutl` command to ensure that the input
3736 length does not exceed the maximum supported digest length when performing
3737 a sign, verify or verifyrecover operation.
3741 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3742 I/O in combination with something like select() or poll() will hang. This
3743 can be turned off again using SSL_CTX_clear_mode().
3744 Many applications do not properly handle non-application data records, and
3745 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3746 around the problems in those applications, but can also break some.
3747 It's recommended to read the manpages about SSL_read(), SSL_write(),
3748 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3749 SSL_CTX_set_read_ahead() again.
3753 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3754 now allow empty (zero character) pass phrases.
3758 * Apply blinding to binary field modular inversion and remove patent
3759 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3763 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3764 binary and prime elliptic curves.
3768 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3769 constant time fixed point multiplication.
3773 * Revise elliptic curve scalar multiplication with timing attack
3774 defenses: ec_wNAF_mul redirects to a constant time implementation
3775 when computing fixed point and variable point multiplication (which
3776 in OpenSSL are mostly used with secret scalars in keygen, sign,
3777 ECDH derive operations).
3778 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3781 * Updated CONTRIBUTING
3785 * Updated DRBG / RAND to request nonce and additional low entropy
3786 randomness from the system.
3788 *Matthias St. Pierre*
3790 * Updated 'openssl rehash' to use OpenSSL consistent default.
3794 * Moved the load of the ssl_conf module to libcrypto, which helps
3795 loading engines that libssl uses before libssl is initialised.
3799 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3803 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3805 *Ingo Schwarze, Rich Salz*
3807 * Added output of accepting IP address and port for 'openssl s_server'
3811 * Added a new API for TLSv1.3 ciphersuites:
3812 SSL_CTX_set_ciphersuites()
3813 SSL_set_ciphersuites()
3817 * Memory allocation failures consistently add an error to the error
3822 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3823 in libcrypto when run as setuid/setgid.
3827 * Load any config file by default when libssl is used.
3831 * Added new public header file <openssl/rand_drbg.h> and documentation
3832 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3834 *Matthias St. Pierre*
3836 * QNX support removed (cannot find contributors to get their approval
3837 for the license change).
3841 * TLSv1.3 replay protection for early data has been implemented. See the
3842 SSL_read_early_data() man page for further details.
3846 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3847 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3848 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3849 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3850 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3851 configuration has been separated out. See the ciphers man page or the
3852 SSL_CTX_set_ciphersuites() man page for more information.
3856 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3857 in responder mode now supports the new "-multi" option, which
3858 spawns the specified number of child processes to handle OCSP
3859 requests. The "-timeout" option now also limits the OCSP
3860 responder's patience to wait to receive the full client request
3861 on a newly accepted connection. Child processes are respawned
3862 as needed, and the CA index file is automatically reloaded
3863 when changed. This makes it possible to run the "ocsp" responder
3864 as a long-running service, making the OpenSSL CA somewhat more
3865 feature-complete. In this mode, most diagnostic messages logged
3866 after entering the event loop are logged via syslog(3) rather than
3871 * Added support for X448 and Ed448. Heavily based on original work by
3876 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3877 objects loaded. This adds the functions OSSL_STORE_expect() and
3878 OSSL_STORE_find() as well as needed tools to construct searches and
3879 get the search data out of them.
3883 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3884 version of OpenSSL should review their configuration settings to ensure
3885 that they are still appropriate for TLSv1.3. For further information see:
3886 <https://wiki.openssl.org/index.php/TLS1.3>
3890 * Grand redesign of the OpenSSL random generator
3892 The default RAND method now utilizes an AES-CTR DRBG according to
3893 NIST standard SP 800-90Ar1. The new random generator is essentially
3894 a port of the default random generator from the OpenSSL FIPS 2.0
3895 object module. It is a hybrid deterministic random bit generator
3896 using an AES-CTR bit stream and which seeds and reseeds itself
3897 automatically using trusted system entropy sources.
3899 Some of its new features are:
3900 - Support for multiple DRBG instances with seed chaining.
3901 - The default RAND method makes use of a DRBG.
3902 - There is a public and private DRBG instance.
3903 - The DRBG instances are fork-safe.
3904 - Keep all global DRBG instances on the secure heap if it is enabled.
3905 - The public and private DRBG instance are per thread for lock free
3908 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3910 * Changed Configure so it only says what it does and doesn't dump
3911 so much data. Instead, ./configdata.pm should be used as a script
3912 to display all sorts of configuration data.
3916 * Added processing of "make variables" to Configure.
3920 * Added SHA512/224 and SHA512/256 algorithm support.
3924 * The last traces of Netware support, first removed in 1.1.0, have
3929 * Get rid of Makefile.shared, and in the process, make the processing
3930 of certain files (rc.obj, or the .def/.map/.opt files produced from
3931 the ordinal files) more visible and hopefully easier to trace and
3932 debug (or make silent).
3936 * Make it possible to have environment variable assignments as
3937 arguments to config / Configure.
3941 * Add multi-prime RSA (RFC 8017) support.
3945 * Add SM3 implemented according to GB/T 32905-2016
3946 *Jack Lloyd <jack.lloyd@ribose.com>,*
3947 *Ronald Tse <ronald.tse@ribose.com>,*
3948 *Erick Borsboom <erick.borsboom@ribose.com>*
3950 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3951 as documented in RFC6066.
3952 Based on a patch from Tomasz Moń
3954 *Filipe Raimundo da Silva*
3956 * Add SM4 implemented according to GB/T 32907-2016.
3957 *Jack Lloyd <jack.lloyd@ribose.com>,*
3958 *Ronald Tse <ronald.tse@ribose.com>,*
3959 *Erick Borsboom <erick.borsboom@ribose.com>*
3961 * Reimplement -newreq-nodes and ERR_error_string_n; the
3962 original author does not agree with the license change.
3966 * Add ARIA AEAD TLS support.
3970 * Some macro definitions to support VS6 have been removed. Visual
3971 Studio 6 has not worked since 1.1.0
3975 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3976 without clearing the errors.
3980 * Add "atfork" functions. If building on a system that without
3981 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3982 requirements. The RAND facility now uses/requires this.
3990 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3991 not possible to disable entirely. However, it's still possible to
3992 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3995 To disable, configure with 'no-ui-console'. 'no-ui' is still
3996 possible to use as an alias. Check at compile time with the
3997 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3998 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
4002 * Add a STORE module, which implements a uniform and URI based reader of
4003 stores that can contain keys, certificates, CRLs and numerous other
4004 objects. The main API is loosely based on a few stdio functions,
4005 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
4006 OSSL_STORE_error and OSSL_STORE_close.
4007 The implementation uses backends called "loaders" to implement arbitrary
4008 URI schemes. There is one built in "loader" for the 'file' scheme.
4012 * Add devcrypto engine. This has been implemented against cryptodev-linux,
4013 then adjusted to work on FreeBSD 8.4 as well.
4014 Enable by configuring with 'enable-devcryptoeng'. This is done by default
4015 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
4019 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
4020 util/mkerr.pl, which is adapted to allow those prefixes, leading to
4021 error code calls like this:
4023 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
4025 With this change, we claim the namespaces OSSL and OPENSSL in a manner
4026 that can be encoded in C. For the foreseeable future, this will only
4029 *Richard Levitte and Tim Hudson*
4031 * Removed BSD cryptodev engine.
4035 * Add a build target 'build_all_generated', to build all generated files
4036 and only that. This can be used to prepare everything that requires
4037 things like perl for a system that lacks perl and then move everything
4038 to that system and do the rest of the build there.
4042 * In the UI interface, make it possible to duplicate the user data. This
4043 can be used by engines that need to retain the data for a longer time
4044 than just the call where this user data is passed.
4048 * Ignore the '-named_curve auto' value for compatibility of applications
4051 *Tomáš Mráz <tmraz@fedoraproject.org>*
4053 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
4054 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
4055 alerts across multiple records (some of which could be empty). In practice
4056 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
4057 prohibits this altogether and other libraries (BoringSSL, NSS) do not
4058 support this at all. Supporting it adds significant complexity to the
4059 record layer, and its removal is unlikely to cause interoperability
4064 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
4065 with Z. These are meant to replace LONG and ZLONG and to be size safe.
4066 The use of LONG and ZLONG is discouraged and scheduled for deprecation
4071 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
4072 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
4074 *Richard Levitte, Andy Polyakov*
4076 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
4081 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4082 platform rather than 'mingw'.
4086 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
4087 success if they are asked to add an object which already exists
4088 in the store. This change cascades to other functions which load
4089 certificates and CRLs.
4093 * x86_64 assembly pack: annotate code with DWARF CFI directives to
4094 facilitate stack unwinding even from assembly subroutines.
4098 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
4099 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
4103 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4104 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4105 which is the minimum version we support.
4109 * Certificate time validation (X509_cmp_time) enforces stricter
4110 compliance with RFC 5280. Fractional seconds and timezone offsets
4111 are no longer allowed.
4115 * Add support for ARIA
4119 * s_client will now send the Server Name Indication (SNI) extension by
4120 default unless the new "-noservername" option is used. The server name is
4121 based on the host provided to the "-connect" option unless overridden by
4122 using "-servername".
4126 * Add support for SipHash
4130 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4131 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4132 prevent issues where no progress is being made and the peer continually
4133 sends unrecognised record types, using up resources processing them.
4137 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4138 using the algorithm defined in
4139 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4143 * Heartbeat support has been removed; the ABI is changed for now.
4145 *Richard Levitte, Rich Salz*
4147 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4151 * The RSA "null" method, which was partially supported to avoid patent
4152 issues, has been replaced to always returns NULL.
4159 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4161 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4162 used even when parsing explicit parameters, when loading a encoded key
4163 or calling `EC_GROUP_new_from_ecpkparameters()`/
4164 `EC_GROUP_new_from_ecparameters()`.
4165 This prevents bypass of security hardening and performance gains,
4166 especially for curves with specialized EC_METHODs.
4167 By default, if a key encoded with explicit parameters is loaded and later
4168 encoded, the output is still encoded with explicit parameters, even if
4169 internally a "named" EC_GROUP is used for computation.
4173 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4174 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4175 NULL. After this change, only the cofactor parameter can be NULL. It also
4176 does some minimal sanity checks on the passed order.
4181 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4182 An attack is simple, if the first CMS_recipientInfo is valid but the
4183 second CMS_recipientInfo is chosen ciphertext. If the second
4184 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4185 encryption key will be replaced by garbage, and the message cannot be
4186 decoded, but if the RSA decryption fails, the correct encryption key is
4187 used and the recipient will not notice the attack.
4188 As a work around for this potential attack the length of the decrypted
4189 key must be equal to the cipher default key length, in case the
4190 certificate is not given and all recipientInfo are tried out.
4191 The old behaviour can be re-enabled in the CMS code by setting the
4192 CMS_DEBUG_DECRYPT flag.
4197 * Use Windows installation paths in the mingw builds
4199 Mingw isn't a POSIX environment per se, which means that Windows
4200 paths should be used for installation.
4205 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4207 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4208 This changes the size when using the `genpkey` command when no size is given.
4209 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4210 generation commands to use 2048 bits by default.
4214 * Prevent over long nonces in ChaCha20-Poly1305.
4216 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4217 for every encryption operation. RFC 7539 specifies that the nonce value
4218 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4219 and front pads the nonce with 0 bytes if it is less than 12
4220 bytes. However it also incorrectly allows a nonce to be set of up to 16
4221 bytes. In this case only the last 12 bytes are significant and any
4222 additional leading bytes are ignored.
4224 It is a requirement of using this cipher that nonce values are
4225 unique. Messages encrypted using a reused nonce value are susceptible to
4226 serious confidentiality and integrity attacks. If an application changes
4227 the default nonce length to be longer than 12 bytes and then makes a
4228 change to the leading bytes of the nonce expecting the new value to be a
4229 new unique nonce then such an application could inadvertently encrypt
4230 messages with a reused nonce.
4232 Additionally the ignored bytes in a long nonce are not covered by the
4233 integrity guarantee of this cipher. Any application that relies on the
4234 integrity of these ignored leading bytes of a long nonce may be further
4235 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4236 is safe because no such use sets such a long nonce value. However user
4237 applications that use this cipher directly and set a non-default nonce
4238 length to be longer than 12 bytes may be vulnerable.
4240 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4246 * Added SCA hardening for modular field inversion in EC_GROUP through
4247 a new dedicated field_inv() pointer in EC_METHOD.
4248 This also addresses a leakage affecting conversions from projective
4249 to affine coordinates.
4251 *Billy Bob Brumley, Nicola Tuveri*
4253 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4254 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4258 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4262 * Remove the 'dist' target and add a tarball building script. The
4263 'dist' target has fallen out of use, and it shouldn't be
4264 necessary to configure just to create a source distribution.
4268 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4270 * Timing vulnerability in DSA signature generation
4272 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4273 timing side channel attack. An attacker could use variations in the signing
4274 algorithm to recover the private key.
4276 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4281 * Timing vulnerability in ECDSA signature generation
4283 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4284 timing side channel attack. An attacker could use variations in the signing
4285 algorithm to recover the private key.
4287 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4292 * Add coordinate blinding for EC_POINT and implement projective
4293 coordinate blinding for generic prime curves as a countermeasure to
4294 chosen point SCA attacks.
4296 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4298 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4300 * Client DoS due to large DH parameter
4302 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4303 malicious server can send a very large prime value to the client. This will
4304 cause the client to spend an unreasonably long period of time generating a
4305 key for this prime resulting in a hang until the client has finished. This
4306 could be exploited in a Denial Of Service attack.
4308 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4313 * Cache timing vulnerability in RSA Key Generation
4315 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4316 a cache timing side channel attack. An attacker with sufficient access to
4317 mount cache timing attacks during the RSA key generation process could
4318 recover the private key.
4320 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4321 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4326 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4327 parameter is no longer accepted, as it leads to a corrupt table. NULL
4328 pem_str is reserved for alias entries only.
4332 * Revert blinding in ECDSA sign and instead make problematic addition
4333 length-invariant. Switch even to fixed-length Montgomery multiplication.
4337 * Change generating and checking of primes so that the error rate of not
4338 being prime depends on the intended use based on the size of the input.
4339 For larger primes this will result in more rounds of Miller-Rabin.
4340 The maximal error rate for primes with more than 1080 bits is lowered
4343 *Kurt Roeckx, Annie Yousar*
4345 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4349 * Add blinding to ECDSA and DSA signatures to protect against side channel
4350 attacks discovered by Keegan Ryan (NCC Group).
4354 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4355 now allow empty (zero character) pass phrases.
4359 * Certificate time validation (X509_cmp_time) enforces stricter
4360 compliance with RFC 5280. Fractional seconds and timezone offsets
4361 are no longer allowed.
4365 * Fixed a text canonicalisation bug in CMS
4367 Where a CMS detached signature is used with text content the text goes
4368 through a canonicalisation process first prior to signing or verifying a
4369 signature. This process strips trailing space at the end of lines, converts
4370 line terminators to CRLF and removes additional trailing line terminators
4371 at the end of a file. A bug in the canonicalisation process meant that
4372 some characters, such as form-feed, were incorrectly treated as whitespace
4373 and removed. This is contrary to the specification (RFC5485). This fix
4374 could mean that detached text data signed with an earlier version of
4375 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4376 signed with a fixed OpenSSL may fail to verify with an earlier version of
4377 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4378 and use the "-binary" flag (for the "cms" command line application) or set
4379 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4383 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4385 * Constructed ASN.1 types with a recursive definition could exceed the stack
4387 Constructed ASN.1 types with a recursive definition (such as can be found
4388 in PKCS7) could eventually exceed the stack given malicious input with
4389 excessive recursion. This could result in a Denial Of Service attack. There
4390 are no such structures used within SSL/TLS that come from untrusted sources
4391 so this is considered safe.
4393 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4399 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4401 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4402 effectively reduced to only comparing the least significant bit of each
4403 byte. This allows an attacker to forge messages that would be considered as
4404 authenticated in an amount of tries lower than that guaranteed by the
4405 security claims of the scheme. The module can only be compiled by the
4406 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4408 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4414 * Add a build target 'build_all_generated', to build all generated files
4415 and only that. This can be used to prepare everything that requires
4416 things like perl for a system that lacks perl and then move everything
4417 to that system and do the rest of the build there.
4421 * Backport SSL_OP_NO_RENGOTIATION
4423 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4424 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4425 changes this is no longer possible in 1.1.0. Therefore, the new
4426 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4427 1.1.0 to provide equivalent functionality.
4429 Note that if an application built against 1.1.0h headers (or above) is run
4430 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4431 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4435 * Removed the OS390-Unix config target. It relied on a script that doesn't
4440 * rsaz_1024_mul_avx2 overflow bug on x86_64
4442 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4443 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4444 Analysis suggests that attacks against RSA and DSA as a result of this
4445 defect would be very difficult to perform and are not believed likely.
4446 Attacks against DH1024 are considered just feasible, because most of the
4447 work necessary to deduce information about a private key may be performed
4448 offline. The amount of resources required for such an attack would be
4449 significant. However, for an attack on TLS to be meaningful, the server
4450 would have to share the DH1024 private key among multiple clients, which is
4451 no longer an option since CVE-2016-0701.
4453 This only affects processors that support the AVX2 but not ADX extensions
4454 like Intel Haswell (4th generation).
4456 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4457 was originally found via the OSS-Fuzz project.
4462 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4464 * bn_sqrx8x_internal carry bug on x86_64
4466 There is a carry propagating bug in the x86_64 Montgomery squaring
4467 procedure. No EC algorithms are affected. Analysis suggests that attacks
4468 against RSA and DSA as a result of this defect would be very difficult to
4469 perform and are not believed likely. Attacks against DH are considered just
4470 feasible (although very difficult) because most of the work necessary to
4471 deduce information about a private key may be performed offline. The amount
4472 of resources required for such an attack would be very significant and
4473 likely only accessible to a limited number of attackers. An attacker would
4474 additionally need online access to an unpatched system using the target
4475 private key in a scenario with persistent DH parameters and a private
4476 key that is shared between multiple clients.
4478 This only affects processors that support the BMI1, BMI2 and ADX extensions
4479 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4481 This issue was reported to OpenSSL by the OSS-Fuzz project.
4486 * Malformed X.509 IPAddressFamily could cause OOB read
4488 If an X.509 certificate has a malformed IPAddressFamily extension,
4489 OpenSSL could do a one-byte buffer overread. The most likely result
4490 would be an erroneous display of the certificate in text format.
4492 This issue was reported to OpenSSL by the OSS-Fuzz project.
4497 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4499 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4500 platform rather than 'mingw'.
4504 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4505 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4506 which is the minimum version we support.
4510 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4512 * Encrypt-Then-Mac renegotiation crash
4514 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4515 negotiated where it was not in the original handshake (or vice-versa) then
4516 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4517 and servers are affected.
4519 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4524 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4526 * Truncated packet could crash via OOB read
4528 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4529 cipher is being used, then a truncated packet can cause that host to
4530 perform an out-of-bounds read, usually resulting in a crash.
4532 This issue was reported to OpenSSL by Robert Święcki of Google.
4537 * Bad (EC)DHE parameters cause a client crash
4539 If a malicious server supplies bad parameters for a DHE or ECDHE key
4540 exchange then this can result in the client attempting to dereference a
4541 NULL pointer leading to a client crash. This could be exploited in a Denial
4544 This issue was reported to OpenSSL by Guido Vranken.
4549 * BN_mod_exp may produce incorrect results on x86_64
4551 There is a carry propagating bug in the x86_64 Montgomery squaring
4552 procedure. No EC algorithms are affected. Analysis suggests that attacks
4553 against RSA and DSA as a result of this defect would be very difficult to
4554 perform and are not believed likely. Attacks against DH are considered just
4555 feasible (although very difficult) because most of the work necessary to
4556 deduce information about a private key may be performed offline. The amount
4557 of resources required for such an attack would be very significant and
4558 likely only accessible to a limited number of attackers. An attacker would
4559 additionally need online access to an unpatched system using the target
4560 private key in a scenario with persistent DH parameters and a private
4561 key that is shared between multiple clients. For example this can occur by
4562 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4563 similar to CVE-2015-3193 but must be treated as a separate problem.
4565 This issue was reported to OpenSSL by the OSS-Fuzz project.
4570 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4572 * ChaCha20/Poly1305 heap-buffer-overflow
4574 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4575 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4576 crash. This issue is not considered to be exploitable beyond a DoS.
4578 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4583 * CMS Null dereference
4585 Applications parsing invalid CMS structures can crash with a NULL pointer
4586 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4587 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4588 structure callback if an attempt is made to free certain invalid encodings.
4589 Only CHOICE structures using a callback which do not handle NULL value are
4592 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4597 * Montgomery multiplication may produce incorrect results
4599 There is a carry propagating bug in the Broadwell-specific Montgomery
4600 multiplication procedure that handles input lengths divisible by, but
4601 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4602 and DH private keys are impossible. This is because the subroutine in
4603 question is not used in operations with the private key itself and an input
4604 of the attacker's direct choice. Otherwise the bug can manifest itself as
4605 transient authentication and key negotiation failures or reproducible
4606 erroneous outcome of public-key operations with specially crafted input.
4607 Among EC algorithms only Brainpool P-512 curves are affected and one
4608 presumably can attack ECDH key negotiation. Impact was not analyzed in
4609 detail, because pre-requisites for attack are considered unlikely. Namely
4610 multiple clients have to choose the curve in question and the server has to
4611 share the private key among them, neither of which is default behaviour.
4612 Even then only clients that chose the curve will be affected.
4614 This issue was publicly reported as transient failures and was not
4615 initially recognized as a security issue. Thanks to Richard Morgan for
4616 providing reproducible case.
4621 * Removed automatic addition of RPATH in shared libraries and executables,
4622 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4626 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4628 * Fix Use After Free for large message sizes
4630 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4631 message larger than approx 16k is received then the underlying buffer to
4632 store the incoming message is reallocated and moved. Unfortunately a
4633 dangling pointer to the old location is left which results in an attempt to
4634 write to the previously freed location. This is likely to result in a
4635 crash, however it could potentially lead to execution of arbitrary code.
4637 This issue only affects OpenSSL 1.1.0a.
4639 This issue was reported to OpenSSL by Robert Święcki.
4644 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4646 * OCSP Status Request extension unbounded memory growth
4648 A malicious client can send an excessively large OCSP Status Request
4649 extension. If that client continually requests renegotiation, sending a
4650 large OCSP Status Request extension each time, then there will be unbounded
4651 memory growth on the server. This will eventually lead to a Denial Of
4652 Service attack through memory exhaustion. Servers with a default
4653 configuration are vulnerable even if they do not support OCSP. Builds using
4654 the "no-ocsp" build time option are not affected.
4656 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4661 * SSL_peek() hang on empty record
4663 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4664 sends an empty record. This could be exploited by a malicious peer in a
4665 Denial Of Service attack.
4667 This issue was reported to OpenSSL by Alex Gaynor.
4672 * Excessive allocation of memory in tls_get_message_header() and
4673 dtls1_preprocess_fragment()
4675 A (D)TLS message includes 3 bytes for its length in the header for the
4676 message. This would allow for messages up to 16Mb in length. Messages of
4677 this length are excessive and OpenSSL includes a check to ensure that a
4678 peer is sending reasonably sized messages in order to avoid too much memory
4679 being consumed to service a connection. A flaw in the logic of version
4680 1.1.0 means that memory for the message is allocated too early, prior to
4681 the excessive message length check. Due to way memory is allocated in
4682 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4683 to service a connection. This could lead to a Denial of Service through
4684 memory exhaustion. However, the excessive message length check still takes
4685 place, and this would cause the connection to immediately fail. Assuming
4686 that the application calls SSL_free() on the failed connection in a timely
4687 manner then the 21Mb of allocated memory will then be immediately freed
4688 again. Therefore, the excessive memory allocation will be transitory in
4689 nature. This then means that there is only a security impact if:
4691 1) The application does not call SSL_free() in a timely manner in the event
4692 that the connection fails
4694 2) The application is working in a constrained environment where there is
4695 very little free memory
4697 3) The attacker initiates multiple connection attempts such that there are
4698 multiple connections in a state where memory has been allocated for the
4699 connection; SSL_free() has not yet been called; and there is insufficient
4700 memory to service the multiple requests.
4702 Except in the instance of (1) above any Denial Of Service is likely to be
4703 transitory because as soon as the connection fails the memory is
4704 subsequently freed again in the SSL_free() call. However there is an
4705 increased risk during this period of application crashes due to the lack of
4706 memory - which would then mean a more serious Denial of Service.
4708 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4709 (CVE-2016-6307 and CVE-2016-6308)
4713 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4714 had to be removed. Primary reason is that vendor assembler can't
4715 assemble our modules with -KPIC flag. As result it, assembly
4716 support, was not even available as option. But its lack means
4717 lack of side-channel resistant code, which is incompatible with
4718 security by todays standards. Fortunately gcc is readily available
4719 prepackaged option, which we firmly point at...
4723 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4725 * Windows command-line tool supports UTF-8 opt-in option for arguments
4726 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4727 (to any value) allows Windows user to access PKCS#12 file generated
4728 with Windows CryptoAPI and protected with non-ASCII password, as well
4729 as files generated under UTF-8 locale on Linux also protected with
4734 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4735 have been disabled by default and removed from DEFAULT, just like RC4.
4736 See the RC4 item below to re-enable both.
4740 * The method for finding the storage location for the Windows RAND seed file
4741 has changed. First we check %RANDFILE%. If that is not set then we check
4742 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4743 all else fails we fall back to C:\.
4747 * The EVP_EncryptUpdate() function has had its return type changed from void
4748 to int. A return of 0 indicates and error while a return of 1 indicates
4753 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4754 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4755 off the constant time implementation for RSA, DSA and DH have been made
4756 no-ops and deprecated.
4760 * Windows RAND implementation was simplified to only get entropy by
4761 calling CryptGenRandom(). Various other RAND-related tickets
4764 *Joseph Wylie Yandle, Rich Salz*
4766 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4767 and `OPENSSL_LH_`, respectively. The old names are available
4768 with API compatibility. They new names are now completely documented.
4772 * Unify TYPE_up_ref(obj) methods signature.
4773 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4774 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4775 int (instead of void) like all others TYPE_up_ref() methods.
4776 So now these methods also check the return value of CRYPTO_atomic_add(),
4777 and the validity of object reference counter.
4779 *fdasilvayy@gmail.com*
4781 * With Windows Visual Studio builds, the .pdb files are installed
4782 alongside the installed libraries and executables. For a static
4783 library installation, ossl_static.pdb is the associate compiler
4784 generated .pdb file to be used when linking programs.
4788 * Remove openssl.spec. Packaging files belong with the packagers.
4792 * Automatic Darwin/OSX configuration has had a refresh, it will now
4793 recognise x86_64 architectures automatically. You can still decide
4794 to build for a different bitness with the environment variable
4795 KERNEL_BITS (can be 32 or 64), for example:
4797 KERNEL_BITS=32 ./config
4801 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4802 256 bit AES and HMAC with SHA256.
4806 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4810 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4814 * To enable users to have their own config files and build file templates,
4815 Configure looks in the directory indicated by the environment variable
4816 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4817 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4818 name and is used as is.
4822 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4823 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4824 X509_CERT_FILE_CTX was removed.
4828 * "shared" builds are now the default. To create only static libraries use
4829 the "no-shared" Configure option.
4833 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4834 All of these option have not worked for some while and are fundamental
4839 * Make various cleanup routines no-ops and mark them as deprecated. Most
4840 global cleanup functions are no longer required because they are handled
4841 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4842 Explicitly de-initing can cause problems (e.g. where a library that uses
4843 OpenSSL de-inits, but an application is still using it). The affected
4844 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4845 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4846 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4847 COMP_zlib_cleanup().
4851 * --strict-warnings no longer enables runtime debugging options
4852 such as REF_DEBUG. Instead, debug options are automatically
4853 enabled with '--debug' builds.
4855 *Andy Polyakov, Emilia Käsper*
4857 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4858 have been moved out of the public header files. New functions for managing
4859 these have been added.
4863 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4864 objects have been moved out of the public header files. New
4865 functions for managing these have been added.
4869 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4870 have been moved out of the public header files. New functions for managing
4871 these have been added.
4875 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4876 moved out of the public header files. New functions for managing these
4881 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4885 * Removed the mk1mf build scripts.
4889 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4890 it is always safe to #include a header now.
4894 * Removed the aged BC-32 config and all its supporting scripts
4898 * Removed support for Ultrix, Netware, and OS/2.
4902 * Add support for HKDF.
4904 *Alessandro Ghedini*
4906 * Add support for blake2b and blake2s
4910 * Added support for "pipelining". Ciphers that have the
4911 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4912 encryptions/decryptions simultaneously. There are currently no built-in
4913 ciphers with this property but the expectation is that engines will be able
4914 to offer it to significantly improve throughput. Support has been extended
4915 into libssl so that multiple records for a single connection can be
4916 processed in one go (for >=TLS 1.1).
4920 * Added the AFALG engine. This is an async capable engine which is able to
4921 offload work to the Linux kernel. In this initial version it only supports
4922 AES128-CBC. The kernel must be version 4.1.0 or greater.
4926 * OpenSSL now uses a new threading API. It is no longer necessary to
4927 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4928 are two supported threading models: pthreads and windows threads. It is
4929 also possible to configure OpenSSL at compile time for "no-threads". The
4930 old threading API should no longer be used. The functions have been
4931 replaced with "no-op" compatibility macros.
4933 *Alessandro Ghedini, Matt Caswell*
4935 * Modify behavior of ALPN to invoke callback after SNI/servername
4936 callback, such that updates to the SSL_CTX affect ALPN.
4940 * Add SSL_CIPHER queries for authentication and key-exchange.
4944 * Changes to the DEFAULT cipherlist:
4945 - Prefer (EC)DHE handshakes over plain RSA.
4946 - Prefer AEAD ciphers over legacy ciphers.
4947 - Prefer ECDSA over RSA when both certificates are available.
4948 - Prefer TLSv1.2 ciphers/PRF.
4949 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4954 * Change the ECC default curve list to be this, in order: x25519,
4955 secp256r1, secp521r1, secp384r1.
4959 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4960 disabled by default. They can be re-enabled using the
4961 enable-weak-ssl-ciphers option to Configure.
4965 * If the server has ALPN configured, but supports no protocols that the
4966 client advertises, send a fatal "no_application_protocol" alert.
4967 This behaviour is SHALL in RFC 7301, though it isn't universally
4968 implemented by other servers.
4972 * Add X25519 support.
4973 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4974 for public and private key encoding using the format documented in
4975 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4976 key generation and key derivation.
4978 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4983 * Deprecate SRP_VBASE_get_by_user.
4984 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4985 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4986 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4987 seed, even if the seed is configured.
4989 Users should use SRP_VBASE_get1_by_user instead. Note that in
4990 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4991 also that even though configuring the SRP seed attempts to hide
4992 invalid usernames by continuing the handshake with fake
4993 credentials, this behaviour is not constant time and no strong
4994 guarantees are made that the handshake is indistinguishable from
4995 that of a valid user.
4999 * Configuration change; it's now possible to build dynamic engines
5000 without having to build shared libraries and vice versa. This
5001 only applies to the engines in `engines/`, those in `crypto/engine/`
5002 will always be built into libcrypto (i.e. "static").
5004 Building dynamic engines is enabled by default; to disable, use
5005 the configuration option "disable-dynamic-engine".
5007 The only requirements for building dynamic engines are the
5008 presence of the DSO module and building with position independent
5009 code, so they will also automatically be disabled if configuring
5010 with "disable-dso" or "disable-pic".
5012 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
5013 are also taken away from openssl/opensslconf.h, as they are
5018 * Configuration change; if there is a known flag to compile
5019 position independent code, it will always be applied on the
5020 libcrypto and libssl object files, and never on the application
5021 object files. This means other libraries that use routines from
5022 libcrypto / libssl can be made into shared libraries regardless
5023 of how OpenSSL was configured.
5025 If this isn't desirable, the configuration options "disable-pic"
5026 or "no-pic" can be used to disable the use of PIC. This will
5027 also disable building shared libraries and dynamic engines.
5031 * Removed JPAKE code. It was experimental and has no wide use.
5035 * The INSTALL_PREFIX Makefile variable has been renamed to
5036 DESTDIR. That makes for less confusion on what this variable
5037 is for. Also, the configuration option --install_prefix is
5042 * Heartbeat for TLS has been removed and is disabled by default
5043 for DTLS; configure with enable-heartbeats. Code that uses the
5044 old #define's might need to be updated.
5046 *Emilia Käsper, Rich Salz*
5048 * Rename REF_CHECK to REF_DEBUG.
5052 * New "unified" build system
5054 The "unified" build system is aimed to be a common system for all
5055 platforms we support. With it comes new support for VMS.
5057 This system builds supports building in a different directory tree
5058 than the source tree. It produces one Makefile (for unix family
5059 or lookalikes), or one descrip.mms (for VMS).
5061 The source of information to make the Makefile / descrip.mms is
5062 small files called 'build.info', holding the necessary
5063 information for each directory with source to compile, and a
5064 template in Configurations, like unix-Makefile.tmpl or
5067 With this change, the library names were also renamed on Windows
5068 and on VMS. They now have names that are closer to the standard
5069 on Unix, and include the major version number, and in certain
5070 cases, the architecture they are built for. See "Notes on shared
5071 libraries" in INSTALL.
5073 We rely heavily on the perl module Text::Template.
5077 * Added support for auto-initialisation and de-initialisation of the library.
5078 OpenSSL no longer requires explicit init or deinit routines to be called,
5079 except in certain circumstances. See the OPENSSL_init_crypto() and
5080 OPENSSL_init_ssl() man pages for further information.
5084 * The arguments to the DTLSv1_listen function have changed. Specifically the
5085 "peer" argument is now expected to be a BIO_ADDR object.
5087 * Rewrite of BIO networking library. The BIO library lacked consistent
5088 support of IPv6, and adding it required some more extensive
5089 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
5090 which hold all types of addresses and chains of address information.
5091 It also introduces a new API, with functions like BIO_socket,
5092 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
5093 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
5094 have been adapted accordingly.
5098 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
5103 * CRIME protection: disable compression by default, even if OpenSSL is
5104 compiled with zlib enabled. Applications can still enable compression
5105 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
5106 using the SSL_CONF library to configure compression.
5110 * The signature of the session callback configured with
5111 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
5112 was explicitly marked as `const unsigned char*` instead of
5117 * Always DPURIFY. Remove the use of uninitialized memory in the
5118 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
5122 * Removed many obsolete configuration items, including
5123 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
5124 MD2_CHAR, MD2_INT, MD2_LONG
5126 IDEA_SHORT, IDEA_LONG
5127 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5129 *Rich Salz, with advice from Andy Polyakov*
5131 * Many BN internals have been moved to an internal header file.
5133 *Rich Salz with help from Andy Polyakov*
5135 * Configuration and writing out the results from it has changed.
5136 Files such as Makefile include/openssl/opensslconf.h and are now
5137 produced through general templates, such as Makefile.in and
5138 crypto/opensslconf.h.in and some help from the perl module
5141 Also, the center of configuration information is no longer
5142 Makefile. Instead, Configure produces a perl module in
5143 configdata.pm which holds most of the config data (in the hash
5144 table %config), the target data that comes from the target
5145 configuration in one of the `Configurations/*.conf` files (in
5150 * To clarify their intended purposes, the Configure options
5151 --prefix and --openssldir change their semantics, and become more
5152 straightforward and less interdependent.
5154 --prefix shall be used exclusively to give the location INSTALLTOP
5155 where programs, scripts, libraries, include files and manuals are
5156 going to be installed. The default is now /usr/local.
5158 --openssldir shall be used exclusively to give the default
5159 location OPENSSLDIR where certificates, private keys, CRLs are
5160 managed. This is also where the default openssl.cnf gets
5162 If the directory given with this option is a relative path, the
5163 values of both the --prefix value and the --openssldir value will
5164 be combined to become OPENSSLDIR.
5165 The default for --openssldir is INSTALLTOP/ssl.
5167 Anyone who uses --openssldir to specify where OpenSSL is to be
5168 installed MUST change to use --prefix instead.
5172 * The GOST engine was out of date and therefore it has been removed. An up
5173 to date GOST engine is now being maintained in an external repository.
5174 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5175 support for GOST ciphersuites (these are only activated if a GOST engine
5180 * EGD is no longer supported by default; use enable-egd when
5183 *Ben Kaduk and Rich Salz*
5185 * The distribution now has Makefile.in files, which are used to
5186 create Makefile's when Configure is run. *Configure must be run
5187 before trying to build now.*
5191 * The return value for SSL_CIPHER_description() for error conditions
5196 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5198 Obtaining and performing DNSSEC validation of TLSA records is
5199 the application's responsibility. The application provides
5200 the TLSA records of its choice to OpenSSL, and these are then
5201 used to authenticate the peer.
5203 The TLSA records need not even come from DNS. They can, for
5204 example, be used to implement local end-entity certificate or
5205 trust-anchor "pinning", where the "pin" data takes the form
5206 of TLSA records, which can augment or replace verification
5207 based on the usual WebPKI public certification authorities.
5211 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5212 continues to support deprecated interfaces in default builds.
5213 However, applications are strongly advised to compile their
5214 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5215 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5216 or the 1.1.0 releases.
5218 In environments in which all applications have been ported to
5219 not use any deprecated interfaces OpenSSL's Configure script
5220 should be used with the --api=1.1.0 option to entirely remove
5221 support for the deprecated features from the library and
5222 unconditionally disable them in the installed headers.
5223 Essentially the same effect can be achieved with the "no-deprecated"
5224 argument to Configure, except that this will always restrict
5225 the build to just the latest API, rather than a fixed API
5228 As applications are ported to future revisions of the API,
5229 they should update their compile-time OPENSSL_API_COMPAT define
5230 accordingly, but in most cases should be able to continue to
5231 compile with later releases.
5233 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5234 0x10000000L and 0x00908000L, respectively. However those
5235 versions did not support the OPENSSL_API_COMPAT feature, and
5236 so applications are not typically tested for explicit support
5237 of just the undeprecated features of either release.
5241 * Add support for setting the minimum and maximum supported protocol.
5242 It can bet set via the SSL_set_min_proto_version() and
5243 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5244 MaxProtocol. It's recommended to use the new APIs to disable
5245 protocols instead of disabling individual protocols using
5246 SSL_set_options() or SSL_CONF's Protocol. This change also
5247 removes support for disabling TLS 1.2 in the OpenSSL TLS
5248 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5252 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5256 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5257 and integrates ECDSA and ECDH functionality into EC. Implementations can
5258 now redirect key generation and no longer need to convert to or from
5261 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5262 include the ec.h header file instead.
5266 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5267 ciphers who are no longer supported and drops support the ephemeral RSA key
5268 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5272 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5273 opaque. For HMAC_CTX, the following constructors and destructors
5276 HMAC_CTX *HMAC_CTX_new(void);
5277 void HMAC_CTX_free(HMAC_CTX *ctx);
5279 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5280 destroy such methods has been added. See EVP_MD_meth_new(3) and
5281 EVP_CIPHER_meth_new(3) for documentation.
5284 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5285 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5286 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5287 an already created structure.
5288 2) For consistency with the majority of our object creators and
5289 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5290 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5291 for deprecated builds.
5295 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5296 cryptographic operations to be performed asynchronously as long as an
5297 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5298 further details. Libssl has also had this capability integrated with the
5299 introduction of the new mode SSL_MODE_ASYNC and associated error
5300 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5301 pages. This work was developed in partnership with Intel Corp.
5305 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5306 always enabled now. If you want to disable the support you should
5307 exclude it using the list of supported ciphers. This also means that the
5308 "-no_ecdhe" option has been removed from s_server.
5312 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5313 SSL_{CTX_}set1_curves() which can set a list.
5317 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5318 curve you want to support using SSL_{CTX_}set1_curves().
5322 * State machine rewrite. The state machine code has been significantly
5323 refactored in order to remove much duplication of code and solve issues
5324 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5325 further details). This change does have some associated API changes.
5326 Notably the SSL_state() function has been removed and replaced by
5327 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5328 SSL_set_state() has been removed altogether. The previous handshake states
5329 defined in ssl.h and ssl3.h have also been removed.
5333 * All instances of the string "ssleay" in the public API were replaced
5334 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5335 Some error codes related to internal RSA_eay API's were renamed.
5339 * The demo files in crypto/threads were moved to demo/threads.
5343 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5346 *Matt Caswell, Rich Salz*
5348 * New ASN.1 embed macro.
5350 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5351 structure is not allocated: it is part of the parent. That is instead of
5359 This reduces memory fragmentation and make it impossible to accidentally
5360 set a mandatory field to NULL.
5362 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5363 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5364 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5369 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5373 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5374 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5375 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5376 DES and RC4 ciphersuites.
5380 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5381 This changes the decoding behaviour for some invalid messages,
5382 though the change is mostly in the more lenient direction, and
5383 legacy behaviour is preserved as much as possible.
5387 * Fix no-stdio build.
5388 *David Woodhouse <David.Woodhouse@intel.com> and also*
5389 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5391 * New testing framework
5392 The testing framework has been largely rewritten and is now using
5393 perl and the perl modules Test::Harness and an extended variant of
5394 Test::More called OpenSSL::Test to do its work. All test scripts in
5395 test/ have been rewritten into test recipes, and all direct calls to
5396 executables in test/Makefile have become individual recipes using the
5397 simplified testing OpenSSL::Test::Simple.
5399 For documentation on our testing modules, do:
5401 perldoc test/testlib/OpenSSL/Test/Simple.pm
5402 perldoc test/testlib/OpenSSL/Test.pm
5406 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5407 are used; the latter aborts on memory leaks (usually checked on exit).
5408 Some undocumented "set malloc, etc., hooks" functions were removed
5409 and others were changed. All are now documented.
5413 * In DSA_generate_parameters_ex, if the provided seed is too short,
5416 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5418 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5419 from RFC4279, RFC4785, RFC5487, RFC5489.
5421 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5422 original RSA_PSK patch.
5426 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5427 era flag was never set throughout the codebase (only read). Also removed
5428 SSL3_FLAGS_POP_BUFFER which was only used if
5429 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5433 * Changed the default name options in the "ca", "crl", "req" and "x509"
5434 to be "oneline" instead of "compat".
5438 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5439 not aware of clients that still exhibit this bug, and the workaround
5440 hasn't been working properly for a while.
5444 * The return type of BIO_number_read() and BIO_number_written() as well as
5445 the corresponding num_read and num_write members in the BIO structure has
5446 changed from unsigned long to uint64_t. On platforms where an unsigned
5447 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5452 * Given the pervasive nature of TLS extensions it is inadvisable to run
5453 OpenSSL without support for them. It also means that maintaining
5454 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5455 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5459 * Removed support for the two export grade static DH ciphersuites
5460 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5461 were newly added (along with a number of other static DH ciphersuites) to
5462 1.0.2. However the two export ones have *never* worked since they were
5463 introduced. It seems strange in any case to be adding new export
5464 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5468 * Version negotiation has been rewritten. In particular SSLv23_method(),
5469 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5470 and turned into macros which simply call the new preferred function names
5471 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5472 should use the new names instead. Also as part of this change the ssl23.h
5473 header file has been removed.
5477 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5478 code and the associated standard is no longer considered fit-for-purpose.
5482 * RT2547 was closed. When generating a private key, try to make the
5483 output file readable only by the owner. This behavior change might
5484 be noticeable when interacting with other software.
5486 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5491 * Added HTTP GET support to the ocsp command.
5495 * Changed default digest for the dgst and enc commands from MD5 to
5500 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5504 * Added support for TLS extended master secret from
5505 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5506 initial patch which was a great help during development.
5510 * All libssl internal structures have been removed from the public header
5511 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5512 now redundant). Users should not attempt to access internal structures
5513 directly. Instead they should use the provided API functions.
5517 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5518 Access to deprecated functions can be re-enabled by running config with
5519 "enable-deprecated". In addition applications wishing to use deprecated
5520 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5521 will, by default, disable some transitive includes that previously existed
5522 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5526 * Added support for OCB mode. OpenSSL has been granted a patent license
5527 compatible with the OpenSSL license for use of OCB. Details are available
5528 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5529 for OCB can be removed by calling config with no-ocb.
5533 * SSLv2 support has been removed. It still supports receiving an SSLv2
5534 compatible client hello.
5538 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5539 done while fixing the error code for the key-too-small case.
5541 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5543 * CA.sh has been removed; use CA.pl instead.
5547 * Removed old DES API.
5551 * Remove various unsupported platforms:
5557 Sinix/ReliantUNIX RM400
5562 16-bit platforms such as WIN16
5566 * Clean up OPENSSL_NO_xxx #define's
5567 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5568 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5569 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5570 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5571 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5572 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5573 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5574 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5575 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5576 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5580 * Cleaned up dead code
5581 Remove all but one '#ifdef undef' which is to be looked at.
5585 * Clean up calling of xxx_free routines.
5586 Just like free(), fix most of the xxx_free routines to accept
5587 NULL. Remove the non-null checks from callers. Save much code.
5591 * Add secure heap for storage of private keys (when possible).
5592 Add BIO_s_secmem(), CBIGNUM, etc.
5593 Contributed by Akamai Technologies under our Corporate CLA.
5597 * Experimental support for a new, fast, unbiased prime candidate generator,
5598 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5600 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5602 * New output format NSS in the sess_id command line tool. This allows
5603 exporting the session id and the master key in NSS keylog format.
5605 *Martin Kaiser <martin@kaiser.cx>*
5607 * Harmonize version and its documentation. -f flag is used to display
5610 *mancha <mancha1@zoho.com>*
5612 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5613 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5615 *mancha <mancha1@zoho.com>*
5617 * Fix some double frees. These are not thought to be exploitable.
5619 *mancha <mancha1@zoho.com>*
5621 * A missing bounds check in the handling of the TLS heartbeat extension
5622 can be used to reveal up to 64k of memory to a connected client or
5625 Thanks for Neel Mehta of Google Security for discovering this bug and to
5626 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5627 preparing the fix ([CVE-2014-0160])
5629 *Adam Langley, Bodo Moeller*
5631 * Fix for the attack described in the paper "Recovering OpenSSL
5632 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5633 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5634 <http://eprint.iacr.org/2014/140>
5636 Thanks to Yuval Yarom and Naomi Benger for discovering this
5637 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5639 *Yuval Yarom and Naomi Benger*
5641 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5642 this fixes a limitation in previous versions of OpenSSL.
5646 * Experimental encrypt-then-mac support.
5648 Experimental support for encrypt then mac from
5649 draft-gutmann-tls-encrypt-then-mac-02.txt
5651 To enable it set the appropriate extension number (0x42 for the test
5652 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5654 For non-compliant peers (i.e. just about everything) this should have no
5657 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5661 * Add EVP support for key wrapping algorithms, to avoid problems with
5662 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5663 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5664 algorithms and include tests cases.
5668 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5673 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5674 MGF1 digest and OAEP label.
5678 * Make openssl verify return errors.
5680 *Chris Palmer <palmer@google.com> and Ben Laurie*
5682 * New function ASN1_TIME_diff to calculate the difference between two
5683 ASN1_TIME structures or one structure and the current time.
5687 * Update fips_test_suite to support multiple command line options. New
5688 test to induce all self test errors in sequence and check expected
5693 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5694 sign or verify all in one operation.
5698 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5699 test programs and fips_test_suite. Includes functionality to parse
5700 the minimal script output of fipsalgest.pl directly.
5704 * Add authorisation parameter to FIPS_module_mode_set().
5708 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5712 * Use separate DRBG fields for internal and external flags. New function
5713 FIPS_drbg_health_check() to perform on demand health checking. Add
5714 generation tests to fips_test_suite with reduced health check interval to
5715 demonstrate periodic health checking. Add "nodh" option to
5716 fips_test_suite to skip very slow DH test.
5720 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5725 * More extensive health check for DRBG checking many more failure modes.
5726 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5727 combination: call this in fips_test_suite.
5731 * Add support for canonical generation of DSA parameter 'g'. See
5734 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5735 POST to handle HMAC cases.
5739 * Add functions FIPS_module_version() and FIPS_module_version_text()
5740 to return numerical and string versions of the FIPS module number.
5744 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5745 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5746 outside the validated module in the FIPS capable OpenSSL.
5750 * Minor change to DRBG entropy callback semantics. In some cases
5751 there is no multiple of the block length between min_len and
5752 max_len. Allow the callback to return more than max_len bytes
5753 of entropy but discard any extra: it is the callback's responsibility
5754 to ensure that the extra data discarded does not impact the
5755 requested amount of entropy.
5759 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5760 information in FIPS186-3, SP800-57 and SP800-131A.
5764 * CCM support via EVP. Interface is very similar to GCM case except we
5765 must supply all data in one chunk (i.e. no update, final) and the
5766 message length must be supplied if AAD is used. Add algorithm test
5771 * Initial version of POST overhaul. Add POST callback to allow the status
5772 of POST to be monitored and/or failures induced. Modify fips_test_suite
5773 to use callback. Always run all selftests even if one fails.
5777 * XTS support including algorithm test driver in the fips_gcmtest program.
5778 Note: this does increase the maximum key length from 32 to 64 bytes but
5779 there should be no binary compatibility issues as existing applications
5780 will never use XTS mode.
5784 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5785 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5786 performs algorithm blocking for unapproved PRNG types. Also do not
5787 set PRNG type in FIPS_mode_set(): leave this to the application.
5788 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5789 the standard OpenSSL PRNG: set additional data to a date time vector.
5793 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5794 This shouldn't present any incompatibility problems because applications
5795 shouldn't be using these directly and any that are will need to rethink
5796 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5800 * Extensive self tests and health checking required by SP800-90 DRBG.
5801 Remove strength parameter from FIPS_drbg_instantiate and always
5802 instantiate at maximum supported strength.
5806 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5810 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5814 * New function DH_compute_key_padded() to compute a DH key and pad with
5815 leading zeroes if needed: this complies with SP800-56A et al.
5819 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5820 anything, incomplete, subject to change and largely untested at present.
5824 * Modify fipscanisteronly build option to only build the necessary object
5825 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5829 * Add experimental option FIPSSYMS to give all symbols in
5830 fipscanister.o and FIPS or fips prefix. This will avoid
5831 conflicts with future versions of OpenSSL. Add perl script
5832 util/fipsas.pl to preprocess assembly language source files
5833 and rename any affected symbols.
5837 * Add selftest checks and algorithm block of non-fips algorithms in
5838 FIPS mode. Remove DES2 from selftests.
5842 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5843 return internal method without any ENGINE dependencies. Add new
5844 tiny fips sign and verify functions.
5848 * New build option no-ec2m to disable characteristic 2 code.
5852 * New build option "fipscanisteronly". This only builds fipscanister.o
5853 and (currently) associated fips utilities. Uses the file Makefile.fips
5854 instead of Makefile.org as the prototype.
5858 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5859 Update fips_gcmtest to use IV generator.
5863 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5864 setting output buffer to NULL. The `*Final` function must be
5865 called although it will not retrieve any additional data. The tag
5866 can be set or retrieved with a ctrl. The IV length is by default 12
5867 bytes (96 bits) but can be set to an alternative value. If the IV
5868 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5873 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5874 underlying do_cipher function handles all cipher semantics itself
5875 including padding and finalisation. This is useful if (for example)
5876 an ENGINE cipher handles block padding itself. The behaviour of
5877 do_cipher is subtly changed if this flag is set: the return value
5878 is the number of characters written to the output buffer (zero is
5879 no longer an error code) or a negative error code. Also if the
5880 input buffer is NULL and length 0 finalisation should be performed.
5884 * If a candidate issuer certificate is already part of the constructed
5885 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5889 * Improve forward-security support: add functions
5891 void SSL_CTX_set_not_resumable_session_callback(
5892 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5893 void SSL_set_not_resumable_session_callback(
5894 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5896 for use by SSL/TLS servers; the callback function will be called whenever a
5897 new session is created, and gets to decide whether the session may be
5898 cached to make it resumable (return 0) or not (return 1). (As by the
5899 SSL/TLS protocol specifications, the session_id sent by the server will be
5900 empty to indicate that the session is not resumable; also, the server will
5901 not generate RFC 4507 (RFC 5077) session tickets.)
5903 A simple reasonable callback implementation is to return is_forward_secure.
5904 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5905 by the SSL/TLS server library, indicating whether it can provide forward
5908 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5910 * New -verify_name option in command line utilities to set verification
5915 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5916 Add CMAC pkey methods.
5920 * Experimental renegotiation in s_server -www mode. If the client
5921 browses /reneg connection is renegotiated. If /renegcert it is
5922 renegotiated requesting a certificate.
5926 * Add an "external" session cache for debugging purposes to s_server. This
5927 should help trace issues which normally are only apparent in deployed
5928 multi-process servers.
5932 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5933 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5934 BIO_set_cipher() and some obscure PEM functions were changed so they
5935 can now return an error. The RAND changes required a change to the
5936 RAND_METHOD structure.
5940 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5941 a gcc attribute to warn if the result of a function is ignored. This
5942 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5943 whose return value is often ignored.
5947 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5948 These allow SCTs (signed certificate timestamps) to be requested and
5949 validated when establishing a connection.
5951 *Rob Percival <robpercival@google.com>*
5956 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5958 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5959 used even when parsing explicit parameters, when loading a encoded key
5960 or calling `EC_GROUP_new_from_ecpkparameters()`/
5961 `EC_GROUP_new_from_ecparameters()`.
5962 This prevents bypass of security hardening and performance gains,
5963 especially for curves with specialized EC_METHODs.
5964 By default, if a key encoded with explicit parameters is loaded and later
5965 encoded, the output is still encoded with explicit parameters, even if
5966 internally a "named" EC_GROUP is used for computation.
5970 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5971 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5972 NULL. After this change, only the cofactor parameter can be NULL. It also
5973 does some minimal sanity checks on the passed order.
5978 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5979 An attack is simple, if the first CMS_recipientInfo is valid but the
5980 second CMS_recipientInfo is chosen ciphertext. If the second
5981 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5982 encryption key will be replaced by garbage, and the message cannot be
5983 decoded, but if the RSA decryption fails, the correct encryption key is
5984 used and the recipient will not notice the attack.
5985 As a work around for this potential attack the length of the decrypted
5986 key must be equal to the cipher default key length, in case the
5987 certificate is not given and all recipientInfo are tried out.
5988 The old behaviour can be re-enabled in the CMS code by setting the
5989 CMS_DEBUG_DECRYPT flag.
5994 * Document issue with installation paths in diverse Windows builds
5996 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5997 binaries and run-time config file.
6002 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
6004 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
6005 This changes the size when using the `genpkey` command when no size is given.
6006 It fixes an omission in earlier changes that changed all RSA, DSA and DH
6007 generation commands to use 2048 bits by default.
6011 * Add FIPS support for Android Arm 64-bit
6013 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
6014 Module in Version 2.0.10. For some reason, the corresponding target
6015 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
6016 built with FIPS support on Android Arm 64-bit. This omission has been
6019 *Matthias St. Pierre*
6021 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
6023 * 0-byte record padding oracle
6025 If an application encounters a fatal protocol error and then calls
6026 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
6027 then OpenSSL can respond differently to the calling application if a 0 byte
6028 record is received with invalid padding compared to if a 0 byte record is
6029 received with an invalid MAC. If the application then behaves differently
6030 based on that in a way that is detectable to the remote peer, then this
6031 amounts to a padding oracle that could be used to decrypt data.
6033 In order for this to be exploitable "non-stitched" ciphersuites must be in
6034 use. Stitched ciphersuites are optimised implementations of certain
6035 commonly used ciphersuites. Also the application must call SSL_shutdown()
6036 twice even if a protocol error has occurred (applications should not do
6037 this but some do anyway).
6039 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
6040 Aviram, with additional investigation by Steven Collison and Andrew
6041 Hourselt. It was reported to OpenSSL on 10th December 2018.
6046 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
6050 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
6052 * Microarchitecture timing vulnerability in ECC scalar multiplication
6054 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
6055 shown to be vulnerable to a microarchitecture timing side channel attack.
6056 An attacker with sufficient access to mount local timing attacks during
6057 ECDSA signature generation could recover the private key.
6059 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
6060 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
6066 * Timing vulnerability in DSA signature generation
6068 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
6069 timing side channel attack. An attacker could use variations in the signing
6070 algorithm to recover the private key.
6072 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
6077 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
6078 Module, accidentally introduced while backporting security fixes from the
6079 development branch and hindering the use of ECC in FIPS mode.
6083 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
6085 * Client DoS due to large DH parameter
6087 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
6088 malicious server can send a very large prime value to the client. This will
6089 cause the client to spend an unreasonably long period of time generating a
6090 key for this prime resulting in a hang until the client has finished. This
6091 could be exploited in a Denial Of Service attack.
6093 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
6098 * Cache timing vulnerability in RSA Key Generation
6100 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
6101 a cache timing side channel attack. An attacker with sufficient access to
6102 mount cache timing attacks during the RSA key generation process could
6103 recover the private key.
6105 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
6106 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
6111 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
6112 parameter is no longer accepted, as it leads to a corrupt table. NULL
6113 pem_str is reserved for alias entries only.
6117 * Revert blinding in ECDSA sign and instead make problematic addition
6118 length-invariant. Switch even to fixed-length Montgomery multiplication.
6122 * Change generating and checking of primes so that the error rate of not
6123 being prime depends on the intended use based on the size of the input.
6124 For larger primes this will result in more rounds of Miller-Rabin.
6125 The maximal error rate for primes with more than 1080 bits is lowered
6128 *Kurt Roeckx, Annie Yousar*
6130 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6134 * Add blinding to ECDSA and DSA signatures to protect against side channel
6135 attacks discovered by Keegan Ryan (NCC Group).
6139 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6140 now allow empty (zero character) pass phrases.
6144 * Certificate time validation (X509_cmp_time) enforces stricter
6145 compliance with RFC 5280. Fractional seconds and timezone offsets
6146 are no longer allowed.
6150 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6152 * Constructed ASN.1 types with a recursive definition could exceed the stack
6154 Constructed ASN.1 types with a recursive definition (such as can be found
6155 in PKCS7) could eventually exceed the stack given malicious input with
6156 excessive recursion. This could result in a Denial Of Service attack. There
6157 are no such structures used within SSL/TLS that come from untrusted sources
6158 so this is considered safe.
6160 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6166 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6168 * Read/write after SSL object in error state
6170 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6171 mechanism. The intent was that if a fatal error occurred during a handshake
6172 then OpenSSL would move into the error state and would immediately fail if
6173 you attempted to continue the handshake. This works as designed for the
6174 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6175 SSL_connect()), however due to a bug it does not work correctly if
6176 SSL_read() or SSL_write() is called directly. In that scenario, if the
6177 handshake fails then a fatal error will be returned in the initial function
6178 call. If SSL_read()/SSL_write() is subsequently called by the application
6179 for the same SSL object then it will succeed and the data is passed without
6180 being decrypted/encrypted directly from the SSL/TLS record layer.
6182 In order to exploit this issue an application bug would have to be present
6183 that resulted in a call to SSL_read()/SSL_write() being issued after having
6184 already received a fatal error.
6186 This issue was reported to OpenSSL by David Benjamin (Google).
6191 * rsaz_1024_mul_avx2 overflow bug on x86_64
6193 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6194 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6195 Analysis suggests that attacks against RSA and DSA as a result of this
6196 defect would be very difficult to perform and are not believed likely.
6197 Attacks against DH1024 are considered just feasible, because most of the
6198 work necessary to deduce information about a private key may be performed
6199 offline. The amount of resources required for such an attack would be
6200 significant. However, for an attack on TLS to be meaningful, the server
6201 would have to share the DH1024 private key among multiple clients, which is
6202 no longer an option since CVE-2016-0701.
6204 This only affects processors that support the AVX2 but not ADX extensions
6205 like Intel Haswell (4th generation).
6207 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6208 was originally found via the OSS-Fuzz project.
6213 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6215 * bn_sqrx8x_internal carry bug on x86_64
6217 There is a carry propagating bug in the x86_64 Montgomery squaring
6218 procedure. No EC algorithms are affected. Analysis suggests that attacks
6219 against RSA and DSA as a result of this defect would be very difficult to
6220 perform and are not believed likely. Attacks against DH are considered just
6221 feasible (although very difficult) because most of the work necessary to
6222 deduce information about a private key may be performed offline. The amount
6223 of resources required for such an attack would be very significant and
6224 likely only accessible to a limited number of attackers. An attacker would
6225 additionally need online access to an unpatched system using the target
6226 private key in a scenario with persistent DH parameters and a private
6227 key that is shared between multiple clients.
6229 This only affects processors that support the BMI1, BMI2 and ADX extensions
6230 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6232 This issue was reported to OpenSSL by the OSS-Fuzz project.
6237 * Malformed X.509 IPAddressFamily could cause OOB read
6239 If an X.509 certificate has a malformed IPAddressFamily extension,
6240 OpenSSL could do a one-byte buffer overread. The most likely result
6241 would be an erroneous display of the certificate in text format.
6243 This issue was reported to OpenSSL by the OSS-Fuzz project.
6247 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6249 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6250 platform rather than 'mingw'.
6254 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6256 * Truncated packet could crash via OOB read
6258 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6259 cipher is being used, then a truncated packet can cause that host to
6260 perform an out-of-bounds read, usually resulting in a crash.
6262 This issue was reported to OpenSSL by Robert Święcki of Google.
6267 * BN_mod_exp may produce incorrect results on x86_64
6269 There is a carry propagating bug in the x86_64 Montgomery squaring
6270 procedure. No EC algorithms are affected. Analysis suggests that attacks
6271 against RSA and DSA as a result of this defect would be very difficult to
6272 perform and are not believed likely. Attacks against DH are considered just
6273 feasible (although very difficult) because most of the work necessary to
6274 deduce information about a private key may be performed offline. The amount
6275 of resources required for such an attack would be very significant and
6276 likely only accessible to a limited number of attackers. An attacker would
6277 additionally need online access to an unpatched system using the target
6278 private key in a scenario with persistent DH parameters and a private
6279 key that is shared between multiple clients. For example this can occur by
6280 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6281 similar to CVE-2015-3193 but must be treated as a separate problem.
6283 This issue was reported to OpenSSL by the OSS-Fuzz project.
6288 * Montgomery multiplication may produce incorrect results
6290 There is a carry propagating bug in the Broadwell-specific Montgomery
6291 multiplication procedure that handles input lengths divisible by, but
6292 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6293 and DH private keys are impossible. This is because the subroutine in
6294 question is not used in operations with the private key itself and an input
6295 of the attacker's direct choice. Otherwise the bug can manifest itself as
6296 transient authentication and key negotiation failures or reproducible
6297 erroneous outcome of public-key operations with specially crafted input.
6298 Among EC algorithms only Brainpool P-512 curves are affected and one
6299 presumably can attack ECDH key negotiation. Impact was not analyzed in
6300 detail, because pre-requisites for attack are considered unlikely. Namely
6301 multiple clients have to choose the curve in question and the server has to
6302 share the private key among them, neither of which is default behaviour.
6303 Even then only clients that chose the curve will be affected.
6305 This issue was publicly reported as transient failures and was not
6306 initially recognized as a security issue. Thanks to Richard Morgan for
6307 providing reproducible case.
6312 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6313 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6314 prevent issues where no progress is being made and the peer continually
6315 sends unrecognised record types, using up resources processing them.
6319 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6321 * Missing CRL sanity check
6323 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6324 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6325 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6327 This issue only affects the OpenSSL 1.0.2i
6332 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6334 * OCSP Status Request extension unbounded memory growth
6336 A malicious client can send an excessively large OCSP Status Request
6337 extension. If that client continually requests renegotiation, sending a
6338 large OCSP Status Request extension each time, then there will be unbounded
6339 memory growth on the server. This will eventually lead to a Denial Of
6340 Service attack through memory exhaustion. Servers with a default
6341 configuration are vulnerable even if they do not support OCSP. Builds using
6342 the "no-ocsp" build time option are not affected.
6344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6349 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6352 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6358 * OOB write in MDC2_Update()
6360 An overflow can occur in MDC2_Update() either if called directly or
6361 through the EVP_DigestUpdate() function using MDC2. If an attacker
6362 is able to supply very large amounts of input data after a previous
6363 call to EVP_EncryptUpdate() with a partial block then a length check
6364 can overflow resulting in a heap corruption.
6366 The amount of data needed is comparable to SIZE_MAX which is impractical
6369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6374 * Malformed SHA512 ticket DoS
6376 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6377 DoS attack where a malformed ticket will result in an OOB read which will
6380 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6381 a custom server callback and ticket lookup mechanism.
6383 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6388 * OOB write in BN_bn2dec()
6390 The function BN_bn2dec() does not check the return value of BN_div_word().
6391 This can cause an OOB write if an application uses this function with an
6392 overly large BIGNUM. This could be a problem if an overly large certificate
6393 or CRL is printed out from an untrusted source. TLS is not affected because
6394 record limits will reject an oversized certificate before it is parsed.
6396 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6401 * OOB read in TS_OBJ_print_bio()
6403 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6404 the total length the OID text representation would use and not the amount
6405 of data written. This will result in OOB reads when large OIDs are
6408 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6413 * Pointer arithmetic undefined behaviour
6415 Avoid some undefined pointer arithmetic
6417 A common idiom in the codebase is to check limits in the following manner:
6420 Where "p" points to some malloc'd data of SIZE bytes and
6423 "len" here could be from some externally supplied data (e.g. from a TLS
6426 The rules of C pointer arithmetic are such that "p + len" is only well
6427 defined where len <= SIZE. Therefore the above idiom is actually
6428 undefined behaviour.
6430 For example this could cause problems if some malloc implementation
6431 provides an address for "p" such that "p + len" actually overflows for
6432 values of len that are too big and therefore p + len < limit.
6434 This issue was reported to OpenSSL by Guido Vranken
6439 * Constant time flag not preserved in DSA signing
6441 Operations in the DSA signing algorithm should run in constant time in
6442 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6443 implementation means that a non-constant time codepath is followed for
6444 certain operations. This has been demonstrated through a cache-timing
6445 attack to be sufficient for an attacker to recover the private DSA key.
6447 This issue was reported by César Pereida (Aalto University), Billy Brumley
6448 (Tampere University of Technology), and Yuval Yarom (The University of
6449 Adelaide and NICTA).
6454 * DTLS buffered message DoS
6456 In a DTLS connection where handshake messages are delivered out-of-order
6457 those messages that OpenSSL is not yet ready to process will be buffered
6458 for later use. Under certain circumstances, a flaw in the logic means that
6459 those messages do not get removed from the buffer even though the handshake
6460 has been completed. An attacker could force up to approx. 15 messages to
6461 remain in the buffer when they are no longer required. These messages will
6462 be cleared when the DTLS connection is closed. The default maximum size for
6463 a message is 100k. Therefore, the attacker could force an additional 1500k
6464 to be consumed per connection. By opening many simultaneous connections an
6465 attacker could cause a DoS attack through memory exhaustion.
6467 This issue was reported to OpenSSL by Quan Luo.
6472 * DTLS replay protection DoS
6474 A flaw in the DTLS replay attack protection mechanism means that records
6475 that arrive for future epochs update the replay protection "window" before
6476 the MAC for the record has been validated. This could be exploited by an
6477 attacker by sending a record for the next epoch (which does not have to
6478 decrypt or have a valid MAC), with a very large sequence number. This means
6479 that all subsequent legitimate packets are dropped causing a denial of
6480 service for a specific DTLS connection.
6482 This issue was reported to OpenSSL by the OCAP audit team.
6487 * Certificate message OOB reads
6489 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6490 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6491 theoretical DoS risk but this has not been observed in practice on common
6494 The messages affected are client certificate, client certificate request
6495 and server certificate. As a result the attack can only be performed
6496 against a client or a server which enables client authentication.
6498 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6503 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6505 * Prevent padding oracle in AES-NI CBC MAC check
6507 A MITM attacker can use a padding oracle attack to decrypt traffic
6508 when the connection uses an AES CBC cipher and the server support
6511 This issue was introduced as part of the fix for Lucky 13 padding
6512 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6513 constant time by making sure that always the same bytes are read and
6514 compared against either the MAC or padding bytes. But it no longer
6515 checked that there was enough data to have both the MAC and padding
6518 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6522 * Fix EVP_EncodeUpdate overflow
6524 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6525 Base64 encoding of binary data. If an attacker is able to supply very large
6526 amounts of input data then a length check can overflow resulting in a heap
6529 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6530 the `PEM_write_bio*` family of functions. These are mainly used within the
6531 OpenSSL command line applications, so any application which processes data
6532 from an untrusted source and outputs it as a PEM file should be considered
6533 vulnerable to this issue. User applications that call these APIs directly
6534 with large amounts of untrusted data may also be vulnerable.
6536 This issue was reported by Guido Vranken.
6541 * Fix EVP_EncryptUpdate overflow
6543 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6544 is able to supply very large amounts of input data after a previous call to
6545 EVP_EncryptUpdate() with a partial block then a length check can overflow
6546 resulting in a heap corruption. Following an analysis of all OpenSSL
6547 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6548 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6549 the first called function after an EVP_EncryptInit(), and therefore that
6550 specific call must be safe. The second form is where the length passed to
6551 EVP_EncryptUpdate() can be seen from the code to be some small value and
6552 therefore there is no possibility of an overflow. Since all instances are
6553 one of these two forms, it is believed that there can be no overflows in
6554 internal code due to this problem. It should be noted that
6555 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6556 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6557 of these calls have also been analysed too and it is believed there are no
6558 instances in internal usage where an overflow could occur.
6560 This issue was reported by Guido Vranken.
6565 * Prevent ASN.1 BIO excessive memory allocation
6567 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6568 a short invalid encoding can cause allocation of large amounts of memory
6569 potentially consuming excessive resources or exhausting memory.
6571 Any application parsing untrusted data through d2i BIO functions is
6572 affected. The memory based functions such as d2i_X509() are *not* affected.
6573 Since the memory based functions are used by the TLS library, TLS
6574 applications are not affected.
6576 This issue was reported by Brian Carpenter.
6583 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6584 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6585 in arbitrary stack data being returned in the buffer.
6587 This issue was reported by Guido Vranken.
6592 * Modify behavior of ALPN to invoke callback after SNI/servername
6593 callback, such that updates to the SSL_CTX affect ALPN.
6597 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6602 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6603 methods are enabled and ssl2 is disabled the methods return NULL.
6607 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6609 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6610 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6611 provide any "EXPORT" or "LOW" strength ciphers.
6615 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6616 is by default disabled at build-time. Builds that are not configured with
6617 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6618 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6619 will need to explicitly call either of:
6621 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6623 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6625 as appropriate. Even if either of those is used, or the application
6626 explicitly uses the version-specific SSLv2_method() or its client and
6627 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6628 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6629 ciphers, and SSLv2 56-bit DES are no longer available.
6634 * Fix a double-free in DSA code
6636 A double free bug was discovered when OpenSSL parses malformed DSA private
6637 keys and could lead to a DoS attack or memory corruption for applications
6638 that receive DSA private keys from untrusted sources. This scenario is
6641 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6647 * Disable SRP fake user seed to address a server memory leak.
6649 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6651 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6652 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6653 was changed to ignore the "fake user" SRP seed, even if the seed
6656 Users should use SRP_VBASE_get1_by_user instead. Note that in
6657 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6658 also that even though configuring the SRP seed attempts to hide
6659 invalid usernames by continuing the handshake with fake
6660 credentials, this behaviour is not constant time and no strong
6661 guarantees are made that the handshake is indistinguishable from
6662 that of a valid user.
6667 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6669 In the BN_hex2bn function the number of hex digits is calculated using an
6670 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6671 large values of `i` this can result in `bn_expand` not allocating any
6672 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6673 field as NULL leading to a subsequent NULL ptr deref. For very large values
6674 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6675 In this case memory is allocated to the internal BIGNUM data field, but it
6676 is insufficiently sized leading to heap corruption. A similar issue exists
6677 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6678 is ever called by user applications with very large untrusted hex/dec data.
6679 This is anticipated to be a rare occurrence.
6681 All OpenSSL internal usage of these functions use data that is not expected
6682 to be untrusted, e.g. config file data or application command line
6683 arguments. If user developed applications generate config file data based
6684 on untrusted data then it is possible that this could also lead to security
6685 consequences. This is also anticipated to be rare.
6687 This issue was reported to OpenSSL by Guido Vranken.
6692 * Fix memory issues in `BIO_*printf` functions
6694 The internal `fmtstr` function used in processing a "%s" format string in
6695 the `BIO_*printf` functions could overflow while calculating the length of a
6696 string and cause an OOB read when printing very long strings.
6698 Additionally the internal `doapr_outch` function can attempt to write to an
6699 OOB memory location (at an offset from the NULL pointer) in the event of a
6700 memory allocation failure. In 1.0.2 and below this could be caused where
6701 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6702 could be in processing a very long "%s" format string. Memory leaks can
6705 The first issue may mask the second issue dependent on compiler behaviour.
6706 These problems could enable attacks where large amounts of untrusted data
6707 is passed to the `BIO_*printf` functions. If applications use these functions
6708 in this way then they could be vulnerable. OpenSSL itself uses these
6709 functions when printing out human-readable dumps of ASN.1 data. Therefore
6710 applications that print this data could be vulnerable if the data is from
6711 untrusted sources. OpenSSL command line applications could also be
6712 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6713 as command line arguments.
6715 Libssl is not considered directly vulnerable. Additionally certificates etc
6716 received via remote connections via libssl are also unlikely to be able to
6717 trigger these issues because of message size limits enforced within libssl.
6719 This issue was reported to OpenSSL Guido Vranken.
6724 * Side channel attack on modular exponentiation
6726 A side-channel attack was found which makes use of cache-bank conflicts on
6727 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6728 of RSA keys. The ability to exploit this issue is limited as it relies on
6729 an attacker who has control of code in a thread running on the same
6730 hyper-threaded core as the victim thread which is performing decryptions.
6732 This issue was reported to OpenSSL by Yuval Yarom, The University of
6733 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6734 Nadia Heninger, University of Pennsylvania with more information at
6735 <http://cachebleed.info>.
6740 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6741 if no keysize is specified with default_bits. This fixes an
6742 omission in an earlier change that changed all RSA/DSA key generation
6743 commands to use 2048 bits by default.
6747 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6749 * DH small subgroups
6751 Historically OpenSSL only ever generated DH parameters based on "safe"
6752 primes. More recently (in version 1.0.2) support was provided for
6753 generating X9.42 style parameter files such as those required for RFC 5114
6754 support. The primes used in such files may not be "safe". Where an
6755 application is using DH configured with parameters based on primes that are
6756 not "safe" then an attacker could use this fact to find a peer's private
6757 DH exponent. This attack requires that the attacker complete multiple
6758 handshakes in which the peer uses the same private DH exponent. For example
6759 this could be used to discover a TLS server's private DH exponent if it's
6760 reusing the private DH exponent or it's using a static DH ciphersuite.
6762 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6763 TLS. It is not on by default. If the option is not set then the server
6764 reuses the same private DH exponent for the life of the server process and
6765 would be vulnerable to this attack. It is believed that many popular
6766 applications do set this option and would therefore not be at risk.
6768 The fix for this issue adds an additional check where a "q" parameter is
6769 available (as is the case in X9.42 based parameters). This detects the
6770 only known attack, and is the only possible defense for static DH
6771 ciphersuites. This could have some performance impact.
6773 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6774 default and cannot be disabled. This could have some performance impact.
6776 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6781 * SSLv2 doesn't block disabled ciphers
6783 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6784 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6785 been disabled, provided that the SSLv2 protocol was not also disabled via
6788 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6789 and Sebastian Schinzel.
6794 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6796 * BN_mod_exp may produce incorrect results on x86_64
6798 There is a carry propagating bug in the x86_64 Montgomery squaring
6799 procedure. No EC algorithms are affected. Analysis suggests that attacks
6800 against RSA and DSA as a result of this defect would be very difficult to
6801 perform and are not believed likely. Attacks against DH are considered just
6802 feasible (although very difficult) because most of the work necessary to
6803 deduce information about a private key may be performed offline. The amount
6804 of resources required for such an attack would be very significant and
6805 likely only accessible to a limited number of attackers. An attacker would
6806 additionally need online access to an unpatched system using the target
6807 private key in a scenario with persistent DH parameters and a private
6808 key that is shared between multiple clients. For example this can occur by
6809 default in OpenSSL DHE based SSL/TLS ciphersuites.
6811 This issue was reported to OpenSSL by Hanno Böck.
6816 * Certificate verify crash with missing PSS parameter
6818 The signature verification routines will crash with a NULL pointer
6819 dereference if presented with an ASN.1 signature using the RSA PSS
6820 algorithm and absent mask generation function parameter. Since these
6821 routines are used to verify certificate signature algorithms this can be
6822 used to crash any certificate verification operation and exploited in a
6823 DoS attack. Any application which performs certificate verification is
6824 vulnerable including OpenSSL clients and servers which enable client
6827 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6832 * X509_ATTRIBUTE memory leak
6834 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6835 memory. This structure is used by the PKCS#7 and CMS routines so any
6836 application which reads PKCS#7 or CMS data from untrusted sources is
6837 affected. SSL/TLS is not affected.
6839 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6845 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6846 This changes the decoding behaviour for some invalid messages,
6847 though the change is mostly in the more lenient direction, and
6848 legacy behaviour is preserved as much as possible.
6852 * In DSA_generate_parameters_ex, if the provided seed is too short,
6855 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6857 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6859 * Alternate chains certificate forgery
6861 During certificate verification, OpenSSL will attempt to find an
6862 alternative certificate chain if the first attempt to build such a chain
6863 fails. An error in the implementation of this logic can mean that an
6864 attacker could cause certain checks on untrusted certificates to be
6865 bypassed, such as the CA flag, enabling them to use a valid leaf
6866 certificate to act as a CA and "issue" an invalid certificate.
6868 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6873 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6875 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6876 incompatibility in the handling of HMAC. The previous ABI has now been
6881 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6883 * Malformed ECParameters causes infinite loop
6885 When processing an ECParameters structure OpenSSL enters an infinite loop
6886 if the curve specified is over a specially malformed binary polynomial
6889 This can be used to perform denial of service against any
6890 system which processes public keys, certificate requests or
6891 certificates. This includes TLS clients and TLS servers with
6892 client authentication enabled.
6894 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6899 * Exploitable out-of-bounds read in X509_cmp_time
6901 X509_cmp_time does not properly check the length of the ASN1_TIME
6902 string and can read a few bytes out of bounds. In addition,
6903 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6906 An attacker can use this to craft malformed certificates and CRLs of
6907 various sizes and potentially cause a segmentation fault, resulting in
6908 a DoS on applications that verify certificates or CRLs. TLS clients
6909 that verify CRLs are affected. TLS clients and servers with client
6910 authentication enabled may be affected if they use custom verification
6913 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6914 independently by Hanno Böck.
6919 * PKCS7 crash with missing EnvelopedContent
6921 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6922 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6923 with missing content and trigger a NULL pointer dereference on parsing.
6925 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6926 structures from untrusted sources are affected. OpenSSL clients and
6927 servers are not affected.
6929 This issue was reported to OpenSSL by Michal Zalewski (Google).
6934 * CMS verify infinite loop with unknown hash function
6936 When verifying a signedData message the CMS code can enter an infinite loop
6937 if presented with an unknown hash function OID. This can be used to perform
6938 denial of service against any system which verifies signedData messages using
6940 This issue was reported to OpenSSL by Johannes Bauer.
6945 * Race condition handling NewSessionTicket
6947 If a NewSessionTicket is received by a multi-threaded client when attempting to
6948 reuse a previous ticket then a race condition can occur potentially leading to
6949 a double free of the ticket data.
6954 * Only support 256-bit or stronger elliptic curves with the
6955 'ecdh_auto' setting (server) or by default (client). Of supported
6956 curves, prefer P-256 (both).
6960 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6962 * ClientHello sigalgs DoS fix
6964 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6965 invalid signature algorithms extension a NULL pointer dereference will
6966 occur. This can be exploited in a DoS attack against the server.
6968 This issue was was reported to OpenSSL by David Ramos of Stanford
6972 *Stephen Henson and Matt Caswell*
6974 * Multiblock corrupted pointer fix
6976 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6977 feature only applies on 64 bit x86 architecture platforms that support AES
6978 NI instructions. A defect in the implementation of "multiblock" can cause
6979 OpenSSL's internal write buffer to become incorrectly set to NULL when
6980 using non-blocking IO. Typically, when the user application is using a
6981 socket BIO for writing, this will only result in a failed connection.
6982 However if some other BIO is used then it is likely that a segmentation
6983 fault will be triggered, thus enabling a potential DoS attack.
6985 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6990 * Segmentation fault in DTLSv1_listen fix
6992 The DTLSv1_listen function is intended to be stateless and processes the
6993 initial ClientHello from many peers. It is common for user code to loop
6994 over the call to DTLSv1_listen until a valid ClientHello is received with
6995 an associated cookie. A defect in the implementation of DTLSv1_listen means
6996 that state is preserved in the SSL object from one invocation to the next
6997 that can lead to a segmentation fault. Errors processing the initial
6998 ClientHello can trigger this scenario. An example of such an error could be
6999 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
7002 This issue was reported to OpenSSL by Per Allansson.
7007 * Segmentation fault in ASN1_TYPE_cmp fix
7009 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7010 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7011 certificate signature algorithm consistency this can be used to crash any
7012 certificate verification operation and exploited in a DoS attack. Any
7013 application which performs certificate verification is vulnerable including
7014 OpenSSL clients and servers which enable client authentication.
7019 * Segmentation fault for invalid PSS parameters fix
7021 The signature verification routines will crash with a NULL pointer
7022 dereference if presented with an ASN.1 signature using the RSA PSS
7023 algorithm and invalid parameters. Since these routines are used to verify
7024 certificate signature algorithms this can be used to crash any
7025 certificate verification operation and exploited in a DoS attack. Any
7026 application which performs certificate verification is vulnerable including
7027 OpenSSL clients and servers which enable client authentication.
7029 This issue was was reported to OpenSSL by Brian Carpenter.
7034 * ASN.1 structure reuse memory corruption fix
7036 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7037 memory corruption via an invalid write. Such reuse is and has been
7038 strongly discouraged and is believed to be rare.
7040 Applications that parse structures containing CHOICE or ANY DEFINED BY
7041 components may be affected. Certificate parsing (d2i_X509 and related
7042 functions) are however not affected. OpenSSL clients and servers are
7048 * PKCS7 NULL pointer dereferences fix
7050 The PKCS#7 parsing code does not handle missing outer ContentInfo
7051 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7052 missing content and trigger a NULL pointer dereference on parsing.
7054 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7055 otherwise parse PKCS#7 structures from untrusted sources are
7056 affected. OpenSSL clients and servers are not affected.
7058 This issue was reported to OpenSSL by Michal Zalewski (Google).
7063 * DoS via reachable assert in SSLv2 servers fix
7065 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7066 servers that both support SSLv2 and enable export cipher suites by sending
7067 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7069 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7070 (OpenSSL development team).
7075 * Empty CKE with client auth and DHE fix
7077 If client auth is used then a server can seg fault in the event of a DHE
7078 ciphersuite being selected and a zero length ClientKeyExchange message
7079 being sent by the client. This could be exploited in a DoS attack.
7084 * Handshake with unseeded PRNG fix
7086 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
7087 with an unseeded PRNG. The conditions are:
7088 - The client is on a platform where the PRNG has not been seeded
7089 automatically, and the user has not seeded manually
7090 - A protocol specific client method version has been used (i.e. not
7091 SSL_client_methodv23)
7092 - A ciphersuite is used that does not require additional random data from
7093 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
7095 If the handshake succeeds then the client random that has been used will
7096 have been generated from a PRNG with insufficient entropy and therefore the
7097 output may be predictable.
7099 For example using the following command with an unseeded openssl will
7100 succeed on an unpatched platform:
7102 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
7107 * Use After Free following d2i_ECPrivatekey error fix
7109 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7110 could cause a use after free condition. This, in turn, could cause a double
7111 free in several private key parsing functions (such as d2i_PrivateKey
7112 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7113 for applications that receive EC private keys from untrusted
7114 sources. This scenario is considered rare.
7116 This issue was discovered by the BoringSSL project and fixed in their
7122 * X509_to_X509_REQ NULL pointer deref fix
7124 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7125 the certificate key is invalid. This function is rarely used in practice.
7127 This issue was discovered by Brian Carpenter.
7132 * Removed the export ciphers from the DEFAULT ciphers
7136 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7138 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7139 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7140 So far those who have to target multiple platforms would compromise
7141 and argue that binary targeting say ARMv5 would still execute on
7142 ARMv8. "Universal" build resolves this compromise by providing
7143 near-optimal performance even on newer platforms.
7147 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7148 (other platforms pending).
7150 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7152 * Add support for the SignedCertificateTimestampList certificate and
7153 OCSP response extensions from RFC6962.
7157 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7158 for corner cases. (Certain input points at infinity could lead to
7159 bogus results, with non-infinity inputs mapped to infinity too.)
7163 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7164 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7165 common cases are optimized and there still is room for further
7166 improvements. Vector Permutation AES for Altivec is also added.
7170 * Add support for little-endian ppc64 Linux target.
7172 *Marcelo Cerri (IBM)*
7174 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7175 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7176 are optimized and there still is room for further improvements.
7177 Both 32- and 64-bit modes are supported.
7179 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7181 * Improved ARMv7 NEON support.
7185 * Support for SPARC Architecture 2011 crypto extensions, first
7186 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7187 SHA256/512, MD5, GHASH and modular exponentiation.
7189 *Andy Polyakov, David Miller*
7191 * Accelerated modular exponentiation for Intel processors, a.k.a.
7194 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7196 * Support for new and upcoming Intel processors, including AVX2,
7197 BMI and SHA ISA extensions. This includes additional "stitched"
7198 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7201 This work was sponsored by Intel Corp.
7205 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7206 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7207 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7211 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7212 this fixes a limitation in previous versions of OpenSSL.
7216 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7217 MGF1 digest and OAEP label.
7221 * Add EVP support for key wrapping algorithms, to avoid problems with
7222 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7223 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7224 algorithms and include tests cases.
7228 * Add functions to allocate and set the fields of an ECDSA_METHOD
7231 *Douglas E. Engert, Steve Henson*
7233 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7234 difference in days and seconds between two tm or ASN1_TIME structures.
7238 * Add -rev test option to s_server to just reverse order of characters
7239 received by client and send back to server. Also prints an abbreviated
7240 summary of the connection parameters.
7244 * New option -brief for s_client and s_server to print out a brief summary
7245 of connection parameters.
7249 * Add callbacks for arbitrary TLS extensions.
7251 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7253 * New option -crl_download in several openssl utilities to download CRLs
7254 from CRLDP extension in certificates.
7258 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7262 * New function X509_CRL_diff to generate a delta CRL from the difference
7263 of two full CRLs. Add support to "crl" utility.
7267 * New functions to set lookup_crls function and to retrieve
7268 X509_STORE from X509_STORE_CTX.
7272 * Print out deprecated issuer and subject unique ID fields in
7277 * Extend OCSP I/O functions so they can be used for simple general purpose
7278 HTTP as well as OCSP. New wrapper function which can be used to download
7279 CRLs using the OCSP API.
7283 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7287 * `SSL_CONF*` functions. These provide a common framework for application
7288 configuration using configuration files or command lines.
7292 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7293 message callback and prints the results. Needs compile time option
7294 "enable-ssl-trace". New options to s_client and s_server to enable
7299 * New ctrl and macro to retrieve supported points extensions.
7300 Print out extension in s_server and s_client.
7304 * New functions to retrieve certificate signature and signature
7309 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7314 * New Suite B modes for TLS code. These use and enforce the requirements
7315 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7316 only use Suite B curves. The Suite B modes can be set by using the
7317 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7321 * New chain verification flags for Suite B levels of security. Check
7322 algorithms are acceptable when flags are set in X509_verify_cert.
7326 * Make tls1_check_chain return a set of flags indicating checks passed
7327 by a certificate chain. Add additional tests to handle client
7328 certificates: checks for matching certificate type and issuer name
7333 * If an attempt is made to use a signature algorithm not in the peer
7334 preference list abort the handshake. If client has no suitable
7335 signature algorithms in response to a certificate request do not
7336 use the certificate.
7340 * If server EC tmp key is not in client preference list abort handshake.
7344 * Add support for certificate stores in CERT structure. This makes it
7345 possible to have different stores per SSL structure or one store in
7346 the parent SSL_CTX. Include distinct stores for certificate chain
7347 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7348 to build and store a certificate chain in CERT structure: returning
7349 an error if the chain cannot be built: this will allow applications
7350 to test if a chain is correctly configured.
7352 Note: if the CERT based stores are not set then the parent SSL_CTX
7353 store is used to retain compatibility with existing behaviour.
7357 * New function ssl_set_client_disabled to set a ciphersuite disabled
7358 mask based on the current session, check mask when sending client
7359 hello and checking the requested ciphersuite.
7363 * New ctrls to retrieve and set certificate types in a certificate
7364 request message. Print out received values in s_client. If certificate
7365 types is not set with custom values set sensible values based on
7366 supported signature algorithms.
7370 * Support for distinct client and server supported signature algorithms.
7374 * Add certificate callback. If set this is called whenever a certificate
7375 is required by client or server. An application can decide which
7376 certificate chain to present based on arbitrary criteria: for example
7377 supported signature algorithms. Add very simple example to s_server.
7378 This fixes many of the problems and restrictions of the existing client
7379 certificate callback: for example you can now clear an existing
7380 certificate and specify the whole chain.
7384 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7385 the certificate can be used for (if anything). Set valid_flags field
7386 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7387 to have similar checks in it.
7389 Add new "cert_flags" field to CERT structure and include a "strict mode".
7390 This enforces some TLS certificate requirements (such as only permitting
7391 certificate signature algorithms contained in the supported algorithms
7392 extension) which some implementations ignore: this option should be used
7393 with caution as it could cause interoperability issues.
7397 * Update and tidy signature algorithm extension processing. Work out
7398 shared signature algorithms based on preferences and peer algorithms
7399 and print them out in s_client and s_server. Abort handshake if no
7400 shared signature algorithms.
7404 * Add new functions to allow customised supported signature algorithms
7405 for SSL and SSL_CTX structures. Add options to s_client and s_server
7410 * New function SSL_certs_clear() to delete all references to certificates
7411 from an SSL structure. Before this once a certificate had been added
7412 it couldn't be removed.
7416 * Integrate hostname, email address and IP address checking with certificate
7417 verification. New verify options supporting checking in openssl utility.
7421 * Fixes and wildcard matching support to hostname and email checking
7422 functions. Add manual page.
7424 *Florian Weimer (Red Hat Product Security Team)*
7426 * New functions to check a hostname email or IP address against a
7427 certificate. Add options x509 utility to print results of checks against
7432 * Fix OCSP checking.
7434 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7436 * Initial experimental support for explicitly trusted non-root CAs.
7437 OpenSSL still tries to build a complete chain to a root but if an
7438 intermediate CA has a trust setting included that is used. The first
7439 setting is used: whether to trust (e.g., -addtrust option to the x509
7444 * Add -trusted_first option which attempts to find certificates in the
7445 trusted store even if an untrusted chain is also supplied.
7449 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7450 platform support for Linux and Android.
7454 * Support for linux-x32, ILP32 environment in x86_64 framework.
7458 * Experimental multi-implementation support for FIPS capable OpenSSL.
7459 When in FIPS mode the approved implementations are used as normal,
7460 when not in FIPS mode the internal unapproved versions are used instead.
7461 This means that the FIPS capable OpenSSL isn't forced to use the
7462 (often lower performance) FIPS implementations outside FIPS mode.
7466 * Transparently support X9.42 DH parameters when calling
7467 PEM_read_bio_DHparameters. This means existing applications can handle
7468 the new parameter format automatically.
7472 * Initial experimental support for X9.42 DH parameter format: mainly
7473 to support use of 'q' parameter for RFC5114 parameters.
7477 * Add DH parameters from RFC5114 including test data to dhtest.
7481 * Support for automatic EC temporary key parameter selection. If enabled
7482 the most preferred EC parameters are automatically used instead of
7483 hardcoded fixed parameters. Now a server just has to call:
7484 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7485 support ECDH and use the most appropriate parameters.
7489 * Enhance and tidy EC curve and point format TLS extension code. Use
7490 static structures instead of allocation if default values are used.
7491 New ctrls to set curves we wish to support and to retrieve shared curves.
7492 Print out shared curves in s_server. New options to s_server and s_client
7493 to set list of supported curves.
7497 * New ctrls to retrieve supported signature algorithms and
7498 supported curve values as an array of NIDs. Extend openssl utility
7499 to print out received values.
7503 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7504 between NIDs and the more common NIST names such as "P-256". Enhance
7505 ecparam utility and ECC method to recognise the NIST names for curves.
7509 * Enhance SSL/TLS certificate chain handling to support different
7510 chains for each certificate instead of one chain in the parent SSL_CTX.
7514 * Support for fixed DH ciphersuite client authentication: where both
7515 server and client use DH certificates with common parameters.
7519 * Support for fixed DH ciphersuites: those requiring DH server
7524 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7526 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7527 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7528 X509_CINF_get_signature were reverted post internal team review.
7533 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7535 * OCSP Status Request extension unbounded memory growth
7537 A malicious client can send an excessively large OCSP Status Request
7538 extension. If that client continually requests renegotiation, sending a
7539 large OCSP Status Request extension each time, then there will be unbounded
7540 memory growth on the server. This will eventually lead to a Denial Of
7541 Service attack through memory exhaustion. Servers with a default
7542 configuration are vulnerable even if they do not support OCSP. Builds using
7543 the "no-ocsp" build time option are not affected.
7545 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7550 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7553 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7559 * OOB write in MDC2_Update()
7561 An overflow can occur in MDC2_Update() either if called directly or
7562 through the EVP_DigestUpdate() function using MDC2. If an attacker
7563 is able to supply very large amounts of input data after a previous
7564 call to EVP_EncryptUpdate() with a partial block then a length check
7565 can overflow resulting in a heap corruption.
7567 The amount of data needed is comparable to SIZE_MAX which is impractical
7570 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7575 * Malformed SHA512 ticket DoS
7577 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7578 DoS attack where a malformed ticket will result in an OOB read which will
7581 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7582 a custom server callback and ticket lookup mechanism.
7584 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7589 * OOB write in BN_bn2dec()
7591 The function BN_bn2dec() does not check the return value of BN_div_word().
7592 This can cause an OOB write if an application uses this function with an
7593 overly large BIGNUM. This could be a problem if an overly large certificate
7594 or CRL is printed out from an untrusted source. TLS is not affected because
7595 record limits will reject an oversized certificate before it is parsed.
7597 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7602 * OOB read in TS_OBJ_print_bio()
7604 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7605 the total length the OID text representation would use and not the amount
7606 of data written. This will result in OOB reads when large OIDs are
7609 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7614 * Pointer arithmetic undefined behaviour
7616 Avoid some undefined pointer arithmetic
7618 A common idiom in the codebase is to check limits in the following manner:
7621 Where "p" points to some malloc'd data of SIZE bytes and
7624 "len" here could be from some externally supplied data (e.g. from a TLS
7627 The rules of C pointer arithmetic are such that "p + len" is only well
7628 defined where len <= SIZE. Therefore, the above idiom is actually
7629 undefined behaviour.
7631 For example this could cause problems if some malloc implementation
7632 provides an address for "p" such that "p + len" actually overflows for
7633 values of len that are too big and therefore p + len < limit.
7635 This issue was reported to OpenSSL by Guido Vranken
7640 * Constant time flag not preserved in DSA signing
7642 Operations in the DSA signing algorithm should run in constant time in
7643 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7644 implementation means that a non-constant time codepath is followed for
7645 certain operations. This has been demonstrated through a cache-timing
7646 attack to be sufficient for an attacker to recover the private DSA key.
7648 This issue was reported by César Pereida (Aalto University), Billy Brumley
7649 (Tampere University of Technology), and Yuval Yarom (The University of
7650 Adelaide and NICTA).
7655 * DTLS buffered message DoS
7657 In a DTLS connection where handshake messages are delivered out-of-order
7658 those messages that OpenSSL is not yet ready to process will be buffered
7659 for later use. Under certain circumstances, a flaw in the logic means that
7660 those messages do not get removed from the buffer even though the handshake
7661 has been completed. An attacker could force up to approx. 15 messages to
7662 remain in the buffer when they are no longer required. These messages will
7663 be cleared when the DTLS connection is closed. The default maximum size for
7664 a message is 100k. Therefore, the attacker could force an additional 1500k
7665 to be consumed per connection. By opening many simultaneous connections an
7666 attacker could cause a DoS attack through memory exhaustion.
7668 This issue was reported to OpenSSL by Quan Luo.
7673 * DTLS replay protection DoS
7675 A flaw in the DTLS replay attack protection mechanism means that records
7676 that arrive for future epochs update the replay protection "window" before
7677 the MAC for the record has been validated. This could be exploited by an
7678 attacker by sending a record for the next epoch (which does not have to
7679 decrypt or have a valid MAC), with a very large sequence number. This means
7680 that all subsequent legitimate packets are dropped causing a denial of
7681 service for a specific DTLS connection.
7683 This issue was reported to OpenSSL by the OCAP audit team.
7688 * Certificate message OOB reads
7690 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7691 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7692 theoretical DoS risk but this has not been observed in practice on common
7695 The messages affected are client certificate, client certificate request
7696 and server certificate. As a result the attack can only be performed
7697 against a client or a server which enables client authentication.
7699 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7704 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7706 * Prevent padding oracle in AES-NI CBC MAC check
7708 A MITM attacker can use a padding oracle attack to decrypt traffic
7709 when the connection uses an AES CBC cipher and the server support
7712 This issue was introduced as part of the fix for Lucky 13 padding
7713 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7714 constant time by making sure that always the same bytes are read and
7715 compared against either the MAC or padding bytes. But it no longer
7716 checked that there was enough data to have both the MAC and padding
7719 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7724 * Fix EVP_EncodeUpdate overflow
7726 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7727 Base64 encoding of binary data. If an attacker is able to supply very large
7728 amounts of input data then a length check can overflow resulting in a heap
7731 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7732 the `PEM_write_bio*` family of functions. These are mainly used within the
7733 OpenSSL command line applications, so any application which processes data
7734 from an untrusted source and outputs it as a PEM file should be considered
7735 vulnerable to this issue. User applications that call these APIs directly
7736 with large amounts of untrusted data may also be vulnerable.
7738 This issue was reported by Guido Vranken.
7743 * Fix EVP_EncryptUpdate overflow
7745 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7746 is able to supply very large amounts of input data after a previous call to
7747 EVP_EncryptUpdate() with a partial block then a length check can overflow
7748 resulting in a heap corruption. Following an analysis of all OpenSSL
7749 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7750 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7751 the first called function after an EVP_EncryptInit(), and therefore that
7752 specific call must be safe. The second form is where the length passed to
7753 EVP_EncryptUpdate() can be seen from the code to be some small value and
7754 therefore there is no possibility of an overflow. Since all instances are
7755 one of these two forms, it is believed that there can be no overflows in
7756 internal code due to this problem. It should be noted that
7757 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7758 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7759 of these calls have also been analysed too and it is believed there are no
7760 instances in internal usage where an overflow could occur.
7762 This issue was reported by Guido Vranken.
7767 * Prevent ASN.1 BIO excessive memory allocation
7769 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7770 a short invalid encoding can casuse allocation of large amounts of memory
7771 potentially consuming excessive resources or exhausting memory.
7773 Any application parsing untrusted data through d2i BIO functions is
7774 affected. The memory based functions such as d2i_X509() are *not* affected.
7775 Since the memory based functions are used by the TLS library, TLS
7776 applications are not affected.
7778 This issue was reported by Brian Carpenter.
7785 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7786 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7787 in arbitrary stack data being returned in the buffer.
7789 This issue was reported by Guido Vranken.
7794 * Modify behavior of ALPN to invoke callback after SNI/servername
7795 callback, such that updates to the SSL_CTX affect ALPN.
7799 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7804 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7805 methods are enabled and ssl2 is disabled the methods return NULL.
7809 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7811 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7812 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7813 provide any "EXPORT" or "LOW" strength ciphers.
7817 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7818 is by default disabled at build-time. Builds that are not configured with
7819 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7820 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7821 will need to explicitly call either of:
7823 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7825 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7827 as appropriate. Even if either of those is used, or the application
7828 explicitly uses the version-specific SSLv2_method() or its client and
7829 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7830 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7831 ciphers, and SSLv2 56-bit DES are no longer available.
7836 * Fix a double-free in DSA code
7838 A double free bug was discovered when OpenSSL parses malformed DSA private
7839 keys and could lead to a DoS attack or memory corruption for applications
7840 that receive DSA private keys from untrusted sources. This scenario is
7843 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7849 * Disable SRP fake user seed to address a server memory leak.
7851 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7853 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7854 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7855 was changed to ignore the "fake user" SRP seed, even if the seed
7858 Users should use SRP_VBASE_get1_by_user instead. Note that in
7859 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7860 also that even though configuring the SRP seed attempts to hide
7861 invalid usernames by continuing the handshake with fake
7862 credentials, this behaviour is not constant time and no strong
7863 guarantees are made that the handshake is indistinguishable from
7864 that of a valid user.
7869 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7871 In the BN_hex2bn function the number of hex digits is calculated using an
7872 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7873 large values of `i` this can result in `bn_expand` not allocating any
7874 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7875 field as NULL leading to a subsequent NULL ptr deref. For very large values
7876 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7877 In this case memory is allocated to the internal BIGNUM data field, but it
7878 is insufficiently sized leading to heap corruption. A similar issue exists
7879 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7880 is ever called by user applications with very large untrusted hex/dec data.
7881 This is anticipated to be a rare occurrence.
7883 All OpenSSL internal usage of these functions use data that is not expected
7884 to be untrusted, e.g. config file data or application command line
7885 arguments. If user developed applications generate config file data based
7886 on untrusted data then it is possible that this could also lead to security
7887 consequences. This is also anticipated to be rare.
7889 This issue was reported to OpenSSL by Guido Vranken.
7894 * Fix memory issues in `BIO_*printf` functions
7896 The internal `fmtstr` function used in processing a "%s" format string in
7897 the `BIO_*printf` functions could overflow while calculating the length of a
7898 string and cause an OOB read when printing very long strings.
7900 Additionally the internal `doapr_outch` function can attempt to write to an
7901 OOB memory location (at an offset from the NULL pointer) in the event of a
7902 memory allocation failure. In 1.0.2 and below this could be caused where
7903 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7904 could be in processing a very long "%s" format string. Memory leaks can
7907 The first issue may mask the second issue dependent on compiler behaviour.
7908 These problems could enable attacks where large amounts of untrusted data
7909 is passed to the `BIO_*printf` functions. If applications use these functions
7910 in this way then they could be vulnerable. OpenSSL itself uses these
7911 functions when printing out human-readable dumps of ASN.1 data. Therefore
7912 applications that print this data could be vulnerable if the data is from
7913 untrusted sources. OpenSSL command line applications could also be
7914 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7915 as command line arguments.
7917 Libssl is not considered directly vulnerable. Additionally certificates etc
7918 received via remote connections via libssl are also unlikely to be able to
7919 trigger these issues because of message size limits enforced within libssl.
7921 This issue was reported to OpenSSL Guido Vranken.
7926 * Side channel attack on modular exponentiation
7928 A side-channel attack was found which makes use of cache-bank conflicts on
7929 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7930 of RSA keys. The ability to exploit this issue is limited as it relies on
7931 an attacker who has control of code in a thread running on the same
7932 hyper-threaded core as the victim thread which is performing decryptions.
7934 This issue was reported to OpenSSL by Yuval Yarom, The University of
7935 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7936 Nadia Heninger, University of Pennsylvania with more information at
7937 <http://cachebleed.info>.
7942 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7943 if no keysize is specified with default_bits. This fixes an
7944 omission in an earlier change that changed all RSA/DSA key generation
7945 commands to use 2048 bits by default.
7949 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7951 * Protection for DH small subgroup attacks
7953 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7954 switched on by default and cannot be disabled. This could have some
7959 * SSLv2 doesn't block disabled ciphers
7961 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7962 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7963 been disabled, provided that the SSLv2 protocol was not also disabled via
7966 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7967 and Sebastian Schinzel.
7972 * Reject DH handshakes with parameters shorter than 1024 bits.
7976 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7978 * Certificate verify crash with missing PSS parameter
7980 The signature verification routines will crash with a NULL pointer
7981 dereference if presented with an ASN.1 signature using the RSA PSS
7982 algorithm and absent mask generation function parameter. Since these
7983 routines are used to verify certificate signature algorithms this can be
7984 used to crash any certificate verification operation and exploited in a
7985 DoS attack. Any application which performs certificate verification is
7986 vulnerable including OpenSSL clients and servers which enable client
7989 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7994 * X509_ATTRIBUTE memory leak
7996 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7997 memory. This structure is used by the PKCS#7 and CMS routines so any
7998 application which reads PKCS#7 or CMS data from untrusted sources is
7999 affected. SSL/TLS is not affected.
8001 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8007 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
8008 This changes the decoding behaviour for some invalid messages,
8009 though the change is mostly in the more lenient direction, and
8010 legacy behaviour is preserved as much as possible.
8014 * In DSA_generate_parameters_ex, if the provided seed is too short,
8015 use a random seed, as already documented.
8017 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
8019 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
8021 * Alternate chains certificate forgery
8023 During certificate verification, OpenSSL will attempt to find an
8024 alternative certificate chain if the first attempt to build such a chain
8025 fails. An error in the implementation of this logic can mean that an
8026 attacker could cause certain checks on untrusted certificates to be
8027 bypassed, such as the CA flag, enabling them to use a valid leaf
8028 certificate to act as a CA and "issue" an invalid certificate.
8030 This issue was reported to OpenSSL by Adam Langley/David Benjamin
8036 * Race condition handling PSK identify hint
8038 If PSK identity hints are received by a multi-threaded client then
8039 the values are wrongly updated in the parent SSL_CTX structure. This can
8040 result in a race condition potentially leading to a double free of the
8046 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
8048 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
8049 incompatibility in the handling of HMAC. The previous ABI has now been
8052 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
8054 * Malformed ECParameters causes infinite loop
8056 When processing an ECParameters structure OpenSSL enters an infinite loop
8057 if the curve specified is over a specially malformed binary polynomial
8060 This can be used to perform denial of service against any
8061 system which processes public keys, certificate requests or
8062 certificates. This includes TLS clients and TLS servers with
8063 client authentication enabled.
8065 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8070 * Exploitable out-of-bounds read in X509_cmp_time
8072 X509_cmp_time does not properly check the length of the ASN1_TIME
8073 string and can read a few bytes out of bounds. In addition,
8074 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8077 An attacker can use this to craft malformed certificates and CRLs of
8078 various sizes and potentially cause a segmentation fault, resulting in
8079 a DoS on applications that verify certificates or CRLs. TLS clients
8080 that verify CRLs are affected. TLS clients and servers with client
8081 authentication enabled may be affected if they use custom verification
8084 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8085 independently by Hanno Böck.
8090 * PKCS7 crash with missing EnvelopedContent
8092 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8093 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8094 with missing content and trigger a NULL pointer dereference on parsing.
8096 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8097 structures from untrusted sources are affected. OpenSSL clients and
8098 servers are not affected.
8100 This issue was reported to OpenSSL by Michal Zalewski (Google).
8105 * CMS verify infinite loop with unknown hash function
8107 When verifying a signedData message the CMS code can enter an infinite loop
8108 if presented with an unknown hash function OID. This can be used to perform
8109 denial of service against any system which verifies signedData messages using
8111 This issue was reported to OpenSSL by Johannes Bauer.
8116 * Race condition handling NewSessionTicket
8118 If a NewSessionTicket is received by a multi-threaded client when attempting to
8119 reuse a previous ticket then a race condition can occur potentially leading to
8120 a double free of the ticket data.
8125 * Reject DH handshakes with parameters shorter than 768 bits.
8127 *Kurt Roeckx and Emilia Kasper*
8129 * dhparam: generate 2048-bit parameters by default.
8131 *Kurt Roeckx and Emilia Kasper*
8133 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8135 * Segmentation fault in ASN1_TYPE_cmp fix
8137 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8138 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8139 certificate signature algorithm consistency this can be used to crash any
8140 certificate verification operation and exploited in a DoS attack. Any
8141 application which performs certificate verification is vulnerable including
8142 OpenSSL clients and servers which enable client authentication.
8147 * ASN.1 structure reuse memory corruption fix
8149 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8150 memory corruption via an invalid write. Such reuse is and has been
8151 strongly discouraged and is believed to be rare.
8153 Applications that parse structures containing CHOICE or ANY DEFINED BY
8154 components may be affected. Certificate parsing (d2i_X509 and related
8155 functions) are however not affected. OpenSSL clients and servers are
8161 * PKCS7 NULL pointer dereferences fix
8163 The PKCS#7 parsing code does not handle missing outer ContentInfo
8164 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8165 missing content and trigger a NULL pointer dereference on parsing.
8167 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8168 otherwise parse PKCS#7 structures from untrusted sources are
8169 affected. OpenSSL clients and servers are not affected.
8171 This issue was reported to OpenSSL by Michal Zalewski (Google).
8176 * DoS via reachable assert in SSLv2 servers fix
8178 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8179 servers that both support SSLv2 and enable export cipher suites by sending
8180 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8182 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8183 (OpenSSL development team).
8188 * Use After Free following d2i_ECPrivatekey error fix
8190 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8191 could cause a use after free condition. This, in turn, could cause a double
8192 free in several private key parsing functions (such as d2i_PrivateKey
8193 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8194 for applications that receive EC private keys from untrusted
8195 sources. This scenario is considered rare.
8197 This issue was discovered by the BoringSSL project and fixed in their
8203 * X509_to_X509_REQ NULL pointer deref fix
8205 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8206 the certificate key is invalid. This function is rarely used in practice.
8208 This issue was discovered by Brian Carpenter.
8213 * Removed the export ciphers from the DEFAULT ciphers
8217 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8219 * Build fixes for the Windows and OpenVMS platforms
8221 *Matt Caswell and Richard Levitte*
8223 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8225 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8226 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8227 dereference. This could lead to a Denial Of Service attack. Thanks to
8228 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8233 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8234 dtls1_buffer_record function under certain conditions. In particular this
8235 could occur if an attacker sent repeated DTLS records with the same
8236 sequence number but for the next epoch. The memory leak could be exploited
8237 by an attacker in a Denial of Service attack through memory exhaustion.
8238 Thanks to Chris Mueller for reporting this issue.
8243 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8244 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8245 method would be set to NULL which could later result in a NULL pointer
8246 dereference. Thanks to Frank Schmirler for reporting this issue.
8251 * Abort handshake if server key exchange message is omitted for ephemeral
8254 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8255 reporting this issue.
8260 * Remove non-export ephemeral RSA code on client and server. This code
8261 violated the TLS standard by allowing the use of temporary RSA keys in
8262 non-export ciphersuites and could be used by a server to effectively
8263 downgrade the RSA key length used to a value smaller than the server
8264 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8265 INRIA or reporting this issue.
8270 * Fixed issue where DH client certificates are accepted without verification.
8271 An OpenSSL server will accept a DH certificate for client authentication
8272 without the certificate verify message. This effectively allows a client to
8273 authenticate without the use of a private key. This only affects servers
8274 which trust a client certificate authority which issues certificates
8275 containing DH keys: these are extremely rare and hardly ever encountered.
8276 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8282 * Ensure that the session ID context of an SSL is updated when its
8283 SSL_CTX is updated via SSL_set_SSL_CTX.
8285 The session ID context is typically set from the parent SSL_CTX,
8286 and can vary with the CTX.
8290 * Fix various certificate fingerprint issues.
8292 By using non-DER or invalid encodings outside the signed portion of a
8293 certificate the fingerprint can be changed without breaking the signature.
8294 Although no details of the signed portion of the certificate can be changed
8295 this can cause problems with some applications: e.g. those using the
8296 certificate fingerprint for blacklists.
8298 1. Reject signatures with non zero unused bits.
8300 If the BIT STRING containing the signature has non zero unused bits reject
8301 the signature. All current signature algorithms require zero unused bits.
8303 2. Check certificate algorithm consistency.
8305 Check the AlgorithmIdentifier inside TBS matches the one in the
8306 certificate signature. NB: this will result in signature failure
8307 errors for some broken certificates.
8309 Thanks to Konrad Kraszewski from Google for reporting this issue.
8311 3. Check DSA/ECDSA signatures use DER.
8313 Re-encode DSA/ECDSA signatures and compare with the original received
8314 signature. Return an error if there is a mismatch.
8316 This will reject various cases including garbage after signature
8317 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8318 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8319 (negative or with leading zeroes).
8321 Further analysis was conducted and fixes were developed by Stephen Henson
8322 of the OpenSSL core team.
8328 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8329 results on some platforms, including x86_64. This bug occurs at random
8330 with a very low probability, and is not known to be exploitable in any
8331 way, though its exact impact is difficult to determine. Thanks to Pieter
8332 Wuille (Blockstream) who reported this issue and also suggested an initial
8333 fix. Further analysis was conducted by the OpenSSL development team and
8334 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8335 the OpenSSL core team.
8340 * Do not resume sessions on the server if the negotiated protocol
8341 version does not match the session's version. Resuming with a different
8342 version, while not strictly forbidden by the RFC, is of questionable
8343 sanity and breaks all known clients.
8345 *David Benjamin, Emilia Käsper*
8347 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8348 early CCS messages during renegotiation. (Note that because
8349 renegotiation is encrypted, this early CCS was not exploitable.)
8353 * Tighten client-side session ticket handling during renegotiation:
8354 ensure that the client only accepts a session ticket if the server sends
8355 the extension anew in the ServerHello. Previously, a TLS client would
8356 reuse the old extension state and thus accept a session ticket if one was
8357 announced in the initial ServerHello.
8359 Similarly, ensure that the client requires a session ticket if one
8360 was advertised in the ServerHello. Previously, a TLS client would
8361 ignore a missing NewSessionTicket message.
8365 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8369 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8370 sends a carefully crafted handshake message, to cause OpenSSL to fail
8371 to free up to 64k of memory causing a memory leak. This could be
8372 exploited in a Denial Of Service attack. This issue affects OpenSSL
8373 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8374 whether SRTP is used or configured. Implementations of OpenSSL that
8375 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8377 The fix was developed by the OpenSSL team.
8382 * Session Ticket Memory Leak.
8384 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8385 integrity of that ticket is first verified. In the event of a session
8386 ticket integrity check failing, OpenSSL will fail to free memory
8387 causing a memory leak. By sending a large number of invalid session
8388 tickets an attacker could exploit this issue in a Denial Of Service
8394 * Build option no-ssl3 is incomplete.
8396 When OpenSSL is configured with "no-ssl3" as a build option, servers
8397 could accept and complete an SSL 3.0 handshake, and clients could be
8398 configured to send them.
8401 *Akamai and the OpenSSL team*
8403 * Add support for TLS_FALLBACK_SCSV.
8404 Client applications doing fallback retries should call
8405 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8408 *Adam Langley, Bodo Moeller*
8410 * Add additional DigestInfo checks.
8412 Re-encode DigestInto in DER and check against the original when
8413 verifying RSA signature: this will reject any improperly encoded
8414 DigestInfo structures.
8416 Note: this is a precautionary measure and no attacks are currently known.
8420 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8422 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8423 SRP code can be overrun an internal buffer. Add sanity check that
8424 g, A, B < N to SRP code.
8426 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8427 Group for discovering this issue.
8432 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8433 TLS 1.0 instead of higher protocol versions when the ClientHello message
8434 is badly fragmented. This allows a man-in-the-middle attacker to force a
8435 downgrade to TLS 1.0 even if both the server and the client support a
8436 higher protocol version, by modifying the client's TLS records.
8438 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8439 researching this issue.
8444 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8445 to a denial of service attack. A malicious server can crash the client
8446 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8447 ciphersuite and sending carefully crafted handshake messages.
8449 Thanks to Felix Gröbert (Google) for discovering and researching this
8455 * By sending carefully crafted DTLS packets an attacker could cause openssl
8456 to leak memory. This can be exploited through a Denial of Service attack.
8457 Thanks to Adam Langley for discovering and researching this issue.
8462 * An attacker can force openssl to consume large amounts of memory whilst
8463 processing DTLS handshake messages. This can be exploited through a
8464 Denial of Service attack.
8465 Thanks to Adam Langley for discovering and researching this issue.
8470 * An attacker can force an error condition which causes openssl to crash
8471 whilst processing DTLS packets due to memory being freed twice. This
8472 can be exploited through a Denial of Service attack.
8473 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8479 * If a multithreaded client connects to a malicious server using a resumed
8480 session and the server sends an ec point format extension it could write
8481 up to 255 bytes to freed memory.
8483 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8489 * A malicious server can crash an OpenSSL client with a null pointer
8490 dereference (read) by specifying an SRP ciphersuite even though it was not
8491 properly negotiated with the client. This can be exploited through a
8492 Denial of Service attack.
8494 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8495 discovering and researching this issue.
8500 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8501 X509_name_oneline, X509_name_print_ex et al. to leak some information
8502 from the stack. Applications may be affected if they echo pretty printing
8503 output to the attacker.
8505 Thanks to Ivan Fratric (Google) for discovering this issue.
8508 *Emilia Käsper, and Steve Henson*
8510 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8511 for corner cases. (Certain input points at infinity could lead to
8512 bogus results, with non-infinity inputs mapped to infinity too.)
8516 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8518 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8519 handshake can force the use of weak keying material in OpenSSL
8520 SSL/TLS clients and servers.
8522 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8523 researching this issue. ([CVE-2014-0224])
8525 *KIKUCHI Masashi, Steve Henson*
8527 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8528 OpenSSL DTLS client the code can be made to recurse eventually crashing
8531 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8534 *Imre Rad, Steve Henson*
8536 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8537 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8538 client or server. This is potentially exploitable to run arbitrary
8539 code on a vulnerable client or server.
8541 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8543 *Jüri Aedla, Steve Henson*
8545 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8546 are subject to a denial of service attack.
8548 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8549 this issue. ([CVE-2014-3470])
8551 *Felix Gröbert, Ivan Fratric, Steve Henson*
8553 * Harmonize version and its documentation. -f flag is used to display
8556 *mancha <mancha1@zoho.com>*
8558 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8559 in i2d_ECPrivateKey.
8561 *mancha <mancha1@zoho.com>*
8563 * Fix some double frees. These are not thought to be exploitable.
8565 *mancha <mancha1@zoho.com>*
8567 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8569 * A missing bounds check in the handling of the TLS heartbeat extension
8570 can be used to reveal up to 64k of memory to a connected client or
8573 Thanks for Neel Mehta of Google Security for discovering this bug and to
8574 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8575 preparing the fix ([CVE-2014-0160])
8577 *Adam Langley, Bodo Moeller*
8579 * Fix for the attack described in the paper "Recovering OpenSSL
8580 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8581 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8582 <http://eprint.iacr.org/2014/140>
8584 Thanks to Yuval Yarom and Naomi Benger for discovering this
8585 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8587 *Yuval Yarom and Naomi Benger*
8589 * TLS pad extension: draft-agl-tls-padding-03
8591 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8592 TLS client Hello record length value would otherwise be > 255 and
8593 less that 512 pad with a dummy extension containing zeroes so it
8594 is at least 512 bytes long.
8596 *Adam Langley, Steve Henson*
8598 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8600 * Fix for TLS record tampering bug. A carefully crafted invalid
8601 handshake could crash OpenSSL with a NULL pointer exception.
8602 Thanks to Anton Johansson for reporting this issues.
8605 * Keep original DTLS digest and encryption contexts in retransmission
8606 structures so we can use the previous session parameters if they need
8607 to be resent. ([CVE-2013-6450])
8611 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8612 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8613 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8614 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8615 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8616 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8618 *Rob Stradling, Adam Langley*
8620 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8622 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8623 supporting platforms or when small records were transferred.
8625 *Andy Polyakov, Steve Henson*
8627 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8629 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8631 This addresses the flaw in CBC record processing discovered by
8632 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8633 at: <http://www.isg.rhul.ac.uk/tls/>
8635 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8636 Security Group at Royal Holloway, University of London
8637 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8638 Emilia Käsper for the initial patch.
8641 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8643 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8644 ciphersuites which can be exploited in a denial of service attack.
8645 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8646 and detecting this bug and to Wolfgang Ettlinger
8647 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8652 * Return an error when checking OCSP signatures when key is NULL.
8653 This fixes a DoS attack. ([CVE-2013-0166])
8657 * Make openssl verify return errors.
8659 *Chris Palmer <palmer@google.com> and Ben Laurie*
8661 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8662 the right response is stapled. Also change SSL_get_certificate()
8663 so it returns the certificate actually sent.
8664 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8666 *Rob Stradling <rob.stradling@comodo.com>*
8668 * Fix possible deadlock when decoding public keys.
8672 * Don't use TLS 1.0 record version number in initial client hello
8677 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8679 * Sanity check record length before skipping explicit IV in TLS
8680 1.2, 1.1 and DTLS to fix DoS attack.
8682 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8683 fuzzing as a service testing platform.
8688 * Initialise tkeylen properly when encrypting CMS messages.
8689 Thanks to Solar Designer of Openwall for reporting this issue.
8693 * In FIPS mode don't try to use composite ciphers as they are not
8698 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8700 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8701 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8702 mean any application compiled against OpenSSL 1.0.0 headers setting
8703 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8704 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8705 0x10000000L Any application which was previously compiled against
8706 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8707 will need to be recompiled as a result. Letting be results in
8708 inability to disable specifically TLS 1.1 and in client context,
8709 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8713 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8714 disable just protocol X, but all protocols above X *if* there are
8715 protocols *below* X still enabled. In more practical terms it means
8716 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8717 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8718 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8723 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8725 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8726 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8727 in CRYPTO_realloc_clean.
8729 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8730 issue and to Adam Langley <agl@chromium.org> for fixing it.
8733 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8735 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8739 * Workarounds for some broken servers that "hang" if a client hello
8740 record length exceeds 255 bytes.
8742 1. Do not use record version number > TLS 1.0 in initial client
8743 hello: some (but not all) hanging servers will now work.
8744 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8745 the number of ciphers sent in the client hello. This should be
8746 set to an even number, such as 50, for example by passing:
8747 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8748 Most broken servers should now work.
8749 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8750 TLS 1.2 client support entirely.
8754 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8758 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8760 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8761 STRING form instead of a DigestInfo.
8765 * The format used for MDC2 RSA signatures is inconsistent between EVP
8766 and the RSA_sign/RSA_verify functions. This was made more apparent when
8767 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8768 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8769 the correct format in RSA_verify so both forms transparently work.
8773 * Some servers which support TLS 1.0 can choke if we initially indicate
8774 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8775 encrypted premaster secret. As a workaround use the maximum permitted
8776 client version in client hello, this should keep such servers happy
8777 and still work with previous versions of OpenSSL.
8781 * Add support for TLS/DTLS heartbeats.
8783 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8785 * Add support for SCTP.
8787 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8789 * Improved PRNG seeding for VOS.
8791 *Paul Green <Paul.Green@stratus.com>*
8793 * Extensive assembler packs updates, most notably:
8795 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8796 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8797 - x86_64: bit-sliced AES implementation;
8798 - ARM: NEON support, contemporary platforms optimizations;
8799 - s390x: z196 support;
8800 - `*`: GHASH and GF(2^m) multiplication implementations;
8804 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8805 (removal of unnecessary code)
8807 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8809 * Add TLS key material exporter from RFC 5705.
8813 * Add DTLS-SRTP negotiation from RFC 5764.
8817 * Add Next Protocol Negotiation,
8818 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8819 disabled with a no-npn flag to config or Configure. Code donated
8822 *Adam Langley <agl@google.com> and Ben Laurie*
8824 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8825 NIST-P256, NIST-P521, with constant-time single point multiplication on
8826 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8827 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8828 Code made available under Apache License version 2.0.
8830 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8831 line to include this in your build of OpenSSL, and run "make depend" (or
8832 "make update"). This enables the following EC_METHODs:
8834 EC_GFp_nistp224_method()
8835 EC_GFp_nistp256_method()
8836 EC_GFp_nistp521_method()
8838 EC_GROUP_new_by_curve_name() will automatically use these (while
8839 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8842 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8844 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8845 all platforms. Move ssize_t definition from e_os.h to the public
8846 header file e_os2.h as it now appears in public header file cms.h
8850 * New -sigopt option to the ca, req and x509 utilities. Additional
8851 signature parameters can be passed using this option and in
8856 * Add RSA PSS signing function. This will generate and set the
8857 appropriate AlgorithmIdentifiers for PSS based on those in the
8858 corresponding EVP_MD_CTX structure. No application support yet.
8862 * Support for companion algorithm specific ASN1 signing routines.
8863 New function ASN1_item_sign_ctx() signs a pre-initialised
8864 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8865 the appropriate parameters.
8869 * Add new algorithm specific ASN1 verification initialisation function
8870 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8871 handling will be the same no matter what EVP_PKEY_METHOD is used.
8872 Add a PSS handler to support verification of PSS signatures: checked
8873 against a number of sample certificates.
8877 * Add signature printing for PSS. Add PSS OIDs.
8879 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8881 * Add algorithm specific signature printing. An individual ASN1 method
8882 can now print out signatures instead of the standard hex dump.
8884 More complex signatures (e.g. PSS) can print out more meaningful
8885 information. Include DSA version that prints out the signature
8890 * Password based recipient info support for CMS library: implementing
8895 * Split password based encryption into PBES2 and PBKDF2 functions. This
8896 neatly separates the code into cipher and PBE sections and is required
8897 for some algorithms that split PBES2 into separate pieces (such as
8898 password based CMS).
8902 * Session-handling fixes:
8903 - Fix handling of connections that are resuming with a session ID,
8904 but also support Session Tickets.
8905 - Fix a bug that suppressed issuing of a new ticket if the client
8906 presented a ticket with an expired session.
8907 - Try to set the ticket lifetime hint to something reasonable.
8908 - Make tickets shorter by excluding irrelevant information.
8909 - On the client side, don't ignore renewed tickets.
8911 *Adam Langley, Bodo Moeller (Google)*
8913 * Fix PSK session representation.
8917 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8919 This work was sponsored by Intel.
8923 * Add GCM support to TLS library. Some custom code is needed to split
8924 the IV between the fixed (from PRF) and explicit (from TLS record)
8925 portions. This adds all GCM ciphersuites supported by RFC5288 and
8926 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8927 add a special AESGCM string for GCM only.
8931 * Expand range of ctrls for AES GCM. Permit setting invocation
8932 field on decrypt and retrieval of invocation field only on encrypt.
8936 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8937 As required by RFC5289 these ciphersuites cannot be used if for
8938 versions of TLS earlier than 1.2.
8942 * For FIPS capable OpenSSL interpret a NULL default public key method
8943 as unset and return the appropriate default but do *not* set the default.
8944 This means we can return the appropriate method in applications that
8945 switch between FIPS and non-FIPS modes.
8949 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8950 ENGINE is used then we cannot handle that in the FIPS module so we
8951 keep original code iff non-FIPS operations are allowed.
8955 * Add -attime option to openssl utilities.
8957 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8959 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8963 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8964 FIPS EC methods unconditionally for now.
8968 * New build option no-ec2m to disable characteristic 2 code.
8972 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8973 all cases can be covered as some introduce binary incompatibilities.
8977 * Redirect RSA operations to FIPS module including keygen,
8978 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8982 * Add similar low-level API blocking to ciphers.
8986 * low-level digest APIs are not approved in FIPS mode: any attempt
8987 to use these will cause a fatal error. Applications that *really* want
8988 to use them can use the `private_*` version instead.
8992 * Redirect cipher operations to FIPS module for FIPS builds.
8996 * Redirect digest operations to FIPS module for FIPS builds.
9000 * Update build system to add "fips" flag which will link in fipscanister.o
9001 for static and shared library builds embedding a signature if needed.
9005 * Output TLS supported curves in preference order instead of numerical
9006 order. This is currently hardcoded for the highest order curves first.
9007 This should be configurable so applications can judge speed vs strength.
9011 * Add TLS v1.2 server support for client authentication.
9015 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
9020 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
9021 FIPS modules versions.
9025 * Add TLS v1.2 client side support for client authentication. Keep cache
9026 of handshake records longer as we don't know the hash algorithm to use
9027 until after the certificate request message is received.
9031 * Initial TLS v1.2 client support. Add a default signature algorithms
9032 extension including all the algorithms we support. Parse new signature
9033 format in client key exchange. Relax some ECC signing restrictions for
9034 TLS v1.2 as indicated in RFC5246.
9038 * Add server support for TLS v1.2 signature algorithms extension. Switch
9039 to new signature format when needed using client digest preference.
9040 All server ciphersuites should now work correctly in TLS v1.2. No client
9041 support yet and no support for client certificates.
9045 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
9046 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
9047 ciphersuites. At present only RSA key exchange ciphersuites work with
9048 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
9049 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
9050 and version checking.
9054 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
9055 with this defined it will not be affected by any changes to ssl internal
9056 structures. Add several utility functions to allow openssl application
9057 to work with OPENSSL_NO_SSL_INTERN defined.
9061 * A long standing patch to add support for SRP from EdelWeb (Peter
9062 Sylvester and Christophe Renou) was integrated.
9063 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
9064 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
9067 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
9071 * Permit abbreviated handshakes when renegotiating using the function
9072 SSL_renegotiate_abbreviated().
9074 *Robin Seggelmann <seggelmann@fh-muenster.de>*
9076 * Add call to ENGINE_register_all_complete() to
9077 ENGINE_load_builtin_engines(), so some implementations get used
9078 automatically instead of needing explicit application support.
9082 * Add support for TLS key exporter as described in RFC5705.
9084 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
9086 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
9087 a few changes are required:
9089 Add SSL_OP_NO_TLSv1_1 flag.
9090 Add TLSv1_1 methods.
9091 Update version checking logic to handle version 1.1.
9092 Add explicit IV handling (ported from DTLS code).
9093 Add command line options to s_client/s_server.
9100 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
9102 * X509_ATTRIBUTE memory leak
9104 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
9105 memory. This structure is used by the PKCS#7 and CMS routines so any
9106 application which reads PKCS#7 or CMS data from untrusted sources is
9107 affected. SSL/TLS is not affected.
9109 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
9115 * Race condition handling PSK identify hint
9117 If PSK identity hints are received by a multi-threaded client then
9118 the values are wrongly updated in the parent SSL_CTX structure. This can
9119 result in a race condition potentially leading to a double free of the
9125 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
9127 * Malformed ECParameters causes infinite loop
9129 When processing an ECParameters structure OpenSSL enters an infinite loop
9130 if the curve specified is over a specially malformed binary polynomial
9133 This can be used to perform denial of service against any
9134 system which processes public keys, certificate requests or
9135 certificates. This includes TLS clients and TLS servers with
9136 client authentication enabled.
9138 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9143 * Exploitable out-of-bounds read in X509_cmp_time
9145 X509_cmp_time does not properly check the length of the ASN1_TIME
9146 string and can read a few bytes out of bounds. In addition,
9147 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9150 An attacker can use this to craft malformed certificates and CRLs of
9151 various sizes and potentially cause a segmentation fault, resulting in
9152 a DoS on applications that verify certificates or CRLs. TLS clients
9153 that verify CRLs are affected. TLS clients and servers with client
9154 authentication enabled may be affected if they use custom verification
9157 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9158 independently by Hanno Böck.
9163 * PKCS7 crash with missing EnvelopedContent
9165 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9166 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9167 with missing content and trigger a NULL pointer dereference on parsing.
9169 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9170 structures from untrusted sources are affected. OpenSSL clients and
9171 servers are not affected.
9173 This issue was reported to OpenSSL by Michal Zalewski (Google).
9178 * CMS verify infinite loop with unknown hash function
9180 When verifying a signedData message the CMS code can enter an infinite loop
9181 if presented with an unknown hash function OID. This can be used to perform
9182 denial of service against any system which verifies signedData messages using
9184 This issue was reported to OpenSSL by Johannes Bauer.
9189 * Race condition handling NewSessionTicket
9191 If a NewSessionTicket is received by a multi-threaded client when attempting to
9192 reuse a previous ticket then a race condition can occur potentially leading to
9193 a double free of the ticket data.
9198 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9200 * Segmentation fault in ASN1_TYPE_cmp fix
9202 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9203 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9204 certificate signature algorithm consistency this can be used to crash any
9205 certificate verification operation and exploited in a DoS attack. Any
9206 application which performs certificate verification is vulnerable including
9207 OpenSSL clients and servers which enable client authentication.
9212 * ASN.1 structure reuse memory corruption fix
9214 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9215 memory corruption via an invalid write. Such reuse is and has been
9216 strongly discouraged and is believed to be rare.
9218 Applications that parse structures containing CHOICE or ANY DEFINED BY
9219 components may be affected. Certificate parsing (d2i_X509 and related
9220 functions) are however not affected. OpenSSL clients and servers are
9226 * PKCS7 NULL pointer dereferences fix
9228 The PKCS#7 parsing code does not handle missing outer ContentInfo
9229 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9230 missing content and trigger a NULL pointer dereference on parsing.
9232 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9233 otherwise parse PKCS#7 structures from untrusted sources are
9234 affected. OpenSSL clients and servers are not affected.
9236 This issue was reported to OpenSSL by Michal Zalewski (Google).
9241 * DoS via reachable assert in SSLv2 servers fix
9243 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9244 servers that both support SSLv2 and enable export cipher suites by sending
9245 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9247 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9248 (OpenSSL development team).
9253 * Use After Free following d2i_ECPrivatekey error fix
9255 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9256 could cause a use after free condition. This, in turn, could cause a double
9257 free in several private key parsing functions (such as d2i_PrivateKey
9258 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9259 for applications that receive EC private keys from untrusted
9260 sources. This scenario is considered rare.
9262 This issue was discovered by the BoringSSL project and fixed in their
9268 * X509_to_X509_REQ NULL pointer deref fix
9270 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9271 the certificate key is invalid. This function is rarely used in practice.
9273 This issue was discovered by Brian Carpenter.
9278 * Removed the export ciphers from the DEFAULT ciphers
9282 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9284 * Build fixes for the Windows and OpenVMS platforms
9286 *Matt Caswell and Richard Levitte*
9288 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9290 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9291 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9292 dereference. This could lead to a Denial Of Service attack. Thanks to
9293 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9298 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9299 dtls1_buffer_record function under certain conditions. In particular this
9300 could occur if an attacker sent repeated DTLS records with the same
9301 sequence number but for the next epoch. The memory leak could be exploited
9302 by an attacker in a Denial of Service attack through memory exhaustion.
9303 Thanks to Chris Mueller for reporting this issue.
9308 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9309 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9310 method would be set to NULL which could later result in a NULL pointer
9311 dereference. Thanks to Frank Schmirler for reporting this issue.
9316 * Abort handshake if server key exchange message is omitted for ephemeral
9319 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9320 reporting this issue.
9325 * Remove non-export ephemeral RSA code on client and server. This code
9326 violated the TLS standard by allowing the use of temporary RSA keys in
9327 non-export ciphersuites and could be used by a server to effectively
9328 downgrade the RSA key length used to a value smaller than the server
9329 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9330 INRIA or reporting this issue.
9335 * Fixed issue where DH client certificates are accepted without verification.
9336 An OpenSSL server will accept a DH certificate for client authentication
9337 without the certificate verify message. This effectively allows a client to
9338 authenticate without the use of a private key. This only affects servers
9339 which trust a client certificate authority which issues certificates
9340 containing DH keys: these are extremely rare and hardly ever encountered.
9341 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9347 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9348 results on some platforms, including x86_64. This bug occurs at random
9349 with a very low probability, and is not known to be exploitable in any
9350 way, though its exact impact is difficult to determine. Thanks to Pieter
9351 Wuille (Blockstream) who reported this issue and also suggested an initial
9352 fix. Further analysis was conducted by the OpenSSL development team and
9353 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9354 the OpenSSL core team.
9359 * Fix various certificate fingerprint issues.
9361 By using non-DER or invalid encodings outside the signed portion of a
9362 certificate the fingerprint can be changed without breaking the signature.
9363 Although no details of the signed portion of the certificate can be changed
9364 this can cause problems with some applications: e.g. those using the
9365 certificate fingerprint for blacklists.
9367 1. Reject signatures with non zero unused bits.
9369 If the BIT STRING containing the signature has non zero unused bits reject
9370 the signature. All current signature algorithms require zero unused bits.
9372 2. Check certificate algorithm consistency.
9374 Check the AlgorithmIdentifier inside TBS matches the one in the
9375 certificate signature. NB: this will result in signature failure
9376 errors for some broken certificates.
9378 Thanks to Konrad Kraszewski from Google for reporting this issue.
9380 3. Check DSA/ECDSA signatures use DER.
9382 Re-encode DSA/ECDSA signatures and compare with the original received
9383 signature. Return an error if there is a mismatch.
9385 This will reject various cases including garbage after signature
9386 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9387 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9388 (negative or with leading zeroes).
9390 Further analysis was conducted and fixes were developed by Stephen Henson
9391 of the OpenSSL core team.
9397 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9399 * Session Ticket Memory Leak.
9401 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9402 integrity of that ticket is first verified. In the event of a session
9403 ticket integrity check failing, OpenSSL will fail to free memory
9404 causing a memory leak. By sending a large number of invalid session
9405 tickets an attacker could exploit this issue in a Denial Of Service
9411 * Build option no-ssl3 is incomplete.
9413 When OpenSSL is configured with "no-ssl3" as a build option, servers
9414 could accept and complete an SSL 3.0 handshake, and clients could be
9415 configured to send them.
9418 *Akamai and the OpenSSL team*
9420 * Add support for TLS_FALLBACK_SCSV.
9421 Client applications doing fallback retries should call
9422 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9425 *Adam Langley, Bodo Moeller*
9427 * Add additional DigestInfo checks.
9429 Re-encode DigestInto in DER and check against the original when
9430 verifying RSA signature: this will reject any improperly encoded
9431 DigestInfo structures.
9433 Note: this is a precautionary measure and no attacks are currently known.
9437 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9439 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9440 to a denial of service attack. A malicious server can crash the client
9441 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9442 ciphersuite and sending carefully crafted handshake messages.
9444 Thanks to Felix Gröbert (Google) for discovering and researching this
9450 * By sending carefully crafted DTLS packets an attacker could cause openssl
9451 to leak memory. This can be exploited through a Denial of Service attack.
9452 Thanks to Adam Langley for discovering and researching this issue.
9457 * An attacker can force openssl to consume large amounts of memory whilst
9458 processing DTLS handshake messages. This can be exploited through a
9459 Denial of Service attack.
9460 Thanks to Adam Langley for discovering and researching this issue.
9465 * An attacker can force an error condition which causes openssl to crash
9466 whilst processing DTLS packets due to memory being freed twice. This
9467 can be exploited through a Denial of Service attack.
9468 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9474 * If a multithreaded client connects to a malicious server using a resumed
9475 session and the server sends an ec point format extension it could write
9476 up to 255 bytes to freed memory.
9478 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9484 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9485 X509_name_oneline, X509_name_print_ex et al. to leak some information
9486 from the stack. Applications may be affected if they echo pretty printing
9487 output to the attacker.
9489 Thanks to Ivan Fratric (Google) for discovering this issue.
9492 *Emilia Käsper, and Steve Henson*
9494 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9495 for corner cases. (Certain input points at infinity could lead to
9496 bogus results, with non-infinity inputs mapped to infinity too.)
9500 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9502 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9503 handshake can force the use of weak keying material in OpenSSL
9504 SSL/TLS clients and servers.
9506 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9507 researching this issue. ([CVE-2014-0224])
9509 *KIKUCHI Masashi, Steve Henson*
9511 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9512 OpenSSL DTLS client the code can be made to recurse eventually crashing
9515 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9518 *Imre Rad, Steve Henson*
9520 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9521 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9522 client or server. This is potentially exploitable to run arbitrary
9523 code on a vulnerable client or server.
9525 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9527 *Jüri Aedla, Steve Henson*
9529 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9530 are subject to a denial of service attack.
9532 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9533 this issue. ([CVE-2014-3470])
9535 *Felix Gröbert, Ivan Fratric, Steve Henson*
9537 * Harmonize version and its documentation. -f flag is used to display
9540 *mancha <mancha1@zoho.com>*
9542 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9543 in i2d_ECPrivateKey.
9545 *mancha <mancha1@zoho.com>*
9547 * Fix some double frees. These are not thought to be exploitable.
9549 *mancha <mancha1@zoho.com>*
9551 * Fix for the attack described in the paper "Recovering OpenSSL
9552 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9553 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9554 <http://eprint.iacr.org/2014/140>
9556 Thanks to Yuval Yarom and Naomi Benger for discovering this
9557 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9559 *Yuval Yarom and Naomi Benger*
9561 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9563 * Keep original DTLS digest and encryption contexts in retransmission
9564 structures so we can use the previous session parameters if they need
9565 to be resent. ([CVE-2013-6450])
9569 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9570 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9571 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9572 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9573 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9574 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9576 *Rob Stradling, Adam Langley*
9578 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9580 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9582 This addresses the flaw in CBC record processing discovered by
9583 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9584 at: <http://www.isg.rhul.ac.uk/tls/>
9586 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9587 Security Group at Royal Holloway, University of London
9588 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9589 Emilia Käsper for the initial patch.
9592 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9594 * Return an error when checking OCSP signatures when key is NULL.
9595 This fixes a DoS attack. ([CVE-2013-0166])
9599 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9600 the right response is stapled. Also change SSL_get_certificate()
9601 so it returns the certificate actually sent.
9602 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9603 (This is a backport)
9605 *Rob Stradling <rob.stradling@comodo.com>*
9607 * Fix possible deadlock when decoding public keys.
9611 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9613 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9616 * Sanity check record length before skipping explicit IV in DTLS
9619 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9620 fuzzing as a service testing platform.
9625 * Initialise tkeylen properly when encrypting CMS messages.
9626 Thanks to Solar Designer of Openwall for reporting this issue.
9630 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9632 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9633 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9634 in CRYPTO_realloc_clean.
9636 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9637 issue and to Adam Langley <agl@chromium.org> for fixing it.
9640 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9642 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9644 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9645 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9646 content decryption and always return the same error. Note: this attack
9647 needs on average 2^20 messages so it only affects automated senders. The
9648 old behaviour can be re-enabled in the CMS code by setting the
9649 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9650 an MMA defence is not necessary.
9651 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9652 this issue. ([CVE-2012-0884])
9656 * Fix CVE-2011-4619: make sure we really are receiving a
9657 client hello before rejecting multiple SGC restarts. Thanks to
9658 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9662 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9664 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9665 Thanks to Antonio Martin, Enterprise Secure Access Research and
9666 Development, Cisco Systems, Inc. for discovering this bug and
9667 preparing a fix. ([CVE-2012-0050])
9671 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9673 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9674 of the Vaudenay padding oracle attack on CBC mode encryption
9675 which enables an efficient plaintext recovery attack against
9676 the OpenSSL implementation of DTLS. Their attack exploits timing
9677 differences arising during decryption processing. A research
9678 paper describing this attack can be found at:
9679 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9680 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9681 Security Group at Royal Holloway, University of London
9682 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9683 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9684 for preparing the fix. ([CVE-2011-4108])
9686 *Robin Seggelmann, Michael Tuexen*
9688 * Clear bytes used for block padding of SSL 3.0 records.
9691 *Adam Langley (Google)*
9693 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9694 Kadianakis <desnacked@gmail.com> for discovering this issue and
9695 Adam Langley for preparing the fix. ([CVE-2011-4619])
9697 *Adam Langley (Google)*
9699 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9701 *Andrey Kulikov <amdeich@gmail.com>*
9703 * Prevent malformed RFC3779 data triggering an assertion failure.
9704 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9705 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9707 *Rob Austein <sra@hactrn.net>*
9709 * Improved PRNG seeding for VOS.
9711 *Paul Green <Paul.Green@stratus.com>*
9713 * Fix ssl_ciph.c set-up race.
9715 *Adam Langley (Google)*
9717 * Fix spurious failures in ecdsatest.c.
9719 *Emilia Käsper (Google)*
9721 * Fix the BIO_f_buffer() implementation (which was mixing different
9722 interpretations of the `..._len` fields).
9724 *Adam Langley (Google)*
9726 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9727 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9728 threads won't reuse the same blinding coefficients.
9730 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9731 lock to call BN_BLINDING_invert_ex, and avoids one use of
9732 BN_BLINDING_update for each BN_BLINDING structure (previously,
9733 the last update always remained unused).
9735 *Emilia Käsper (Google)*
9737 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9739 *Bob Buckholz (Google)*
9741 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9743 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9744 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9746 *Kaspar Brand <ossl@velox.ch>*
9748 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9749 for multi-threaded use of ECDH. ([CVE-2011-3210])
9751 *Adam Langley (Google)*
9753 * Fix x509_name_ex_d2i memory leak on bad inputs.
9757 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9758 signature public key algorithm by using OID xref utilities instead.
9759 Before this you could only use some ECC ciphersuites with SHA1 only.
9763 * Add protection against ECDSA timing attacks as mentioned in the paper
9764 by Billy Bob Brumley and Nicola Tuveri, see:
9765 <http://eprint.iacr.org/2011/232.pdf>
9767 *Billy Bob Brumley and Nicola Tuveri*
9769 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9771 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9773 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9775 * Fix bug in string printing code: if *any* escaping is enabled we must
9776 escape the escape character (backslash) or the resulting string is
9781 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9783 * Disable code workaround for ancient and obsolete Netscape browsers
9784 and servers: an attacker can use it in a ciphersuite downgrade attack.
9785 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9789 * Fixed J-PAKE implementation error, originally discovered by
9790 Sebastien Martini, further info and confirmation from Stefan
9791 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9795 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9797 * Fix extension code to avoid race conditions which can result in a buffer
9798 overrun vulnerability: resumed sessions must not be modified as they can
9799 be shared by multiple threads. CVE-2010-3864
9803 * Fix WIN32 build system to correctly link an ENGINE directory into
9808 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9810 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9813 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9815 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9817 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9818 context. The operation can be customised via the ctrl mechanism in
9819 case ENGINEs want to include additional functionality.
9823 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9827 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9828 output hashes compatible with older versions of OpenSSL.
9830 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9832 * Fix compression algorithm handling: if resuming a session use the
9833 compression algorithm of the resumed session instead of determining
9834 it from client hello again. Don't allow server to change algorithm.
9838 * Add load_crls() function to commands tidying load_certs() too. Add option
9839 to verify utility to allow additional CRLs to be included.
9843 * Update OCSP request code to permit adding custom headers to the request:
9844 some responders need this.
9848 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9851 *Julia Lawall <julia@diku.dk>*
9853 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9854 needlessly dereferenced structures, used obsolete functions and
9855 didn't handle all updated verify codes correctly.
9859 * Disable MD2 in the default configuration.
9863 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9864 indicate the initial BIO being pushed or popped. This makes it possible
9865 to determine whether the BIO is the one explicitly called or as a result
9866 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9867 it handles reference counts correctly and doesn't zero out the I/O bio
9868 when it is not being explicitly popped. WARNING: applications which
9869 included workarounds for the old buggy behaviour will need to be modified
9870 or they could free up already freed BIOs.
9874 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9875 renaming to all platforms (within the 0.9.8 branch, this was
9876 done conditionally on Netware platforms to avoid a name clash).
9878 *Guenter <lists@gknw.net>*
9880 * Add ECDHE and PSK support to DTLS.
9882 *Michael Tuexen <tuexen@fh-muenster.de>*
9884 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9889 * Add "missing" function EVP_MD_flags() (without this the only way to
9890 retrieve a digest flags is by accessing the structure directly. Update
9891 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9892 or cipher is registered as in the "from" argument. Print out all
9893 registered digests in the dgst usage message instead of manually
9894 attempting to work them out.
9898 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9899 this allows the use of compression and extensions. Change default cipher
9900 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9901 by default unless an application cipher string requests it.
9905 * Alter match criteria in PKCS12_parse(). It used to try to use local
9906 key ids to find matching certificates and keys but some PKCS#12 files
9907 don't follow the (somewhat unwritten) rules and this strategy fails.
9908 Now just gather all certificates together and the first private key
9909 then look for the first certificate that matches the key.
9913 * Support use of registered digest and cipher names for dgst and cipher
9914 commands instead of having to add each one as a special case. So now
9921 openssl dgst -sha256 foo
9923 and this works for ENGINE based algorithms too.
9927 * Update Gost ENGINE to support parameter files.
9929 *Victor B. Wagner <vitus@cryptocom.ru>*
9931 * Support GeneralizedTime in ca utility.
9933 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9935 * Enhance the hash format used for certificate directory links. The new
9936 form uses the canonical encoding (meaning equivalent names will work
9937 even if they aren't identical) and uses SHA1 instead of MD5. This form
9938 is incompatible with the older format and as a result c_rehash should
9939 be used to rebuild symbolic links.
9943 * Make PKCS#8 the default write format for private keys, replacing the
9944 traditional format. This form is standardised, more secure and doesn't
9945 include an implicit MD5 dependency.
9949 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9950 committed to OpenSSL should pass this lot as a minimum.
9954 * Add session ticket override functionality for use by EAP-FAST.
9956 *Jouni Malinen <j@w1.fi>*
9958 * Modify HMAC functions to return a value. Since these can be implemented
9959 in an ENGINE errors can occur.
9963 * Type-checked OBJ_bsearch_ex.
9967 * Type-checked OBJ_bsearch. Also some constification necessitated
9968 by type-checking. Still to come: TXT_DB, bsearch(?),
9969 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9974 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9975 seconds to a tm structure directly, instead of going through OS
9976 specific date routines. This avoids any issues with OS routines such
9977 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9978 and X509_time_adj_ex() to cover the extended range. The existing
9979 X509_time_adj() is still usable and will no longer have any date issues.
9983 * Delta CRL support. New use deltas option which will attempt to locate
9984 and search any appropriate delta CRLs available.
9986 This work was sponsored by Google.
9990 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9991 code and add additional score elements. Validate alternate CRL paths
9992 as part of the CRL checking and indicate a new error "CRL path validation
9993 error" in this case. Applications wanting additional details can use
9994 the verify callback and check the new "parent" field. If this is not
9995 NULL CRL path validation is taking place. Existing applications won't
9996 see this because it requires extended CRL support which is off by
9999 This work was sponsored by Google.
10003 * Support for freshest CRL extension.
10005 This work was sponsored by Google.
10009 * Initial indirect CRL support. Currently only supported in the CRLs
10010 passed directly and not via lookup. Process certificate issuer
10011 CRL entry extension and lookup CRL entries by bother issuer name
10012 and serial number. Check and process CRL issuer entry in IDP extension.
10014 This work was sponsored by Google.
10018 * Add support for distinct certificate and CRL paths. The CRL issuer
10019 certificate is validated separately in this case. Only enabled if
10020 an extended CRL support flag is set: this flag will enable additional
10021 CRL functionality in future.
10023 This work was sponsored by Google.
10027 * Add support for policy mappings extension.
10029 This work was sponsored by Google.
10033 * Fixes to pathlength constraint, self issued certificate handling,
10034 policy processing to align with RFC3280 and PKITS tests.
10036 This work was sponsored by Google.
10040 * Support for name constraints certificate extension. DN, email, DNS
10041 and URI types are currently supported.
10043 This work was sponsored by Google.
10047 * To cater for systems that provide a pointer-based thread ID rather
10048 than numeric, deprecate the current numeric thread ID mechanism and
10049 replace it with a structure and associated callback type. This
10050 mechanism allows a numeric "hash" to be extracted from a thread ID in
10051 either case, and on platforms where pointers are larger than 'long',
10052 mixing is done to help ensure the numeric 'hash' is usable even if it
10053 can't be guaranteed unique. The default mechanism is to use "&errno"
10054 as a pointer-based thread ID to distinguish between threads.
10056 Applications that want to provide their own thread IDs should now use
10057 CRYPTO_THREADID_set_callback() to register a callback that will call
10058 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
10060 Note that ERR_remove_state() is now deprecated, because it is tied
10061 to the assumption that thread IDs are numeric. ERR_remove_state(0)
10062 to free the current thread's error state should be replaced by
10063 ERR_remove_thread_state(NULL).
10065 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
10066 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
10067 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
10068 application was previously providing a numeric thread callback that
10069 was inappropriate for distinguishing threads, then uniqueness might
10070 have been obtained with &errno that happened immediately in the
10071 intermediate development versions of OpenSSL; this is no longer the
10072 case, the numeric thread callback will now override the automatic use
10075 *Geoff Thorpe, with help from Bodo Moeller*
10077 * Initial support for different CRL issuing certificates. This covers a
10078 simple case where the self issued certificates in the chain exist and
10079 the real CRL issuer is higher in the existing chain.
10081 This work was sponsored by Google.
10085 * Removed effectively defunct crypto/store from the build.
10089 * Revamp of STACK to provide stronger type-checking. Still to come:
10090 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
10091 ASN1_STRING, CONF_VALUE.
10095 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
10096 RAM on SSL connections. This option can save about 34k per idle SSL.
10100 * Revamp of LHASH to provide stronger type-checking. Still to come:
10101 STACK, TXT_DB, bsearch, qsort.
10105 * Initial support for Cryptographic Message Syntax (aka CMS) based
10106 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
10107 support for data, signedData, compressedData, digestedData and
10108 encryptedData, envelopedData types included. Scripts to check against
10109 RFC4134 examples draft and interop and consistency checks of many
10110 content types and variants.
10114 * Add options to enc utility to support use of zlib compression BIO.
10118 * Extend mk1mf to support importing of options and assembly language
10119 files from Configure script, currently only included in VC-WIN32.
10120 The assembly language rules can now optionally generate the source
10121 files from the associated perl scripts.
10125 * Implement remaining functionality needed to support GOST ciphersuites.
10126 Interop testing has been performed using CryptoPro implementations.
10128 *Victor B. Wagner <vitus@cryptocom.ru>*
10130 * s390x assembler pack.
10134 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10139 * Implement Opaque PRF Input TLS extension as specified in
10140 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10141 official specification yet and no extension type assignment by
10142 IANA exists, this extension (for now) will have to be explicitly
10143 enabled when building OpenSSL by providing the extension number
10144 to use. For example, specify an option
10146 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10148 to the "config" or "Configure" script to enable the extension,
10149 assuming extension number 0x9527 (which is a completely arbitrary
10150 and unofficial assignment based on the MD5 hash of the Internet
10151 Draft). Note that by doing so, you potentially lose
10152 interoperability with other TLS implementations since these might
10153 be using the same extension number for other purposes.
10155 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10156 opaque PRF input value to use in the handshake. This will create
10157 an internal copy of the length-'len' string at 'src', and will
10158 return non-zero for success.
10160 To get more control and flexibility, provide a callback function
10163 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10164 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10168 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10171 Callback function 'cb' will be called in handshakes, and is
10172 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10173 Argument 'arg' is for application purposes (the value as given to
10174 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10175 be provided to the callback function). The callback function
10176 has to return non-zero to report success: usually 1 to use opaque
10177 PRF input just if possible, or 2 to enforce use of the opaque PRF
10178 input. In the latter case, the library will abort the handshake
10179 if opaque PRF input is not successfully negotiated.
10181 Arguments 'peerinput' and 'len' given to the callback function
10182 will always be NULL and 0 in the case of a client. A server will
10183 see the client's opaque PRF input through these variables if
10184 available (NULL and 0 otherwise). Note that if the server
10185 provides an opaque PRF input, the length must be the same as the
10186 length of the client's opaque PRF input.
10188 Note that the callback function will only be called when creating
10189 a new session (session resumption can resume whatever was
10190 previously negotiated), and will not be called in SSL 2.0
10191 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10192 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10193 for applications that need to enforce opaque PRF input.
10197 * Update ssl code to support digests other than SHA1+MD5 for handshake
10200 *Victor B. Wagner <vitus@cryptocom.ru>*
10202 * Add RFC4507 support to OpenSSL. This includes the corrections in
10203 RFC4507bis. The encrypted ticket format is an encrypted encoded
10204 SSL_SESSION structure, that way new session features are automatically
10207 If a client application caches session in an SSL_SESSION structure
10208 support is transparent because tickets are now stored in the encoded
10211 The SSL_CTX structure automatically generates keys for ticket
10212 protection in servers so again support should be possible
10213 with no application modification.
10215 If a client or server wishes to disable RFC4507 support then the option
10216 SSL_OP_NO_TICKET can be set.
10218 Add a TLS extension debugging callback to allow the contents of any client
10219 or server extensions to be examined.
10221 This work was sponsored by Google.
10225 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10226 OpenSSL should now compile cleanly on gcc 4.2
10228 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10230 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10231 support including streaming MAC support: this is required for GOST
10232 ciphersuite support.
10234 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10236 * Add option -stream to use PKCS#7 streaming in smime utility. New
10237 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10238 to output in BER and PEM format.
10242 * Experimental support for use of HMAC via EVP_PKEY interface. This
10243 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10244 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10245 ENGINE support for HMAC keys which are unextractable. New -mac and
10246 -macopt options to dgst utility.
10250 * New option -sigopt to dgst utility. Update dgst to use
10251 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10252 alternative signing parameters such as X9.31 or PSS in the dgst
10257 * Change ssl_cipher_apply_rule(), the internal function that does
10258 the work each time a ciphersuite string requests enabling
10259 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10260 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10261 the order of disabled ciphersuites such that those ciphersuites
10262 that most recently went from enabled to disabled not only stay
10263 in order with respect to each other, but also have higher priority
10264 than other disabled ciphersuites the next time ciphersuites are
10267 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10268 the same ciphersuites as with "HIGH" alone, but in a specific
10269 order where the PSK ciphersuites come first (since they are the
10270 most recently disabled ciphersuites when "HIGH" is parsed).
10272 Also, change ssl_create_cipher_list() (using this new
10273 functionality) such that between otherwise identical
10274 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10279 * Change ssl_create_cipher_list() so that it automatically
10280 arranges the ciphersuites in reasonable order before starting
10281 to process the rule string. Thus, the definition for "DEFAULT"
10282 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10283 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10284 This makes it much easier to arrive at a reasonable default order
10285 in applications for which anonymous ciphers are OK (meaning
10286 that you can't actually use DEFAULT).
10288 *Bodo Moeller; suggested by Victor Duchovni*
10290 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10291 processing) into multiple integers instead of setting
10292 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10293 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10294 (These masks as well as the individual bit definitions are hidden
10295 away into the non-exported interface ssl/ssl_locl.h, so this
10296 change to the definition of the SSL_CIPHER structure shouldn't
10297 affect applications.) This give us more bits for each of these
10298 categories, so there is no longer a need to coagulate AES128 and
10299 AES256 into a single algorithm bit, and to coagulate Camellia128
10300 and Camellia256 into a single algorithm bit, which has led to all
10303 Thus, among other things, the kludge introduced in 0.9.7m and
10304 0.9.8e for masking out AES256 independently of AES128 or masking
10305 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10307 With the change, we also introduce new ciphersuite aliases that
10308 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10313 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10314 Use the leftmost N bytes of the signature input if the input is
10315 larger than the prime q (with N being the size in bytes of q).
10319 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10320 it yet and it is largely untested.
10324 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10328 * Initial incomplete changes to avoid need for function casts in OpenSSL
10329 some compilers (gcc 4.2 and later) reject their use. Safestack is
10330 reimplemented. Update ASN1 to avoid use of legacy functions.
10334 * Win32/64 targets are linked with Winsock2.
10338 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10339 to external functions. This can be used to increase CRL handling
10340 efficiency especially when CRLs are very large by (for example) storing
10341 the CRL revoked certificates in a database.
10345 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10346 new CRLs added to a directory can be used. New command line option
10347 -verify_return_error to s_client and s_server. This causes real errors
10348 to be returned by the verify callback instead of carrying on no matter
10349 what. This reflects the way a "real world" verify callback would behave.
10353 * GOST engine, supporting several GOST algorithms and public key formats.
10354 Kindly donated by Cryptocom.
10358 * Partial support for Issuing Distribution Point CRL extension. CRLs
10359 partitioned by DP are handled but no indirect CRL or reason partitioning
10360 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10361 selected via a scoring technique which handles IDP and AKID in CRLs.
10365 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10366 will ultimately be used for all verify operations: this will remove the
10367 X509_STORE dependency on certificate verification and allow alternative
10368 lookup methods. X509_STORE based implementations of these two callbacks.
10372 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10373 Modify get_crl() to find a valid (unexpired) CRL if possible.
10377 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10378 this would be called X509_CRL_cmp() but that name is already used by
10379 a function that just compares CRL issuer names. Cache several CRL
10380 extensions in X509_CRL structure and cache CRLDP in X509.
10384 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10385 this maps equivalent X509_NAME structures into a consistent structure.
10386 Name comparison can then be performed rapidly using memcmp().
10390 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10395 * Allow digests to supply their own micalg string for S/MIME type using
10396 the ctrl EVP_MD_CTRL_MICALG.
10400 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10401 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10402 ctrl. It can then customise the structure before and/or after signing
10407 * New function OBJ_add_sigid() to allow application defined signature OIDs
10408 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10409 to free up any added signature OIDs.
10413 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10414 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10415 digest and cipher tables. New options added to openssl utility:
10416 list-message-digest-algorithms and list-cipher-algorithms.
10420 * Change the array representation of binary polynomials: the list
10421 of degrees of non-zero coefficients is now terminated with -1.
10422 Previously it was terminated with 0, which was also part of the
10423 value; thus, the array representation was not applicable to
10424 polynomials where t^0 has coefficient zero. This change makes
10425 the array representation useful in a more general context.
10429 * Various modifications and fixes to SSL/TLS cipher string
10430 handling. For ECC, the code now distinguishes between fixed ECDH
10431 with RSA certificates on the one hand and with ECDSA certificates
10432 on the other hand, since these are separate ciphersuites. The
10433 unused code for Fortezza ciphersuites has been removed.
10435 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10436 (not "ECDHE"). For consistency with the code for DH
10437 certificates, use of ECDH certificates is now considered ECDH
10438 authentication, not RSA or ECDSA authentication (the latter is
10439 merely the CA's signing algorithm and not actively used in the
10442 The temporary ciphersuite alias "ECCdraft" is no longer
10443 available, and ECC ciphersuites are no longer excluded from "ALL"
10444 and "DEFAULT". The following aliases now exist for RFC 4492
10445 ciphersuites, most of these by analogy with the DH case:
10447 kECDHr - ECDH cert, signed with RSA
10448 kECDHe - ECDH cert, signed with ECDSA
10449 kECDH - ECDH cert (signed with either RSA or ECDSA)
10450 kEECDH - ephemeral ECDH
10451 ECDH - ECDH cert or ephemeral ECDH
10454 aECDSA - ECDSA cert
10457 AECDH - anonymous ECDH
10458 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10462 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10463 Use correct micalg parameters depending on digest(s) in signed message.
10467 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10468 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10472 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10473 an engine to register a method. Add ENGINE lookups for methods and
10474 functional reference processing.
10478 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10479 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10484 * New -resign option to smime utility. This adds one or more signers
10485 to an existing PKCS#7 signedData structure. Also -md option to use an
10486 alternative message digest algorithm for signing.
10490 * Tidy up PKCS#7 routines and add new functions to make it easier to
10491 create PKCS7 structures containing multiple signers. Update smime
10492 application to support multiple signers.
10496 * New -macalg option to pkcs12 utility to allow setting of an alternative
10501 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10502 Reorganize PBE internals to lookup from a static table using NIDs,
10503 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10504 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10505 PRF which will be automatically used with PBES2.
10509 * Replace the algorithm specific calls to generate keys in "req" with the
10514 * Update PKCS#7 enveloped data routines to use new API. This is now
10515 supported by any public key method supporting the encrypt operation. A
10516 ctrl is added to allow the public key algorithm to examine or modify
10517 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10522 * Add a ctrl to asn1 method to allow a public key algorithm to express
10523 a default digest type to use. In most cases this will be SHA1 but some
10524 algorithms (such as GOST) need to specify an alternative digest. The
10525 return value indicates how strong the preference is 1 means optional and
10526 2 is mandatory (that is it is the only supported type). Modify
10527 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10528 use the default md. Update openssl utilities to use the default digest
10529 type for signing if it is not explicitly indicated.
10533 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10534 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10535 signing method from the key type. This effectively removes the link
10536 between digests and public key types.
10540 * Add an OID cross reference table and utility functions. Its purpose is to
10541 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10542 rsaEncryption. This will allow some of the algorithm specific hackery
10543 needed to use the correct OID to be removed.
10547 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10548 structures for PKCS7_sign(). They are now set up by the relevant public
10553 * Add provisional EC pkey method with support for ECDSA and ECDH.
10557 * Add support for key derivation (agreement) in the API, DH method and
10562 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10563 public and private key formats. As a side effect these add additional
10564 command line functionality not previously available: DSA signatures can be
10565 generated and verified using pkeyutl and DH key support and generation in
10572 *Oliver Tappe <zooey@hirschkaefer.de>*
10574 * New make target "install_html_docs" installs HTML renditions of the
10577 *Oliver Tappe <zooey@hirschkaefer.de>*
10579 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10580 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10581 support key and parameter generation and add initial key generation
10582 functionality for RSA.
10586 * Add functions for main EVP_PKEY_method operations. The undocumented
10587 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10588 `EVP_PKEY_{encrypt,decrypt}_old`.
10592 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10593 key API, doesn't do much yet.
10597 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10598 public key algorithms. New option to openssl utility:
10599 "list-public-key-algorithms" to print out info.
10603 * Implement the Supported Elliptic Curves Extension for
10604 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10608 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10609 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10613 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10614 utilities such as rsa, dsa, dsaparam etc except they process any key
10619 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10620 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10621 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10626 * Initial support for pluggable public key ASN1.
10627 De-spaghettify the public key ASN1 handling. Move public and private
10628 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10629 algorithm specific handling to a single module within the relevant
10630 algorithm directory. Add functions to allow (near) opaque processing
10631 of public and private key structures.
10635 * Implement the Supported Point Formats Extension for
10636 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10640 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10641 for the psk identity [hint] and the psk callback functions to the
10642 SSL_SESSION, SSL and SSL_CTX structure.
10645 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10649 SSL_CTX_use_psk_identity_hint
10650 SSL_get_psk_identity_hint
10651 SSL_get_psk_identity
10652 SSL_use_psk_identity_hint
10654 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10656 * Add RFC 3161 compliant time stamp request creation, response generation
10657 and response verification functionality.
10659 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10661 * Add initial support for TLS extensions, specifically for the server_name
10662 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10663 have new members for a hostname. The SSL data structure has an
10664 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10665 stored in that context to allow for session resumption, even after the
10666 SSL has been switched to a new SSL_CTX in reaction to a client's
10667 server_name extension.
10669 New functions (subject to change):
10671 SSL_get_servername()
10672 SSL_get_servername_type()
10675 New CTRL codes and macros (subject to change):
10677 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10678 - SSL_CTX_set_tlsext_servername_callback()
10679 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10680 - SSL_CTX_set_tlsext_servername_arg()
10681 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10683 openssl s_client has a new '-servername ...' option.
10685 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10686 '-key2 ...', '-servername_fatal' (subject to change). This allows
10687 testing the HostName extension for a specific single hostname ('-cert'
10688 and '-key' remain fallbacks for handshakes without HostName
10689 negotiation). If the unrecognized_name alert has to be sent, this by
10690 default is a warning; it becomes fatal with the '-servername_fatal'
10693 *Peter Sylvester, Remy Allais, Christophe Renou*
10695 * Whirlpool hash implementation is added.
10699 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10700 bn(64,32). Because of instruction set limitations it doesn't have
10701 any negative impact on performance. This was done mostly in order
10702 to make it possible to share assembler modules, such as bn_mul_mont
10703 implementations, between 32- and 64-bit builds without hassle.
10707 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10708 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10713 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10714 dedicated Montgomery multiplication procedure, is introduced.
10715 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10716 "64-bit" performance on certain 32-bit targets.
10720 * New option SSL_OP_NO_COMP to disable use of compression selectively
10721 in SSL structures. New SSL ctrl to set maximum send fragment size.
10722 Save memory by setting the I/O buffer sizes dynamically instead of
10723 using the maximum available value.
10727 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10728 in addition to the text details.
10732 * Very, very preliminary EXPERIMENTAL support for printing of general
10733 ASN1 structures. This currently produces rather ugly output and doesn't
10734 handle several customised structures at all.
10738 * Integrated support for PVK file format and some related formats such
10739 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10740 these in the 'rsa' and 'dsa' utilities.
10744 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10748 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10749 place for the (very old) "NETSCAPE" format certificates which are now
10750 handled using new ASN1 code equivalents.
10754 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10755 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10756 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10760 * Modify CRL distribution points extension code to print out previously
10761 unsupported fields. Enhance extension setting code to allow setting of
10766 * Add print and set support for Issuing Distribution Point CRL extension.
10770 * Change 'Configure' script to enable Camellia by default.
10777 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10779 * When rejecting SSL/TLS records due to an incorrect version number, never
10780 update s->server with a new major version number. As of
10781 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10782 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10783 the previous behavior could result in a read attempt at NULL when
10784 receiving specific incorrect SSL/TLS records once record payload
10785 protection is active. ([CVE-2010-0740])
10787 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10789 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10790 could be crashed if the relevant tables were not present (e.g. chrooted).
10792 *Tomas Hoger <thoger@redhat.com>*
10794 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10796 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10798 *Martin Olsson, Neel Mehta*
10800 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10801 accommodate for stack sorting, always a write lock!).
10805 * On some versions of WIN32 Heap32Next is very slow. This can cause
10806 excessive delays in the RAND_poll(): over a minute. As a workaround
10807 include a time check in the inner Heap32Next loop too.
10811 * The code that handled flushing of data in SSL/TLS originally used the
10812 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10813 the problem outlined in PR#1949. The fix suggested there however can
10814 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10815 of Apache). So instead simplify the code to flush unconditionally.
10816 This should be fine since flushing with no data to flush is a no op.
10820 * Handle TLS versions 2.0 and later properly and correctly use the
10821 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10822 off ancient servers have a habit of sticking around for a while...
10826 * Modify compression code so it frees up structures without using the
10827 ex_data callbacks. This works around a problem where some applications
10828 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10829 restarting) then use compression (e.g. SSL with compression) later.
10830 This results in significant per-connection memory leaks and
10831 has caused some security issues including CVE-2008-1678 and
10836 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10837 change when encrypting or decrypting.
10841 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10842 connect and renegotiate with servers which do not support RI.
10843 Until RI is more widely deployed this option is enabled by default.
10847 * Add "missing" ssl ctrls to clear options and mode.
10851 * If client attempts to renegotiate and doesn't support RI respond with
10852 a no_renegotiation alert as required by RFC5746. Some renegotiating
10853 TLS clients will continue a connection gracefully when they receive
10854 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10855 waiting for a server hello which it will never receive. Now we treat a
10856 received no_renegotiation alert as a fatal error. This is because
10857 applications requesting a renegotiation might well expect it to succeed
10858 and would have no code in place to handle the server denying it so the
10859 only safe thing to do is to terminate the connection.
10863 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10864 peer supports secure renegotiation and 0 otherwise. Print out peer
10865 renegotiation support in s_client/s_server.
10869 * Replace the highly broken and deprecated SPKAC certification method with
10870 the updated NID creation version. This should correctly handle UTF8.
10874 * Implement RFC5746. Re-enable renegotiation but require the extension
10875 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10876 turns out to be a bad idea. It has been replaced by
10877 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10878 SSL_CTX_set_options(). This is really not recommended unless you
10879 know what you are doing.
10881 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10883 * Fixes to stateless session resumption handling. Use initial_ctx when
10884 issuing and attempting to decrypt tickets in case it has changed during
10885 servername handling. Use a non-zero length session ID when attempting
10886 stateless session resumption: this makes it possible to determine if
10887 a resumption has occurred immediately after receiving server hello
10888 (several places in OpenSSL subtly assume this) instead of later in
10893 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10894 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10895 fixes for a few places where the return code is not checked
10898 *Julia Lawall <julia@diku.dk>*
10900 * Add --strict-warnings option to Configure script to include devteam
10901 warnings in other configurations.
10905 * Add support for --libdir option and LIBDIR variable in makefiles. This
10906 makes it possible to install openssl libraries in locations which
10907 have names other than "lib", for example "/usr/lib64" which some
10910 *Steve Henson, based on patch from Jeremy Utley*
10912 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10913 X690 8.9.12 and can produce some misleading textual output of OIDs.
10915 *Steve Henson, reported by Dan Kaminsky*
10917 * Delete MD2 from algorithm tables. This follows the recommendation in
10918 several standards that it is not used in new applications due to
10919 several cryptographic weaknesses. For binary compatibility reasons
10920 the MD2 API is still compiled in by default.
10924 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10929 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10930 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10933 *Guenter <lists@gknw.net>*
10935 * Fix the server certificate chain building code to use X509_verify_cert(),
10936 it used to have an ad-hoc builder which was unable to cope with anything
10937 other than a simple chain.
10939 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10941 * Don't check self signed certificate signatures in X509_verify_cert()
10942 by default (a flag can override this): it just wastes time without
10943 adding any security. As a useful side effect self signed root CAs
10944 with non-FIPS digests are now usable in FIPS mode.
10948 * In dtls1_process_out_of_seq_message() the check if the current message
10949 is already buffered was missing. For every new message was memory
10950 allocated, allowing an attacker to perform an denial of service attack
10951 with sending out of seq handshake messages until there is no memory
10952 left. Additionally every future message was buffered, even if the
10953 sequence number made no sense and would be part of another handshake.
10954 So only messages with sequence numbers less than 10 in advance will be
10955 buffered. ([CVE-2009-1378])
10957 *Robin Seggelmann, discovered by Daniel Mentz*
10959 * Records are buffered if they arrive with a future epoch to be
10960 processed after finishing the corresponding handshake. There is
10961 currently no limitation to this buffer allowing an attacker to perform
10962 a DOS attack with sending records with future epochs until there is no
10963 memory left. This patch adds the pqueue_size() function to determine
10964 the size of a buffer and limits the record buffer to 100 entries.
10967 *Robin Seggelmann, discovered by Daniel Mentz*
10969 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10970 parent structure is freed. ([CVE-2009-1379])
10974 * Handle non-blocking I/O properly in SSL_shutdown() call.
10976 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10978 * Add `2.5.4.*` OIDs
10980 *Ilya O. <vrghost@gmail.com>*
10982 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10984 * Disable renegotiation completely - this fixes a severe security
10985 problem ([CVE-2009-3555]) at the cost of breaking all
10986 renegotiation. Renegotiation can be re-enabled by setting
10987 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10988 run-time. This is really not recommended unless you know what
10993 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10995 * Don't set val to NULL when freeing up structures, it is freed up by
10996 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10997 zeroing past the valid field. ([CVE-2009-0789])
10999 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
11001 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
11002 checked correctly. This would allow some invalid signed attributes to
11003 appear to verify correctly. ([CVE-2009-0591])
11005 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11007 * Reject UniversalString and BMPString types with invalid lengths. This
11008 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
11009 a legal length. ([CVE-2009-0590])
11013 * Set S/MIME signing as the default purpose rather than setting it
11014 unconditionally. This allows applications to override it at the store
11019 * Permit restricted recursion of ASN1 strings. This is needed in practice
11020 to handle some structures.
11024 * Improve efficiency of mem_gets: don't search whole buffer each time
11027 *Jeremy Shapiro <jnshapir@us.ibm.com>*
11029 * New -hex option for openssl rand.
11033 * Print out UTF8String and NumericString when parsing ASN1.
11037 * Support NumericString type for name components.
11041 * Allow CC in the environment to override the automatically chosen
11042 compiler. Note that nothing is done to ensure flags work with the
11047 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
11049 * Properly check EVP_VerifyFinal() and similar return values
11052 *Ben Laurie, Bodo Moeller, Google Security Team*
11054 * Enable TLS extensions by default.
11058 * Allow the CHIL engine to be loaded, whether the application is
11059 multithreaded or not. (This does not release the developer from the
11060 obligation to set up the dynamic locking callbacks.)
11062 *Sander Temme <sander@temme.net>*
11064 * Use correct exit code if there is an error in dgst command.
11066 *Steve Henson; problem pointed out by Roland Dirlewanger*
11068 * Tweak Configure so that you need to say "experimental-jpake" to enable
11069 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
11073 * Add experimental JPAKE support, including demo authentication in
11074 s_client and s_server.
11078 * Set the comparison function in v3_addr_canonize().
11080 *Rob Austein <sra@hactrn.net>*
11082 * Add support for XMPP STARTTLS in s_client.
11084 *Philip Paeps <philip@freebsd.org>*
11086 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
11087 to ensure that even with this option, only ciphersuites in the
11088 server's preference list will be accepted. (Note that the option
11089 applies only when resuming a session, so the earlier behavior was
11090 just about the algorithm choice for symmetric cryptography.)
11094 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
11096 * Fix NULL pointer dereference if a DTLS server received
11097 ChangeCipherSpec as first record ([CVE-2009-1386]).
11101 * Fix a state transition in s3_srvr.c and d1_srvr.c
11102 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
11104 *Nagendra Modadugu*
11106 * The fix in 0.9.8c that supposedly got rid of unsafe
11107 double-checked locking was incomplete for RSA blinding,
11108 addressing just one layer of what turns out to have been
11109 doubly unsafe triple-checked locking.
11111 So now fix this for real by retiring the MONT_HELPER macro
11112 in crypto/rsa/rsa_eay.c.
11114 *Bodo Moeller; problem pointed out by Marius Schilder*
11116 * Various precautionary measures:
11118 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
11120 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
11121 (NB: This would require knowledge of the secret session ticket key
11122 to exploit, in which case you'd be SOL either way.)
11124 - Change bn_nist.c so that it will properly handle input BIGNUMs
11125 outside the expected range.
11127 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11130 *Neel Mehta, Bodo Moeller*
11132 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11133 the load fails. Useful for distros.
11135 *Ben Laurie and the FreeBSD team*
11137 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11141 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11145 * Expand ENGINE to support engine supplied SSL client certificate functions.
11147 This work was sponsored by Logica.
11151 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11152 keystores. Support for SSL/TLS client authentication too.
11153 Not compiled unless enable-capieng specified to Configure.
11155 This work was sponsored by Logica.
11159 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11160 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11161 attribute creation routines such as certificate requests and PKCS#12
11166 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11168 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11169 handshake which could lead to a client crash as found using the
11170 Codenomicon TLS test suite ([CVE-2008-1672])
11172 *Steve Henson, Mark Cox*
11174 * Fix double free in TLS server name extensions which could lead to
11175 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11179 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11181 Clear the error queue to ensure that error entries left from
11182 older function calls do not interfere with the correct operation.
11184 *Lutz Jaenicke, Erik de Castro Lopo*
11186 * Remove root CA certificates of commercial CAs:
11188 The OpenSSL project does not recommend any specific CA and does not
11189 have any policy with respect to including or excluding any CA.
11190 Therefore, it does not make any sense to ship an arbitrary selection
11191 of root CA certificates with the OpenSSL software.
11195 * RSA OAEP patches to fix two separate invalid memory reads.
11196 The first one involves inputs when 'lzero' is greater than
11197 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11198 before the beginning of from). The second one involves inputs where
11199 the 'db' section contains nothing but zeroes (there is a one-byte
11200 invalid read after the end of 'db').
11202 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11204 * Partial backport from 0.9.9-dev:
11206 Introduce bn_mul_mont (dedicated Montgomery multiplication
11207 procedure) as a candidate for BIGNUM assembler implementation.
11208 While 0.9.9-dev uses assembler for various architectures, only
11209 x86_64 is available by default here in the 0.9.8 branch, and
11210 32-bit x86 is available through a compile-time setting.
11212 To try the 32-bit x86 assembler implementation, use Configure
11213 option "enable-montasm" (which exists only for this backport).
11215 As "enable-montasm" for 32-bit x86 disclaims code stability
11216 anyway, in this constellation we activate additional code
11217 backported from 0.9.9-dev for further performance improvements,
11218 namely BN_from_montgomery_word. (To enable this otherwise,
11219 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11221 *Andy Polyakov (backport partially by Bodo Moeller)*
11223 * Add TLS session ticket callback. This allows an application to set
11224 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11225 values. This is useful for key rollover for example where several key
11226 sets may exist with different names.
11230 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11231 This was broken until now in 0.9.8 releases, such that the only way
11232 a registered ENGINE could be used (assuming it initialises
11233 successfully on the host) was to explicitly set it as the default
11234 for the relevant algorithms. This is in contradiction with 0.9.7
11235 behaviour and the documentation. With this fix, when an ENGINE is
11236 registered into a given algorithm's table of implementations, the
11237 'uptodate' flag is reset so that auto-discovery will be used next
11238 time a new context for that algorithm attempts to select an
11241 *Ian Lister (tweaked by Geoff Thorpe)*
11243 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11244 implementation in the following ways:
11246 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11249 Lack of BER streaming support means one pass streaming processing is
11250 only supported if data is detached: setting the streaming flag is
11251 ignored for embedded content.
11253 CMS support is disabled by default and must be explicitly enabled
11254 with the enable-cms configuration option.
11258 * Update the GMP engine glue to do direct copies between BIGNUM and
11259 mpz_t when openssl and GMP use the same limb size. Otherwise the
11260 existing "conversion via a text string export" trick is still used.
11262 *Paul Sheer <paulsheer@gmail.com>*
11264 * Zlib compression BIO. This is a filter BIO which compressed and
11265 uncompresses any data passed through it.
11269 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11270 RFC3394 compatible AES key wrapping.
11274 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11275 sets string data without copying. X509_ALGOR_set0() and
11276 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11277 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11278 from an X509_ATTRIBUTE structure optionally checking it occurs only
11279 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11284 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11285 to get the expected BN_FLG_CONSTTIME behavior.
11287 *Bodo Moeller (Google)*
11291 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11292 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11293 - added some more tests to do_tests.pl
11294 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11295 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11296 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11297 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11298 - various changes to netware.pl to enable gcc-cross builds on Win32
11300 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11301 - various changes to fix missing prototype warnings
11302 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11303 - added AES, WHIRLPOOL and CPUID assembler code to build files
11304 - added missing AES assembler make rules to mk1mf.pl
11305 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11307 *Guenter Knauf <eflash@gmx.net>*
11309 * Implement certificate status request TLS extension defined in RFC3546.
11310 A client can set the appropriate parameters and receive the encoded
11311 OCSP response via a callback. A server can query the supplied parameters
11312 and set the encoded OCSP response in the callback. Add simplified examples
11313 to s_client and s_server.
11317 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11319 * Fix various bugs:
11320 + Binary incompatibility of ssl_ctx_st structure
11321 + DTLS interoperation with non-compliant servers
11322 + Don't call get_session_cb() without proposed session
11323 + Fix ia64 assembler code
11325 *Andy Polyakov, Steve Henson*
11327 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11329 * DTLS Handshake overhaul. There were longstanding issues with
11330 OpenSSL DTLS implementation, which were making it impossible for
11331 RFC 4347 compliant client to communicate with OpenSSL server.
11332 Unfortunately just fixing these incompatibilities would "cut off"
11333 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11334 server keeps tolerating non RFC compliant syntax. The opposite is
11335 not true, 0.9.8f client can not communicate with earlier server.
11336 This update even addresses CVE-2007-4995.
11340 * Changes to avoid need for function casts in OpenSSL: some compilers
11341 (gcc 4.2 and later) reject their use.
11342 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11345 * Add RFC4507 support to OpenSSL. This includes the corrections in
11346 RFC4507bis. The encrypted ticket format is an encrypted encoded
11347 SSL_SESSION structure, that way new session features are automatically
11350 If a client application caches session in an SSL_SESSION structure
11351 support is transparent because tickets are now stored in the encoded
11354 The SSL_CTX structure automatically generates keys for ticket
11355 protection in servers so again support should be possible
11356 with no application modification.
11358 If a client or server wishes to disable RFC4507 support then the option
11359 SSL_OP_NO_TICKET can be set.
11361 Add a TLS extension debugging callback to allow the contents of any client
11362 or server extensions to be examined.
11364 This work was sponsored by Google.
11368 * Add initial support for TLS extensions, specifically for the server_name
11369 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11370 have new members for a hostname. The SSL data structure has an
11371 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11372 stored in that context to allow for session resumption, even after the
11373 SSL has been switched to a new SSL_CTX in reaction to a client's
11374 server_name extension.
11376 New functions (subject to change):
11378 SSL_get_servername()
11379 SSL_get_servername_type()
11382 New CTRL codes and macros (subject to change):
11384 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11385 - SSL_CTX_set_tlsext_servername_callback()
11386 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11387 - SSL_CTX_set_tlsext_servername_arg()
11388 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11390 openssl s_client has a new '-servername ...' option.
11392 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11393 '-key2 ...', '-servername_fatal' (subject to change). This allows
11394 testing the HostName extension for a specific single hostname ('-cert'
11395 and '-key' remain fallbacks for handshakes without HostName
11396 negotiation). If the unrecognized_name alert has to be sent, this by
11397 default is a warning; it becomes fatal with the '-servername_fatal'
11400 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11402 * Add AES and SSE2 assembly language support to VC++ build.
11406 * Mitigate attack on final subtraction in Montgomery reduction.
11410 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11411 (which previously caused an internal error).
11415 * Squeeze another 10% out of IGE mode when in != out.
11419 * AES IGE mode speedup.
11421 *Dean Gaudet (Google)*
11423 * Add the Korean symmetric 128-bit cipher SEED (see
11424 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11425 add SEED ciphersuites from RFC 4162:
11427 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11428 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11429 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11430 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11432 To minimize changes between patchlevels in the OpenSSL 0.9.8
11433 series, SEED remains excluded from compilation unless OpenSSL
11434 is configured with 'enable-seed'.
11436 *KISA, Bodo Moeller*
11438 * Mitigate branch prediction attacks, which can be practical if a
11439 single processor is shared, allowing a spy process to extract
11440 information. For detailed background information, see
11441 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11442 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11443 and Necessary Software Countermeasures"). The core of the change
11444 are new versions BN_div_no_branch() and
11445 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11446 respectively, which are slower, but avoid the security-relevant
11447 conditional branches. These are automatically called by BN_div()
11448 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11449 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11450 remove a conditional branch.
11452 BN_FLG_CONSTTIME is the new name for the previous
11453 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11454 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11455 in the exponent causes BN_mod_exp_mont() to use the alternative
11456 implementation in BN_mod_exp_mont_consttime().) The old name
11457 remains as a deprecated alias.
11459 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11460 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11461 constant-time implementations for more than just exponentiation.
11462 Here too the old name is kept as a deprecated alias.
11464 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11465 the BN_BLINDING structure gets an independent copy of the
11466 modulus. This means that the previous `BIGNUM *m` argument to
11467 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11468 essentially becomes `const BIGNUM *m`, although we can't actually
11469 change this in the header file before 0.9.9. It allows
11470 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11471 enable BN_FLG_CONSTTIME.
11473 *Matthew D Wood (Intel Corp)*
11475 * In the SSL/TLS server implementation, be strict about session ID
11476 context matching (which matters if an application uses a single
11477 external cache for different purposes). Previously,
11478 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11479 set. This did ensure strict client verification, but meant that,
11480 with applications using a single external cache for quite
11481 different requirements, clients could circumvent ciphersuite
11482 restrictions for a given session ID context by starting a session
11483 in a different context.
11487 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11488 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11489 authentication-only ciphersuites.
11493 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11494 not complete and could lead to a possible single byte overflow
11495 ([CVE-2007-5135]) [Ben Laurie]
11497 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11499 * Since AES128 and AES256 (and similarly Camellia128 and
11500 Camellia256) share a single mask bit in the logic of
11501 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11502 kludge to work properly if AES128 is available and AES256 isn't
11503 (or if Camellia128 is available and Camellia256 isn't).
11507 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11508 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11509 When a point or a seed is encoded in a BIT STRING, we need to
11510 prevent the removal of trailing zero bits to get the proper DER
11511 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11512 of a NamedBitList, for which trailing 0 bits need to be removed.)
11516 * Have SSL/TLS server implementation tolerate "mismatched" record
11517 protocol version while receiving ClientHello even if the
11518 ClientHello is fragmented. (The server can't insist on the
11519 particular protocol version it has chosen before the ServerHello
11520 message has informed the client about his choice.)
11524 * Add RFC 3779 support.
11526 *Rob Austein for ARIN, Ben Laurie*
11528 * Load error codes if they are not already present instead of using a
11529 static variable. This allows them to be cleanly unloaded and reloaded.
11530 Improve header file function name parsing.
11534 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11535 or CAPABILITY handshake as required by RFCs.
11537 *Goetz Babin-Ebell*
11539 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11541 * Introduce limits to prevent malicious keys being able to
11542 cause a denial of service. ([CVE-2006-2940])
11544 *Steve Henson, Bodo Moeller*
11546 * Fix ASN.1 parsing of certain invalid structures that can result
11547 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11549 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11550 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11552 * Fix SSL client code which could crash if connecting to a
11553 malicious SSLv2 server. ([CVE-2006-4343])
11555 *Tavis Ormandy and Will Drewry, Google Security Team*
11557 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11558 match only those. Before that, "AES256-SHA" would be interpreted
11559 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11560 the same strength classification in 0.9.7h) as we currently only
11561 have a single AES bit in the ciphersuite description bitmap.
11562 That change, however, also applied to ciphersuite strings such as
11563 "RC4-MD5" that intentionally matched multiple ciphersuites --
11564 namely, SSL 2.0 ciphersuites in addition to the more common ones
11565 from SSL 3.0/TLS 1.0.
11567 So we change the selection algorithm again: Naming an explicit
11568 ciphersuite selects this one ciphersuite, and any other similar
11569 ciphersuite (same bitmap) from *other* protocol versions.
11570 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11571 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11573 Since SSL 2.0 does not have any ciphersuites for which the
11574 128/256 bit distinction would be relevant, this works for now.
11575 The proper fix will be to use different bits for AES128 and
11576 AES256, which would have avoided the problems from the beginning;
11577 however, bits are scarce, so we can only do this in a new release
11578 (not just a patchlevel) when we can change the SSL_CIPHER
11579 definition to split the single 'unsigned long mask' bitmap into
11580 multiple values to extend the available space.
11584 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11586 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11587 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11589 * Add AES IGE and biIGE modes.
11593 * Change the Unix randomness entropy gathering to use poll() when
11594 possible instead of select(), since the latter has some
11595 undesirable limitations.
11597 *Darryl Miles via Richard Levitte and Bodo Moeller*
11599 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11600 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11601 cannot be implicitly activated as part of, e.g., the "AES" alias.
11602 However, please upgrade to OpenSSL 0.9.9[-dev] for
11603 non-experimental use of the ECC ciphersuites to get TLS extension
11604 support, which is required for curve and point format negotiation
11605 to avoid potential handshake problems.
11609 * Disable rogue ciphersuites:
11611 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11612 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11613 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11615 The latter two were purportedly from
11616 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11619 Also deactivate the remaining ciphersuites from
11620 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11621 unofficial, and the ID has long expired.
11625 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11626 dual-core machines) and other potential thread-safety issues.
11630 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11631 versions), which is now available for royalty-free use
11632 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11633 Also, add Camellia TLS ciphersuites from RFC 4132.
11635 To minimize changes between patchlevels in the OpenSSL 0.9.8
11636 series, Camellia remains excluded from compilation unless OpenSSL
11637 is configured with 'enable-camellia'.
11641 * Disable the padding bug check when compression is in use. The padding
11642 bug check assumes the first packet is of even length, this is not
11643 necessarily true if compression is enabled and can result in false
11644 positives causing handshake failure. The actual bug test is ancient
11645 code so it is hoped that implementations will either have fixed it by
11646 now or any which still have the bug do not support compression.
11650 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11652 * When applying a cipher rule check to see if string match is an explicit
11653 cipher suite and only match that one cipher suite if it is.
11657 * Link in manifests for VC++ if needed.
11659 *Austin Ziegler <halostatue@gmail.com>*
11661 * Update support for ECC-based TLS ciphersuites according to
11662 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11663 TLS extensions, which are supported starting with the 0.9.9
11664 branch, not in the OpenSSL 0.9.8 branch).
11668 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11669 opaque EVP_CIPHER_CTX handling.
11673 * Fixes and enhancements to zlib compression code. We now only use
11674 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11675 to conform with the standards mentioned here:
11676 <http://www.zlib.net/DLL_FAQ.txt>
11677 Static zlib linking now works on Windows and the new --with-zlib-include
11678 --with-zlib-lib options to Configure can be used to supply the location
11679 of the headers and library. Gracefully handle case where zlib library
11684 * Several fixes and enhancements to the OID generation code. The old code
11685 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11686 handle numbers larger than ULONG_MAX, truncated printing and had a
11687 non standard OBJ_obj2txt() behaviour.
11691 * Add support for building of engines under engine/ as shared libraries
11692 under VC++ build system.
11696 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11697 Hopefully, we will not see any false combination of paths any more.
11701 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11703 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11704 (part of SSL_OP_ALL). This option used to disable the
11705 countermeasure against man-in-the-middle protocol-version
11706 rollback in the SSL 2.0 server implementation, which is a bad
11707 idea. ([CVE-2005-2969])
11709 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11710 for Information Security, National Institute of Advanced Industrial
11711 Science and Technology [AIST], Japan)*
11713 * Add two function to clear and return the verify parameter flags.
11717 * Keep cipherlists sorted in the source instead of sorting them at
11718 runtime, thus removing the need for a lock.
11722 * Avoid some small subgroup attacks in Diffie-Hellman.
11724 *Nick Mathewson and Ben Laurie*
11726 * Add functions for well-known primes.
11730 * Extended Windows CE support.
11732 *Satoshi Nakamura and Andy Polyakov*
11734 * Initialize SSL_METHOD structures at compile time instead of during
11735 runtime, thus removing the need for a lock.
11739 * Make PKCS7_decrypt() work even if no certificate is supplied by
11740 attempting to decrypt each encrypted key in turn. Add support to
11745 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11747 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11750 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11754 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11755 key into the same file any more.
11759 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11763 * Add -utf8 command line and config file option to 'ca'.
11765 *Stefan <stf@udoma.org*
11767 * Removed the macro des_crypt(), as it seems to conflict with some
11768 libraries. Use DES_crypt().
11772 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11773 involves renaming the source and generated shared-libs for
11774 both. The engines will accept the corrected or legacy ids
11775 ('ncipher' and '4758_cca' respectively) when binding. NB,
11776 this only applies when building 'shared'.
11778 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11780 * Add attribute functions to EVP_PKEY structure. Modify
11781 PKCS12_create() to recognize a CSP name attribute and
11782 use it. Make -CSP option work again in pkcs12 utility.
11786 * Add new functionality to the bn blinding code:
11787 - automatic re-creation of the BN_BLINDING parameters after
11788 a fixed number of uses (currently 32)
11789 - add new function for parameter creation
11790 - introduce flags to control the update behaviour of the
11791 BN_BLINDING parameters
11792 - hide BN_BLINDING structure
11793 Add a second BN_BLINDING slot to the RSA structure to improve
11794 performance when a single RSA object is shared among several
11799 * Add support for DTLS.
11801 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11803 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11804 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11808 * Remove buggy and incomplete DH cert support from
11809 ssl/ssl_rsa.c and ssl/s3_both.c
11813 * Use SHA-1 instead of MD5 as the default digest algorithm for
11814 the `apps/openssl` commands.
11818 * Compile clean with "-Wall -Wmissing-prototypes
11819 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11820 DEBUG_SAFESTACK must also be set.
11824 * Change ./Configure so that certain algorithms can be disabled by default.
11825 The new counterpiece to "no-xxx" is "enable-xxx".
11827 The patented RC5 and MDC2 algorithms will now be disabled unless
11828 "enable-rc5" and "enable-mdc2", respectively, are specified.
11830 (IDEA remains enabled despite being patented. This is because IDEA
11831 is frequently required for interoperability, and there is no license
11832 fee for non-commercial use. As before, "no-idea" can be used to
11833 avoid this algorithm.)
11837 * Add processing of proxy certificates (see RFC 3820). This work was
11838 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11839 EGEE (Enabling Grids for E-science in Europe).
11843 * RC4 performance overhaul on modern architectures/implementations, such
11844 as Intel P4, IA-64 and AMD64.
11848 * New utility extract-section.pl. This can be used specify an alternative
11849 section number in a pod file instead of having to treat each file as
11850 a separate case in Makefile. This can be done by adding two lines to the
11853 =for comment openssl_section:XXX
11855 The blank line is mandatory.
11859 * New arguments -certform, -keyform and -pass for s_client and s_server
11860 to allow alternative format key and certificate files and passphrase
11865 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11866 update associated structures and add various utility functions.
11868 Add new policy related verify parameters, include policy checking in
11869 standard verify code. Enhance 'smime' application with extra parameters
11870 to support policy checking and print out.
11874 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11875 Nehemiah processors. These extensions support AES encryption in hardware
11876 as well as RNG (though RNG support is currently disabled).
11878 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11880 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11884 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11886 *Andy Polyakov and a number of other people*
11888 * Improved PowerPC platform support. Most notably BIGNUM assembler
11889 implementation contributed by IBM.
11891 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11893 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11894 exponent rather than 'unsigned long'. There is a corresponding change to
11895 the new 'rsa_keygen' element of the RSA_METHOD structure.
11897 *Jelte Jansen, Geoff Thorpe*
11899 * Functionality for creating the initial serial number file is now
11900 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11902 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11903 number file to 1, which is bound to cause problems. To avoid
11904 the problems while respecting compatibility between different 0.9.7
11905 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11906 CA.pl for serial number initialization. With the new release 0.9.8,
11907 we can fix the problem directly in the 'ca' utility.)
11911 * Reduced header interdependencies by declaring more opaque objects in
11912 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11913 give fewer recursive includes, which could break lazy source code - so
11914 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11915 developers should define this symbol when building and using openssl to
11916 ensure they track the recommended behaviour, interfaces, [etc], but
11917 backwards-compatible behaviour prevails when this isn't defined.
11921 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11925 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11926 This will generate a random key of the appropriate length based on the
11927 cipher context. The EVP_CIPHER can provide its own random key generation
11928 routine to support keys of a specific form. This is used in the des and
11929 3des routines to generate a key of the correct parity. Update S/MIME
11930 code to use new functions and hence generate correct parity DES keys.
11931 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11932 valid (weak or incorrect parity).
11936 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11937 as looking them up. This is useful when the verified structure may contain
11938 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11939 present unless the new PKCS7_NO_CRL flag is asserted.
11943 * Extend ASN1 oid configuration module. It now additionally accepts the
11946 shortName = some long name, 1.2.3.4
11950 * Reimplemented the BN_CTX implementation. There is now no more static
11951 limitation on the number of variables it can handle nor the depth of the
11952 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11953 information can now expand as required, and rather than having a single
11954 static array of bignums, BN_CTX now uses a linked-list of such arrays
11955 allowing it to expand on demand whilst maintaining the usefulness of
11956 BN_CTX's "bundling".
11960 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11961 to allow all RSA operations to function using a single BN_CTX.
11965 * Preliminary support for certificate policy evaluation and checking. This
11966 is initially intended to pass the tests outlined in "Conformance Testing
11967 of Relying Party Client Certificate Path Processing Logic" v1.07.
11971 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11972 remained unused and not that useful. A variety of other little bignum
11973 tweaks and fixes have also been made continuing on from the audit (see
11978 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11979 associated ASN1, EVP and SSL functions and old ASN1 macros.
11983 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11984 and this should never fail. So the return value from the use of
11985 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11986 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11990 * BN_CTX_get() should return zero-valued bignums, providing the same
11991 initialised value as BN_new().
11993 *Geoff Thorpe, suggested by Ulf Möller*
11995 * Support for inhibitAnyPolicy certificate extension.
11999 * An audit of the BIGNUM code is underway, for which debugging code is
12000 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
12001 is considered valid when processing BIGNUMs, and causes execution to
12002 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
12003 further steps are taken to deliberately pollute unused data in BIGNUM
12004 structures to try and expose faulty code further on. For now, openssl will
12005 (in its default mode of operation) continue to tolerate the inconsistent
12006 forms that it has tolerated in the past, but authors and packagers should
12007 consider trying openssl and their own applications when compiled with
12008 these debugging symbols defined. It will help highlight potential bugs in
12009 their own code, and will improve the test coverage for OpenSSL itself. At
12010 some point, these tighter rules will become openssl's default to improve
12011 maintainability, though the assert()s and other overheads will remain only
12012 in debugging configurations. See bn.h for more details.
12014 *Geoff Thorpe, Nils Larsch, Ulf Möller*
12016 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
12017 that can only be obtained through BN_CTX_new() (which implicitly
12018 initialises it). The presence of this function only made it possible
12019 to overwrite an existing structure (and cause memory leaks).
12023 * Because of the callback-based approach for implementing LHASH as a
12024 template type, lh_insert() adds opaque objects to hash-tables and
12025 lh_doall() or lh_doall_arg() are typically used with a destructor callback
12026 to clean up those corresponding objects before destroying the hash table
12027 (and losing the object pointers). So some over-zealous constifications in
12028 LHASH have been relaxed so that lh_insert() does not take (nor store) the
12029 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
12030 prototyped to have "const" restrictions on the object pointers they are
12031 given (and so aren't required to cast them away any more).
12035 * The tmdiff.h API was so ugly and minimal that our own timing utility
12036 (speed) prefers to use its own implementation. The two implementations
12037 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
12038 its object type properly exposed (MS_TM) instead of casting to/from
12039 `char *`. This may still change yet if someone realises MS_TM and
12041 aren't necessarily the greatest nomenclatures - but this is what was used
12042 internally to the implementation so I've used that for now.
12046 * Ensure that deprecated functions do not get compiled when
12047 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
12048 the self-tests were still using deprecated key-generation functions so
12049 these have been updated also.
12053 * Reorganise PKCS#7 code to separate the digest location functionality
12054 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
12055 New function PKCS7_set_digest() to set the digest type for PKCS#7
12056 digestedData type. Add additional code to correctly generate the
12057 digestedData type and add support for this type in PKCS7 initialization
12062 * New function PKCS7_set0_type_other() this initializes a PKCS7
12063 structure of type "other".
12067 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
12068 sure the loop does correctly stop and breaking ("division by zero")
12069 modulus operations are not performed. The (pre-generated) prime
12070 table crypto/bn/bn_prime.h was already correct, but it could not be
12071 re-generated on some platforms because of the "division by zero"
12072 situation in the script.
12074 *Ralf S. Engelschall*
12076 * Update support for ECC-based TLS ciphersuites according to
12077 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
12078 SHA-1 now is only used for "small" curves (where the
12079 representation of a field element takes up to 24 bytes); for
12080 larger curves, the field element resulting from ECDH is directly
12081 used as premaster secret.
12083 *Douglas Stebila (Sun Microsystems Laboratories)*
12085 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
12086 curve secp160r1 to the tests.
12088 *Douglas Stebila (Sun Microsystems Laboratories)*
12090 * Add the possibility to load symbols globally with DSO.
12092 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
12094 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
12095 control of the error stack.
12099 * Add support for STORE in ENGINE.
12103 * Add the STORE type. The intention is to provide a common interface
12104 to certificate and key stores, be they simple file-based stores, or
12105 HSM-type store, or LDAP stores, or...
12106 NOTE: The code is currently UNTESTED and isn't really used anywhere.
12110 * Add a generic structure called OPENSSL_ITEM. This can be used to
12111 pass a list of arguments to any function as well as provide a way
12112 for a function to pass data back to the caller.
12116 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
12117 works like BUF_strdup() but can be used to duplicate a portion of
12118 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
12123 * Add the function sk_find_ex() which works like sk_find(), but will
12124 return an index to an element even if an exact match couldn't be
12125 found. The index is guaranteed to point at the element where the
12126 searched-for key would be inserted to preserve sorting order.
12130 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12131 takes an extra flags argument for optional functionality. Currently,
12132 the following flags are defined:
12134 OBJ_BSEARCH_VALUE_ON_NOMATCH
12135 This one gets OBJ_bsearch_ex() to return a pointer to the first
12136 element where the comparing function returns a negative or zero
12139 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12140 This one gets OBJ_bsearch_ex() to return a pointer to the first
12141 element where the comparing function returns zero. This is useful
12142 if there are more than one element where the comparing function
12147 * Make it possible to create self-signed certificates with 'openssl ca'
12148 in such a way that the self-signed certificate becomes part of the
12149 CA database and uses the same mechanisms for serial number generation
12150 as all other certificate signing. The new flag '-selfsign' enables
12151 this functionality. Adapt CA.sh and CA.pl.in.
12155 * Add functionality to check the public key of a certificate request
12156 against a given private. This is useful to check that a certificate
12157 request can be signed by that key (self-signing).
12161 * Make it possible to have multiple active certificates with the same
12162 subject in the CA index file. This is done only if the keyword
12163 'unique_subject' is set to 'no' in the main CA section (default
12164 if 'CA_default') of the configuration file. The value is saved
12165 with the database itself in a separate index attribute file,
12166 named like the index file with '.attr' appended to the name.
12170 * Generate multi-valued AVAs using '+' notation in config files for
12175 * Support for nameConstraints certificate extension.
12179 * Support for policyConstraints certificate extension.
12183 * Support for policyMappings certificate extension.
12187 * Make sure the default DSA_METHOD implementation only uses its
12188 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12189 and change its own handlers to be NULL so as to remove unnecessary
12190 indirection. This lets alternative implementations fallback to the
12191 default implementation more easily.
12195 * Support for directoryName in GeneralName related extensions
12200 * Make it possible to link applications using Makefile.shared.
12201 Make that possible even when linking against static libraries!
12205 * Support for single pass processing for S/MIME signing. This now
12206 means that S/MIME signing can be done from a pipe, in addition
12207 cleartext signing (multipart/signed type) is effectively streaming
12208 and the signed data does not need to be all held in memory.
12210 This is done with a new flag PKCS7_STREAM. When this flag is set
12211 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12212 is done after the data is output (and digests calculated) in
12213 SMIME_write_PKCS7().
12217 * Add full support for -rpath/-R, both in shared libraries and
12218 applications, at least on the platforms where it's known how
12223 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12224 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12225 will now compute a table of multiples of the generator that
12226 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12227 faster (notably in the case of a single point multiplication,
12228 scalar * generator).
12230 *Nils Larsch, Bodo Moeller*
12232 * IPv6 support for certificate extensions. The various extensions
12233 which use the IP:a.b.c.d can now take IPv6 addresses using the
12234 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12239 * Added an ENGINE that implements RSA by performing private key
12240 exponentiations with the GMP library. The conversions to and from
12241 GMP's mpz_t format aren't optimised nor are any montgomery forms
12242 cached, and on x86 it appears OpenSSL's own performance has caught up.
12243 However there are likely to be other architectures where GMP could
12244 provide a boost. This ENGINE is not built in by default, but it can be
12245 specified at Configure time and should be accompanied by the necessary
12246 linker additions, eg;
12247 ./config -DOPENSSL_USE_GMP -lgmp
12251 * "openssl engine" will not display ENGINE/DSO load failure errors when
12252 testing availability of engines with "-t" - the old behaviour is
12253 produced by increasing the feature's verbosity with "-tt".
12257 * ECDSA routines: under certain error conditions uninitialized BN objects
12258 could be freed. Solution: make sure initialization is performed early
12259 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12264 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12265 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12266 software implementations. For DSA and DH, parameter generation can
12267 also be overridden by providing the appropriate method callbacks.
12271 * Change the "progress" mechanism used in key-generation and
12272 primality testing to functions that take a new BN_GENCB pointer in
12273 place of callback/argument pairs. The new API functions have `_ex`
12274 postfixes and the older functions are reimplemented as wrappers for
12275 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12276 declarations of the old functions to help (graceful) attempts to
12277 migrate to the new functions. Also, the new key-generation API
12278 functions operate on a caller-supplied key-structure and return
12279 success/failure rather than returning a key or NULL - this is to
12280 help make "keygen" another member function of RSA_METHOD etc.
12282 Example for using the new callback interface:
12284 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12285 void *my_arg = ...;
12288 BN_GENCB_set(&my_cb, my_callback, my_arg);
12290 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12291 /* For the meaning of a, b in calls to my_callback(), see the
12292 * documentation of the function that calls the callback.
12293 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12294 * my_callback should return 1 if it wants BN_is_prime_ex()
12295 * to continue, or 0 to stop.
12300 * Change the ZLIB compression method to be stateful, and make it
12301 available to TLS with the number defined in
12302 draft-ietf-tls-compression-04.txt.
12306 * Add the ASN.1 structures and functions for CertificatePair, which
12307 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12309 CertificatePair ::= SEQUENCE {
12310 forward [0] Certificate OPTIONAL,
12311 reverse [1] Certificate OPTIONAL,
12312 -- at least one of the pair shall be present -- }
12314 Also implement the PEM functions to read and write certificate
12315 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12317 This needed to be defined, mostly for the sake of the LDAP
12318 attribute crossCertificatePair, but may prove useful elsewhere as
12323 * Make it possible to inhibit symlinking of shared libraries in
12324 Makefile.shared, for Cygwin's sake.
12328 * Extend the BIGNUM API by creating a function
12329 void BN_set_negative(BIGNUM *a, int neg);
12330 and a macro that behave like
12331 int BN_is_negative(const BIGNUM *a);
12333 to avoid the need to access 'a->neg' directly in applications.
12337 * Implement fast modular reduction for pseudo-Mersenne primes
12338 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12339 EC_GROUP_new_curve_GFp() will now automatically use this
12342 *Nils Larsch <nla@trustcenter.de>*
12344 * Add new lock type (CRYPTO_LOCK_BN).
12348 * Change the ENGINE framework to automatically load engines
12349 dynamically from specific directories unless they could be
12350 found to already be built in or loaded. Move all the
12351 current engines except for the cryptodev one to a new
12352 directory engines/.
12353 The engines in engines/ are built as shared libraries if
12354 the "shared" options was given to ./Configure or ./config.
12355 Otherwise, they are inserted in libcrypto.a.
12356 /usr/local/ssl/engines is the default directory for dynamic
12357 engines, but that can be overridden at configure time through
12358 the usual use of --prefix and/or --openssldir, and at run
12359 time with the environment variable OPENSSL_ENGINES.
12361 *Geoff Thorpe and Richard Levitte*
12363 * Add Makefile.shared, a helper makefile to build shared
12364 libraries. Adapt Makefile.org.
12368 * Add version info to Win32 DLLs.
12370 *Peter 'Luna' Runestig" <peter@runestig.com>*
12372 * Add new 'medium level' PKCS#12 API. Certificates and keys
12373 can be added using this API to created arbitrary PKCS#12
12374 files while avoiding the low-level API.
12376 New options to PKCS12_create(), key or cert can be NULL and
12377 will then be omitted from the output file. The encryption
12378 algorithm NIDs can be set to -1 for no encryption, the mac
12379 iteration count can be set to 0 to omit the mac.
12381 Enhance pkcs12 utility by making the -nokeys and -nocerts
12382 options work when creating a PKCS#12 file. New option -nomac
12383 to omit the mac, NONE can be set for an encryption algorithm.
12384 New code is modified to use the enhanced PKCS12_create()
12385 instead of the low-level API.
12389 * Extend ASN1 encoder to support indefinite length constructed
12390 encoding. This can output sequences tags and octet strings in
12391 this form. Modify pk7_asn1.c to support indefinite length
12392 encoding. This is experimental and needs additional code to
12393 be useful, such as an ASN1 bio and some enhanced streaming
12396 Extend template encode functionality so that tagging is passed
12397 down to the template encoder.
12401 * Let 'openssl req' fail if an argument to '-newkey' is not
12402 recognized instead of using RSA as a default.
12406 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12407 As these are not official, they are not included in "ALL";
12408 the "ECCdraft" ciphersuite group alias can be used to select them.
12410 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12412 * Add ECDH engine support.
12414 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12416 * Add ECDH in new directory crypto/ecdh/.
12418 *Douglas Stebila (Sun Microsystems Laboratories)*
12420 * Let BN_rand_range() abort with an error after 100 iterations
12421 without success (which indicates a broken PRNG).
12425 * Change BN_mod_sqrt() so that it verifies that the input value
12426 is really the square of the return value. (Previously,
12427 BN_mod_sqrt would show GIGO behaviour.)
12431 * Add named elliptic curves over binary fields from X9.62, SECG,
12432 and WAP/WTLS; add OIDs that were still missing.
12434 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12436 * Extend the EC library for elliptic curves over binary fields
12437 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12440 EC_GF2m_simple_method
12444 EC_GROUP_new_curve_GF2m
12445 EC_GROUP_set_curve_GF2m
12446 EC_GROUP_get_curve_GF2m
12447 EC_POINT_set_affine_coordinates_GF2m
12448 EC_POINT_get_affine_coordinates_GF2m
12449 EC_POINT_set_compressed_coordinates_GF2m
12451 Point compression for binary fields is disabled by default for
12452 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12455 As binary polynomials are represented as BIGNUMs, various members
12456 of the EC_GROUP and EC_POINT data structures can be shared
12457 between the implementations for prime fields and binary fields;
12458 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12459 are essentially identical to their `..._GFp` counterparts.
12460 (For simplicity, the `..._GFp` prefix has been dropped from
12461 various internal method names.)
12463 An internal 'field_div' method (similar to 'field_mul' and
12464 'field_sqr') has been added; this is used only for binary fields.
12466 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12468 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12469 through methods ('mul', 'precompute_mult').
12471 The generic implementations (now internally called 'ec_wNAF_mul'
12472 and 'ec_wNAF_precomputed_mult') remain the default if these
12473 methods are undefined.
12475 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12477 * New function EC_GROUP_get_degree, which is defined through
12478 EC_METHOD. For curves over prime fields, this returns the bit
12479 length of the modulus.
12481 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12483 * New functions EC_GROUP_dup, EC_POINT_dup.
12484 (These simply call ..._new and ..._copy).
12486 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12488 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12489 Polynomials are represented as BIGNUMs (where the sign bit is not
12490 used) in the following functions [macros]:
12493 BN_GF2m_sub [= BN_GF2m_add]
12494 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12495 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12496 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12498 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12499 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12500 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12501 BN_GF2m_cmp [= BN_ucmp]
12503 (Note that only the 'mod' functions are actually for fields GF(2^m).
12504 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12506 For some functions, an the irreducible polynomial defining a
12507 field can be given as an 'unsigned int[]' with strictly
12508 decreasing elements giving the indices of those bits that are set;
12509 i.e., p[] represents the polynomial
12510 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12512 p[0] > p[1] > ... > p[k] = 0.
12513 This applies to the following functions:
12516 BN_GF2m_mod_mul_arr
12517 BN_GF2m_mod_sqr_arr
12518 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12519 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12520 BN_GF2m_mod_exp_arr
12521 BN_GF2m_mod_sqrt_arr
12522 BN_GF2m_mod_solve_quad_arr
12526 Conversion can be performed by the following functions:
12531 bntest.c has additional tests for binary polynomial arithmetic.
12533 Two implementations for BN_GF2m_mod_div() are available.
12534 The default algorithm simply uses BN_GF2m_mod_inv() and
12535 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12536 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12537 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12539 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12541 * Add new error code 'ERR_R_DISABLED' that can be used when some
12542 functionality is disabled at compile-time.
12544 *Douglas Stebila <douglas.stebila@sun.com>*
12546 * Change default behaviour of 'openssl asn1parse' so that more
12547 information is visible when viewing, e.g., a certificate:
12549 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12550 mode the content of non-printable OCTET STRINGs is output in a
12551 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12552 avoid the appearance of a printable string.
12554 *Nils Larsch <nla@trustcenter.de>*
12556 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12558 EC_GROUP_set_asn1_flag()
12559 EC_GROUP_get_asn1_flag()
12560 EC_GROUP_set_point_conversion_form()
12561 EC_GROUP_get_point_conversion_form()
12562 These control ASN1 encoding details:
12563 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12564 has been set to OPENSSL_EC_NAMED_CURVE.
12565 - Points are encoded in uncompressed form by default; options for
12566 asn1_for are as for point2oct, namely
12567 POINT_CONVERSION_COMPRESSED
12568 POINT_CONVERSION_UNCOMPRESSED
12569 POINT_CONVERSION_HYBRID
12571 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12573 EC_GROUP_set_seed()
12574 EC_GROUP_get0_seed()
12575 EC_GROUP_get_seed_len()
12576 This is used only for ASN1 purposes (so far).
12578 *Nils Larsch <nla@trustcenter.de>*
12580 * Add 'field_type' member to EC_METHOD, which holds the NID
12581 of the appropriate field type OID. The new function
12582 EC_METHOD_get_field_type() returns this value.
12584 *Nils Larsch <nla@trustcenter.de>*
12587 EC_POINT_point2bn()
12588 EC_POINT_bn2point()
12589 EC_POINT_point2hex()
12590 EC_POINT_hex2point()
12591 providing useful interfaces to EC_POINT_point2oct() and
12592 EC_POINT_oct2point().
12594 *Nils Larsch <nla@trustcenter.de>*
12596 * Change internals of the EC library so that the functions
12597 EC_GROUP_set_generator()
12598 EC_GROUP_get_generator()
12599 EC_GROUP_get_order()
12600 EC_GROUP_get_cofactor()
12601 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12602 to methods, which would lead to unnecessary code duplication when
12603 adding different types of curves.
12605 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12607 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12608 arithmetic, and such that modified wNAFs are generated
12609 (which avoid length expansion in many cases).
12613 * Add a function EC_GROUP_check_discriminant() (defined via
12614 EC_METHOD) that verifies that the curve discriminant is non-zero.
12616 Add a function EC_GROUP_check() that makes some sanity tests
12617 on a EC_GROUP, its generator and order. This includes
12618 EC_GROUP_check_discriminant().
12620 *Nils Larsch <nla@trustcenter.de>*
12622 * Add ECDSA in new directory crypto/ecdsa/.
12624 Add applications 'openssl ecparam' and 'openssl ecdsa'
12625 (these are based on 'openssl dsaparam' and 'openssl dsa').
12627 ECDSA support is also included in various other files across the
12628 library. Most notably,
12629 - 'openssl req' now has a '-newkey ecdsa:file' option;
12630 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12631 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12632 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12633 them suitable for ECDSA where domain parameters must be
12634 extracted before the specific public key;
12635 - ECDSA engine support has been added.
12637 *Nils Larsch <nla@trustcenter.de>*
12639 * Include some named elliptic curves, and add OIDs from X9.62,
12640 SECG, and WAP/WTLS. Each curve can be obtained from the new
12642 EC_GROUP_new_by_curve_name(),
12643 and the list of available named curves can be obtained with
12644 EC_get_builtin_curves().
12645 Also add a 'curve_name' member to EC_GROUP objects, which can be
12647 EC_GROUP_set_curve_name()
12648 EC_GROUP_get_curve_name()
12650 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12652 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12653 was actually never needed) and in BN_mul(). The removal in BN_mul()
12654 required a small change in bn_mul_part_recursive() and the addition
12655 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12656 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12657 bn_sub_words() and bn_add_words() except they take arrays with
12662 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12664 * Cleanse PEM buffers before freeing them since they may contain
12667 *Benjamin Bennett <ben@psc.edu>*
12669 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12670 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12671 authentication-only ciphersuites.
12675 * Since AES128 and AES256 share a single mask bit in the logic of
12676 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12677 kludge to work properly if AES128 is available and AES256 isn't.
12681 * Expand security boundary to match 1.1.1 module.
12685 * Remove redundant features: hash file source, editing of test vectors
12686 modify fipsld to use external fips_premain.c signature.
12690 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12691 run algorithm test programs.
12695 * Make algorithm test programs more tolerant of whitespace.
12699 * Have SSL/TLS server implementation tolerate "mismatched" record
12700 protocol version while receiving ClientHello even if the
12701 ClientHello is fragmented. (The server can't insist on the
12702 particular protocol version it has chosen before the ServerHello
12703 message has informed the client about his choice.)
12707 * Load error codes if they are not already present instead of using a
12708 static variable. This allows them to be cleanly unloaded and reloaded.
12712 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12714 * Introduce limits to prevent malicious keys being able to
12715 cause a denial of service. ([CVE-2006-2940])
12717 *Steve Henson, Bodo Moeller*
12719 * Fix ASN.1 parsing of certain invalid structures that can result
12720 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12722 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12723 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12725 * Fix SSL client code which could crash if connecting to a
12726 malicious SSLv2 server. ([CVE-2006-4343])
12728 *Tavis Ormandy and Will Drewry, Google Security Team*
12730 * Change ciphersuite string processing so that an explicit
12731 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12732 will no longer include "AES128-SHA"), and any other similar
12733 ciphersuite (same bitmap) from *other* protocol versions (so that
12734 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12735 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12736 changes from 0.9.8b and 0.9.8d.
12740 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12742 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12743 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12745 * Change the Unix randomness entropy gathering to use poll() when
12746 possible instead of select(), since the latter has some
12747 undesirable limitations.
12749 *Darryl Miles via Richard Levitte and Bodo Moeller*
12751 * Disable rogue ciphersuites:
12753 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12754 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12755 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12757 The latter two were purportedly from
12758 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12761 Also deactivate the remaining ciphersuites from
12762 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12763 unofficial, and the ID has long expired.
12767 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12768 dual-core machines) and other potential thread-safety issues.
12772 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12774 * Adapt fipsld and the build system to link against the validated FIPS
12775 module in FIPS mode.
12779 * Fixes for VC++ 2005 build under Windows.
12783 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12784 from a Windows bash shell such as MSYS. It is autodetected from the
12785 "config" script when run from a VC++ environment. Modify standard VC++
12786 build to use fipscanister.o from the GNU make build.
12790 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12792 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12793 The value now differs depending on if you build for FIPS or not.
12794 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12795 safely run with a non-FIPSed libcrypto, as it may crash because of
12796 the difference induced by this change.
12800 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12802 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12803 (part of SSL_OP_ALL). This option used to disable the
12804 countermeasure against man-in-the-middle protocol-version
12805 rollback in the SSL 2.0 server implementation, which is a bad
12806 idea. ([CVE-2005-2969])
12808 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12809 for Information Security, National Institute of Advanced Industrial
12810 Science and Technology [AIST, Japan)]*
12812 * Minimal support for X9.31 signatures and PSS padding modes. This is
12813 mainly for FIPS compliance and not fully integrated at this stage.
12817 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12818 the exponentiation using a fixed-length exponent. (Otherwise,
12819 the information leaked through timing could expose the secret key
12820 after many signatures; cf. Bleichenbacher's attack on DSA with
12825 * Make a new fixed-window mod_exp implementation the default for
12826 RSA, DSA, and DH private-key operations so that the sequence of
12827 squares and multiplies and the memory access pattern are
12828 independent of the particular secret key. This will mitigate
12829 cache-timing and potential related attacks.
12831 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12832 and this is automatically used by BN_mod_exp_mont() if the new flag
12833 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12834 will use this BN flag for private exponents unless the flag
12835 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12836 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12838 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12840 * Change the client implementation for SSLv23_method() and
12841 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12842 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12843 (Previously, the SSL 2.0 backwards compatible Client Hello
12844 message format would be used even with SSL_OP_NO_SSLv2.)
12848 * Add support for smime-type MIME parameter in S/MIME messages which some
12853 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12854 a threadsafe manner. Modify rsa code to use new function and add calls
12855 to dsa and dh code (which had race conditions before).
12859 * Include the fixed error library code in the C error file definitions
12860 instead of fixing them up at runtime. This keeps the error code
12861 structures constant.
12865 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12867 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12870 * Fixes for newer kerberos headers. NB: the casts are needed because
12871 the 'length' field is signed on one version and unsigned on another
12872 with no (?) obvious way to tell the difference, without these VC++
12873 complains. Also the "definition" of FAR (blank) is no longer included
12874 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12875 some needed definitions.
12879 * Undo Cygwin change.
12883 * Added support for proxy certificates according to RFC 3820.
12884 Because they may be a security thread to unaware applications,
12885 they must be explicitly allowed in run-time. See
12886 docs/HOWTO/proxy_certificates.txt for further information.
12890 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12892 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12893 server and client random values. Previously
12894 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12895 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12897 This change has negligible security impact because:
12899 1. Server and client random values still have 24 bytes of pseudo random
12902 2. Server and client random values are sent in the clear in the initial
12905 3. The master secret is derived using the premaster secret (48 bytes in
12906 size for static RSA ciphersuites) as well as client server and random
12909 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12912 *Stephen Henson, reported by UK NISCC*
12914 * Use Windows randomness collection on Cygwin.
12918 * Fix hang in EGD/PRNGD query when communication socket is closed
12919 prematurely by EGD/PRNGD.
12921 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12923 * Prompt for pass phrases when appropriate for PKCS12 input format.
12927 * Back-port of selected performance improvements from development
12928 branch, as well as improved support for PowerPC platforms.
12932 * Add lots of checks for memory allocation failure, error codes to indicate
12933 failure and freeing up memory if a failure occurs.
12935 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12937 * Add new -passin argument to dgst.
12941 * Perform some character comparisons of different types in X509_NAME_cmp:
12942 this is needed for some certificates that re-encode DNs into UTF8Strings
12943 (in violation of RFC3280) and can't or won't issue name rollover
12948 * Make an explicit check during certificate validation to see that
12949 the CA setting in each certificate on the chain is correct. As a
12950 side effect always do the following basic checks on extensions,
12951 not just when there's an associated purpose to the check:
12953 - if there is an unhandled critical extension (unless the user
12954 has chosen to ignore this fault)
12955 - if the path length has been exceeded (if one is set at all)
12956 - that certain extensions fit the associated purpose (if one has
12961 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12963 * Avoid a race condition when CRLs are checked in a multi threaded
12964 environment. This would happen due to the reordering of the revoked
12965 entries during signature checking and serial number lookup. Now the
12966 encoding is cached and the serial number sort performed under a lock.
12967 Add new STACK function sk_is_sorted().
12971 * Add Delta CRL to the extension code.
12975 * Various fixes to s3_pkt.c so alerts are sent properly.
12977 *David Holmes <d.holmes@f5.com>*
12979 * Reduce the chances of duplicate issuer name and serial numbers (in
12980 violation of RFC3280) using the OpenSSL certificate creation utilities.
12981 This is done by creating a random 64 bit value for the initial serial
12982 number when a serial number file is created or when a self signed
12983 certificate is created using 'openssl req -x509'. The initial serial
12984 number file is created using 'openssl x509 -next_serial' in CA.pl
12985 rather than being initialized to 1.
12989 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12991 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12992 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12994 *Joe Orton, Steve Henson*
12996 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12999 *Joe Orton, Steve Henson*
13001 * Make it possible to have multiple active certificates with the same
13002 subject in the CA index file. This is done only if the keyword
13003 'unique_subject' is set to 'no' in the main CA section (default
13004 if 'CA_default') of the configuration file. The value is saved
13005 with the database itself in a separate index attribute file,
13006 named like the index file with '.attr' appended to the name.
13010 * X509 verify fixes. Disable broken certificate workarounds when
13011 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
13012 keyUsage extension present. Don't accept CRLs with unhandled critical
13013 extensions: since verify currently doesn't process CRL extensions this
13014 rejects a CRL with *any* critical extensions. Add new verify error codes
13019 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
13020 A clarification of RFC2560 will require the use of OCTET STRINGs and
13021 some implementations cannot handle the current raw format. Since OpenSSL
13022 copies and compares OCSP nonces as opaque blobs without any attempt at
13023 parsing them this should not create any compatibility issues.
13027 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
13028 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
13029 this HMAC (and other) operations are several times slower than OpenSSL
13034 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
13036 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13038 * Use the correct content when signing type "other".
13042 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
13044 * Fix various bugs revealed by running the NISCC test suite:
13046 Stop out of bounds reads in the ASN1 code when presented with
13047 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13049 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
13051 If verify callback ignores invalid public key errors don't try to check
13052 certificate signature with the NULL public key.
13056 * New -ignore_err option in ocsp application to stop the server
13057 exiting on the first error in a request.
13061 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13062 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13067 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13068 extra data after the compression methods not only for TLS 1.0
13069 but also for SSL 3.0 (as required by the specification).
13071 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13073 * Change X509_certificate_type() to mark the key as exported/exportable
13074 when it's 512 *bits* long, not 512 bytes.
13078 * Change AES_cbc_encrypt() so it outputs exact multiple of
13079 blocks during encryption.
13083 * Various fixes to base64 BIO and non blocking I/O. On write
13084 flushes were not handled properly if the BIO retried. On read
13085 data was not being buffered properly and had various logic bugs.
13086 This also affects blocking I/O when the data being decoded is a
13091 * Various S/MIME bugfixes and compatibility changes:
13092 output correct application/pkcs7 MIME type if
13093 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
13094 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
13095 of files as .eml work). Correctly handle very long lines in MIME
13100 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
13102 * Countermeasure against the Klima-Pokorny-Rosa extension of
13103 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13104 a protocol version number mismatch like a decryption error
13105 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13109 * Turn on RSA blinding by default in the default implementation
13110 to avoid a timing attack. Applications that don't want it can call
13111 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13112 They would be ill-advised to do so in most cases.
13114 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13116 * Change RSA blinding code so that it works when the PRNG is not
13117 seeded (in this case, the secret RSA exponent is abused as
13118 an unpredictable seed -- if it is not unpredictable, there
13119 is no point in blinding anyway). Make RSA blinding thread-safe
13120 by remembering the creator's thread ID in rsa->blinding and
13121 having all other threads use local one-time blinding factors
13122 (this requires more computation than sharing rsa->blinding, but
13123 avoids excessive locking; and if an RSA object is not shared
13124 between threads, blinding will still be very fast).
13128 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13129 ENGINE as defaults for all supported algorithms irrespective of
13130 the 'flags' parameter. 'flags' is now honoured, so applications
13131 should make sure they are passing it correctly.
13135 * Target "mingw" now allows native Windows code to be generated in
13136 the Cygwin environment as well as with the MinGW compiler.
13140 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13142 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13143 via timing by performing a MAC computation even if incorrect
13144 block cipher padding has been found. This is a countermeasure
13145 against active attacks where the attacker has to distinguish
13146 between bad padding and a MAC verification error. ([CVE-2003-0078])
13148 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13149 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13150 Martin Vuagnoux (EPFL, Ilion)*
13152 * Make the no-err option work as intended. The intention with no-err
13153 is not to have the whole error stack handling routines removed from
13154 libcrypto, it's only intended to remove all the function name and
13155 reason texts, thereby removing some of the footprint that may not
13156 be interesting if those errors aren't displayed anyway.
13158 NOTE: it's still possible for any application or module to have its
13159 own set of error texts inserted. The routines are there, just not
13160 used by default when no-err is given.
13164 * Add support for FreeBSD on IA64.
13166 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13168 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13169 Kerberos function mit_des_cbc_cksum(). Before this change,
13170 the value returned by DES_cbc_cksum() was like the one from
13171 mit_des_cbc_cksum(), except the bytes were swapped.
13173 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13175 * Allow an application to disable the automatic SSL chain building.
13176 Before this a rather primitive chain build was always performed in
13177 ssl3_output_cert_chain(): an application had no way to send the
13178 correct chain if the automatic operation produced an incorrect result.
13180 Now the chain builder is disabled if either:
13182 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13184 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13186 The reasoning behind this is that an application would not want the
13187 auto chain building to take place if extra chain certificates are
13188 present and it might also want a means of sending no additional
13189 certificates (for example the chain has two certificates and the
13194 * Add the possibility to build without the ENGINE framework.
13196 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13198 * Under Win32 gmtime() can return NULL: check return value in
13199 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13203 * DSA routines: under certain error conditions uninitialized BN objects
13204 could be freed. Solution: make sure initialization is performed early
13205 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13206 Nils Larsch <nla@trustcenter.de> via PR#459)
13210 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13211 checked on reconnect on the client side, therefore session resumption
13212 could still fail with a "ssl session id is different" error. This
13213 behaviour is masked when SSL_OP_ALL is used due to
13214 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13215 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13216 followup to PR #377.
13220 * IA-32 assembler support enhancements: unified ELF targets, support
13221 for SCO/Caldera platforms, fix for Cygwin shared build.
13225 * Add support for FreeBSD on sparc64. As a consequence, support for
13226 FreeBSD on non-x86 processors is separate from x86 processors on
13227 the config script, much like the NetBSD support.
13229 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13231 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13233 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13236 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13237 code (06) was taken as the first octet of the session ID and the last
13238 octet was ignored consequently. As a result SSLv2 client side session
13239 caching could not have worked due to the session ID mismatch between
13241 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13246 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13247 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13252 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13253 seems that in spite of existing for more than a year, many application
13254 author have done nothing to provide the necessary callbacks, which
13255 means that this particular engine will not work properly anywhere.
13256 This is a very unfortunate situation which forces us, in the name
13257 of usability, to give the hw_ncipher.c a static lock, which is part
13259 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13260 appear in 0.9.8 or later. We EXPECT application authors to have
13261 dealt properly with this when 0.9.8 is released (unless we actually
13262 make such changes in the libcrypto locking code that changes will
13263 have to be made anyway).
13267 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13268 octets have been read, EOF or an error occurs. Without this change
13269 some truncated ASN1 structures will not produce an error.
13273 * Disable Heimdal support, since it hasn't been fully implemented.
13274 Still give the possibility to force the use of Heimdal, but with
13275 warnings and a request that patches get sent to openssl-dev.
13279 * Add the VC-CE target, introduce the WINCE sysname, and add
13280 INSTALL.WCE and appropriate conditionals to make it build.
13282 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13284 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13285 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13286 edit numbers of the version.
13288 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13290 * Introduce safe string copy and catenation functions
13291 (BUF_strlcpy() and BUF_strlcat()).
13293 *Ben Laurie (CHATS) and Richard Levitte*
13295 * Avoid using fixed-size buffers for one-line DNs.
13297 *Ben Laurie (CHATS)*
13299 * Add BUF_MEM_grow_clean() to avoid information leakage when
13300 resizing buffers containing secrets, and use where appropriate.
13302 *Ben Laurie (CHATS)*
13304 * Avoid using fixed size buffers for configuration file location.
13306 *Ben Laurie (CHATS)*
13308 * Avoid filename truncation for various CA files.
13310 *Ben Laurie (CHATS)*
13312 * Use sizeof in preference to magic numbers.
13314 *Ben Laurie (CHATS)*
13316 * Avoid filename truncation in cert requests.
13318 *Ben Laurie (CHATS)*
13320 * Add assertions to check for (supposedly impossible) buffer
13323 *Ben Laurie (CHATS)*
13325 * Don't cache truncated DNS entries in the local cache (this could
13326 potentially lead to a spoofing attack).
13328 *Ben Laurie (CHATS)*
13330 * Fix various buffers to be large enough for hex/decimal
13331 representations in a platform independent manner.
13333 *Ben Laurie (CHATS)*
13335 * Add CRYPTO_realloc_clean() to avoid information leakage when
13336 resizing buffers containing secrets, and use where appropriate.
13338 *Ben Laurie (CHATS)*
13340 * Add BIO_indent() to avoid much slightly worrying code to do
13343 *Ben Laurie (CHATS)*
13345 * Convert sprintf()/BIO_puts() to BIO_printf().
13347 *Ben Laurie (CHATS)*
13349 * buffer_gets() could terminate with the buffer only half
13352 *Ben Laurie (CHATS)*
13354 * Add assertions to prevent user-supplied crypto functions from
13355 overflowing internal buffers by having large block sizes, etc.
13357 *Ben Laurie (CHATS)*
13359 * New OPENSSL_assert() macro (similar to assert(), but enabled
13362 *Ben Laurie (CHATS)*
13364 * Eliminate unused copy of key in RC4.
13366 *Ben Laurie (CHATS)*
13368 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13370 *Ben Laurie (CHATS)*
13372 * Fix off-by-one error in EGD path.
13374 *Ben Laurie (CHATS)*
13376 * If RANDFILE path is too long, ignore instead of truncating.
13378 *Ben Laurie (CHATS)*
13380 * Eliminate unused and incorrectly sized X.509 structure
13383 *Ben Laurie (CHATS)*
13385 * Eliminate unused and dangerous function knumber().
13387 *Ben Laurie (CHATS)*
13389 * Eliminate unused and dangerous structure, KSSL_ERR.
13391 *Ben Laurie (CHATS)*
13393 * Protect against overlong session ID context length in an encoded
13394 session object. Since these are local, this does not appear to be
13397 *Ben Laurie (CHATS)*
13399 * Change from security patch (see 0.9.6e below) that did not affect
13400 the 0.9.6 release series:
13402 Remote buffer overflow in SSL3 protocol - an attacker could
13403 supply an oversized master key in Kerberos-enabled versions.
13406 *Ben Laurie (CHATS)*
13408 * Change the SSL kerb5 codes to match RFC 2712.
13412 * Make -nameopt work fully for req and add -reqopt switch.
13414 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13416 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13418 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13420 * Make sure tests can be performed even if the corresponding algorithms
13421 have been removed entirely. This was also the last step to make
13422 OpenSSL compilable with DJGPP under all reasonable conditions.
13424 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13426 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13427 to allow version independent disabling of normally unselected ciphers,
13428 which may be activated as a side-effect of selecting a single cipher.
13430 (E.g., cipher list string "RSA" enables ciphersuites that are left
13431 out of "ALL" because they do not provide symmetric encryption.
13432 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13434 *Lutz Jaenicke, Bodo Moeller*
13436 * Add appropriate support for separate platform-dependent build
13437 directories. The recommended way to make a platform-dependent
13438 build directory is the following (tested on Linux), maybe with
13441 # Place yourself outside of the OpenSSL source tree. In
13442 # this example, the environment variable OPENSSL_SOURCE
13443 # is assumed to contain the absolute OpenSSL source directory.
13444 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13445 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13446 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13447 mkdir -p `dirname $F`
13448 ln -s $OPENSSL_SOURCE/$F $F
13451 To be absolutely sure not to disturb the source tree, a "make clean"
13452 is a good thing. If it isn't successful, don't worry about it,
13453 it probably means the source directory is very clean.
13457 * Make sure any ENGINE control commands make local copies of string
13458 pointers passed to them whenever necessary. Otherwise it is possible
13459 the caller may have overwritten (or deallocated) the original string
13460 data when a later ENGINE operation tries to use the stored values.
13462 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13464 * Improve diagnostics in file reading and command-line digests.
13466 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13468 * Add AES modes CFB and OFB to the object database. Correct an
13469 error in AES-CFB decryption.
13473 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13474 allows existing EVP_CIPHER_CTX structures to be reused after
13475 calling `EVP_*Final()`. This behaviour is used by encryption
13476 BIOs and some applications. This has the side effect that
13477 applications must explicitly clean up cipher contexts with
13478 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13482 * Check the values of dna and dnb in bn_mul_recursive before calling
13483 bn_mul_comba (a non zero value means the a or b arrays do not contain
13484 n2 elements) and fallback to bn_mul_normal if either is not zero.
13488 * Fix escaping of non-ASCII characters when using the -subj option
13489 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13493 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13494 form for "surname", serialNumber has no short form.
13495 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13496 therefore remove "mail" short name for "internet 7".
13497 The OID for unique identifiers in X509 certificates is
13498 x500UniqueIdentifier, not uniqueIdentifier.
13499 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13503 * Add an "init" command to the ENGINE config module and auto initialize
13504 ENGINEs. Without any "init" command the ENGINE will be initialized
13505 after all ctrl commands have been executed on it. If init=1 the
13506 ENGINE is initialized at that point (ctrls before that point are run
13507 on the uninitialized ENGINE and after on the initialized one). If
13508 init=0 then the ENGINE will not be initialized at all.
13512 * Fix the 'app_verify_callback' interface so that the user-defined
13513 argument is actually passed to the callback: In the
13514 SSL_CTX_set_cert_verify_callback() prototype, the callback
13515 declaration has been changed from
13518 int (*cb)(X509_STORE_CTX *,void *);
13519 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13520 i=s->ctx->app_verify_callback(&ctx)
13521 has been changed into
13522 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13524 To update applications using SSL_CTX_set_cert_verify_callback(),
13525 a dummy argument can be added to their callback functions.
13527 *D. K. Smetters <smetters@parc.xerox.com>*
13529 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13531 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13533 * Add and OPENSSL_LOAD_CONF define which will cause
13534 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13535 This allows older applications to transparently support certain
13536 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13537 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13538 load the config file and OPENSSL_add_all_algorithms_conf() which will
13539 always load it have also been added.
13543 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13544 Adjust NIDs and EVP layer.
13546 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13548 * Config modules support in openssl utility.
13550 Most commands now load modules from the config file,
13551 though in a few (such as version) this isn't done
13552 because it couldn't be used for anything.
13554 In the case of ca and req the config file used is
13555 the same as the utility itself: that is the -config
13556 command line option can be used to specify an
13561 * Move default behaviour from OPENSSL_config(). If appname is NULL
13562 use "openssl_conf" if filename is NULL use default openssl config file.
13566 * Add an argument to OPENSSL_config() to allow the use of an alternative
13567 config section name. Add a new flag to tolerate a missing config file
13568 and move code to CONF_modules_load_file().
13572 * Support for crypto accelerator cards from Accelerated Encryption
13573 Processing, www.aep.ie. (Use engine 'aep')
13574 The support was copied from 0.9.6c [engine] and adapted/corrected
13575 to work with the new engine framework.
13577 *AEP Inc. and Richard Levitte*
13579 * Support for SureWare crypto accelerator cards from Baltimore
13580 Technologies. (Use engine 'sureware')
13581 The support was copied from 0.9.6c [engine] and adapted
13582 to work with the new engine framework.
13586 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13587 make the newer ENGINE framework commands for the CHIL engine work.
13589 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13591 * Make it possible to produce shared libraries on ReliantUNIX.
13593 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13595 * Add the configuration target debug-linux-ppro.
13596 Make 'openssl rsa' use the general key loading routines
13597 implemented in `apps.c`, and make those routines able to
13598 handle the key format FORMAT_NETSCAPE and the variant
13601 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13603 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13605 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13607 * Add -keyform to rsautl, and document -engine.
13609 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13611 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13612 BIO_R_NO_SUCH_FILE error code rather than the generic
13613 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13617 * Add new functions
13618 ERR_peek_last_error
13619 ERR_peek_last_error_line
13620 ERR_peek_last_error_line_data.
13621 These are similar to
13623 ERR_peek_error_line
13624 ERR_peek_error_line_data,
13625 but report on the latest error recorded rather than the first one
13626 still in the error queue.
13628 *Ben Laurie, Bodo Moeller*
13630 * default_algorithms option in ENGINE config module. This allows things
13632 default_algorithms = ALL
13633 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13637 * Preliminary ENGINE config module.
13641 * New experimental application configuration code.
13645 * Change the AES code to follow the same name structure as all other
13646 symmetric ciphers, and behave the same way. Move everything to
13647 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13649 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13651 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13653 *Ben Laurie and Theo de Raadt*
13655 * Add option to output public keys in req command.
13657 *Massimiliano Pala madwolf@openca.org*
13659 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13660 (up to about 10% better than before for P-192 and P-224).
13664 * New functions/macros
13666 SSL_CTX_set_msg_callback(ctx, cb)
13667 SSL_CTX_set_msg_callback_arg(ctx, arg)
13668 SSL_set_msg_callback(ssl, cb)
13669 SSL_set_msg_callback_arg(ssl, arg)
13671 to request calling a callback function
13673 void cb(int write_p, int version, int content_type,
13674 const void *buf, size_t len, SSL *ssl, void *arg)
13676 whenever a protocol message has been completely received
13677 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13678 protocol version according to which the SSL library interprets
13679 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13680 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13681 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13682 specification (change_cipher_spec(20), alert(21), handshake(22)).
13683 'buf' and 'len' point to the actual message, 'ssl' to the
13684 SSL object, and 'arg' is the application-defined value set by
13685 SSL[_CTX]_set_msg_callback_arg().
13687 'openssl s_client' and 'openssl s_server' have new '-msg' options
13688 to enable a callback that displays all protocol messages.
13692 * Change the shared library support so shared libraries are built as
13693 soon as the corresponding static library is finished, and thereby get
13694 openssl and the test programs linked against the shared library.
13695 This still only happens when the keyword "shard" has been given to
13696 the configuration scripts.
13698 NOTE: shared library support is still an experimental thing, and
13699 backward binary compatibility is still not guaranteed.
13701 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13703 * Add support for Subject Information Access extension.
13705 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13707 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13708 additional bytes when new memory had to be allocated, not just
13709 when reusing an existing buffer.
13713 * New command line and configuration option 'utf8' for the req command.
13714 This allows field values to be specified as UTF8 strings.
13718 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13719 runs for the former and machine-readable output for the latter.
13723 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13724 of the e-mail address in the DN (i.e., it will go into a certificate
13725 extension only). The new configuration file option 'email_in_dn = no'
13726 has the same effect.
13728 *Massimiliano Pala madwolf@openca.org*
13730 * Change all functions with names starting with `des_` to be starting
13731 with `DES_` instead. Add wrappers that are compatible with libdes,
13732 but are named `_ossl_old_des_*`. Finally, add macros that map the
13733 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13734 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13735 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13738 Since we provide two compatibility mappings, the user needs to
13739 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13740 compatibility is desired. The default (i.e., when that macro
13741 isn't defined) is OpenSSL 0.9.6c compatibility.
13743 There are also macros that enable and disable the support of old
13744 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13745 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13746 are defined, the default will apply: to support the old des routines.
13748 In either case, one must include openssl/des.h to get the correct
13749 definitions. Do not try to just include openssl/des_old.h, that
13752 NOTE: This is a major break of an old API into a new one. Software
13753 authors are encouraged to switch to the `DES_` style functions. Some
13754 time in the future, des_old.h and the libdes compatibility functions
13755 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13756 default), and then completely removed.
13760 * Test for certificates which contain unsupported critical extensions.
13761 If such a certificate is found during a verify operation it is
13762 rejected by default: this behaviour can be overridden by either
13763 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13764 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13765 X509_supported_extension() has also been added which returns 1 if a
13766 particular extension is supported.
13770 * Modify the behaviour of EVP cipher functions in similar way to digests
13771 to retain compatibility with existing code.
13775 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13776 compatibility with existing code. In particular the 'ctx' parameter does
13777 not have to be to be initialized before the call to EVP_DigestInit() and
13778 it is tidied up after a call to EVP_DigestFinal(). New function
13779 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13780 EVP_MD_CTX_copy() changed to not require the destination to be
13781 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13782 requires the destination to be valid.
13784 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13785 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13789 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13790 so that complete 'Handshake' protocol structures are kept in memory
13791 instead of overwriting 'msg_type' and 'length' with 'body' data.
13795 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13797 *Massimo Santin via Richard Levitte*
13799 * Major restructuring to the underlying ENGINE code. This includes
13800 reduction of linker bloat, separation of pure "ENGINE" manipulation
13801 (initialisation, etc) from functionality dealing with implementations
13802 of specific crypto interfaces. This change also introduces integrated
13803 support for symmetric ciphers and digest implementations - so ENGINEs
13804 can now accelerate these by providing EVP_CIPHER and EVP_MD
13805 implementations of their own. This is detailed in
13806 [crypto/engine/README.md](crypto/engine/README.md)
13807 as it couldn't be adequately described here. However, there are a few
13808 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13809 were changed in the original introduction of ENGINE code have now
13810 reverted back - the hooking from this code to ENGINE is now a good
13811 deal more passive and at run-time, operations deal directly with
13812 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13813 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13814 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13815 they were not being used by the framework as there is no concept of a
13816 BIGNUM_METHOD and they could not be generalised to the new
13817 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13818 ENGINE_cpy() has been removed as it cannot be consistently defined in
13823 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13827 * Change mkdef.pl to sort symbols that get the same entry number,
13828 and make sure the automatically generated functions `ERR_load_*`
13829 become part of libeay.num as well.
13833 * New function SSL_renegotiate_pending(). This returns true once
13834 renegotiation has been requested (either SSL_renegotiate() call
13835 or HelloRequest/ClientHello received from the peer) and becomes
13836 false once a handshake has been completed.
13837 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13838 sends a HelloRequest, but does not ensure that a handshake takes
13839 place. SSL_renegotiate_pending() is useful for checking if the
13840 client has followed the request.)
13844 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13845 By default, clients may request session resumption even during
13846 renegotiation (if session ID contexts permit); with this option,
13847 session resumption is possible only in the first handshake.
13849 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13850 more bits available for options that should not be part of
13851 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13855 * Add some demos for certificate and certificate request creation.
13859 * Make maximum certificate chain size accepted from the peer application
13860 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13861 "Douglas E. Engert" <deengert@anl.gov>.
13865 * Add support for shared libraries for Unixware-7
13866 (Boyd Lynn Gerber <gerberb@zenez.com>).
13870 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13871 be done prior to destruction. Use this to unload error strings from
13872 ENGINEs that load their own error strings. NB: This adds two new API
13873 functions to "get" and "set" this destroy handler in an ENGINE.
13877 * Alter all existing ENGINE implementations (except "openssl" and
13878 "openbsd") to dynamically instantiate their own error strings. This
13879 makes them more flexible to be built both as statically-linked ENGINEs
13880 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13881 Also, add stub code to each that makes building them as self-contained
13882 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13886 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13887 implementations into applications that are completely implemented in
13888 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13889 commands that can be used to configure what shared-library to load and
13890 to control aspects of the way it is handled. Also, made an update to
13891 the [README-Engine.md](README-Engine.md) file
13892 that brings its information up-to-date and
13893 provides some information and instructions on the "dynamic" ENGINE
13894 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13898 * Make it possible to unload ranges of ERR strings with a new
13899 "ERR_unload_strings" function.
13903 * Add a copy() function to EVP_MD.
13907 * Make EVP_MD routines take a context pointer instead of just the
13908 md_data void pointer.
13912 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13913 that the digest can only process a single chunk of data
13914 (typically because it is provided by a piece of
13915 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13916 is only going to provide a single chunk of data, and hence the
13917 framework needn't accumulate the data for oneshot drivers.
13921 * As with "ERR", make it possible to replace the underlying "ex_data"
13922 functions. This change also alters the storage and management of global
13923 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13924 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13925 index counters. The API functions that use this state have been changed
13926 to take a "class_index" rather than pointers to the class's local STACK
13927 and counter, and there is now an API function to dynamically create new
13928 classes. This centralisation allows us to (a) plug a lot of the
13929 thread-safety problems that existed, and (b) makes it possible to clean
13930 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13931 such data would previously have always leaked in application code and
13932 workarounds were in place to make the memory debugging turn a blind eye
13933 to it. Application code that doesn't use this new function will still
13934 leak as before, but their memory debugging output will announce it now
13935 rather than letting it slide.
13937 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13938 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13939 has a return value to indicate success or failure.
13943 * Make it possible to replace the underlying "ERR" functions such that the
13944 global state (2 LHASH tables and 2 locks) is only used by the "default"
13945 implementation. This change also adds two functions to "get" and "set"
13946 the implementation prior to it being automatically set the first time
13947 any other ERR function takes place. Ie. an application can call "get",
13948 pass the return value to a module it has just loaded, and that module
13949 can call its own "set" function using that value. This means the
13950 module's "ERR" operations will use (and modify) the error state in the
13951 application and not in its own statically linked copy of OpenSSL code.
13955 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13956 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13957 the operation, and provides a more encapsulated way for external code
13958 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13959 to use these functions rather than manually incrementing the counts.
13961 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13965 * Add EVP test program.
13969 * Add symmetric cipher support to ENGINE. Expect the API to change!
13973 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13974 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13975 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13976 These allow a CRL to be built without having to access X509_CRL fields
13977 directly. Modify 'ca' application to use new functions.
13981 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13982 bug workarounds. Rollback attack detection is a security feature.
13983 The problem will only arise on OpenSSL servers when TLSv1 is not
13984 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13985 Software authors not wanting to support TLSv1 will have special reasons
13986 for their choice and can explicitly enable this option.
13988 *Bodo Moeller, Lutz Jaenicke*
13990 * Rationalise EVP so it can be extended: don't include a union of
13991 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13992 (similar to those existing for EVP_CIPHER_CTX).
13997 EVP_MD_CTX_init(&md); /* new function call */
13998 EVP_DigestInit(&md, EVP_sha1());
13999 EVP_DigestUpdate(&md, in, len);
14000 EVP_DigestFinal(&md, out, NULL);
14001 EVP_MD_CTX_cleanup(&md); /* new function call */
14005 * Make DES key schedule conform to the usual scheme, as well as
14006 correcting its structure. This means that calls to DES functions
14007 now have to pass a pointer to a des_key_schedule instead of a
14008 plain des_key_schedule (which was actually always a pointer
14011 des_key_schedule ks;
14013 des_set_key_checked(..., &ks);
14014 des_ncbc_encrypt(..., &ks, ...);
14016 (Note that a later change renames 'des_...' into 'DES_...'.)
14020 * Initial reduction of linker bloat: the use of some functions, such as
14021 PEM causes large amounts of unused functions to be linked in due to
14022 poor organisation. For example pem_all.c contains every PEM function
14023 which has a knock on effect of linking in large amounts of (unused)
14024 ASN1 code. Grouping together similar functions and splitting unrelated
14025 functions prevents this.
14029 * Cleanup of EVP macros.
14033 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
14034 correct `_ecb suffix`.
14038 * Add initial OCSP responder support to ocsp application. The
14039 revocation information is handled using the text based index
14040 use by the ca application. The responder can either handle
14041 requests generated internally, supplied in files (for example
14042 via a CGI script) or using an internal minimal server.
14046 * Add configuration choices to get zlib compression for TLS.
14050 * Changes to Kerberos SSL for RFC 2712 compliance:
14051 1. Implemented real KerberosWrapper, instead of just using
14052 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
14053 2. Implemented optional authenticator field of KerberosWrapper.
14055 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
14056 and authenticator structs; see crypto/krb5/.
14058 Generalized Kerberos calls to support multiple Kerberos libraries.
14059 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
14060 via Richard Levitte*
14062 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
14063 already does with RSA. testdsa.h now has 'priv_key/pub_key'
14064 values for each of the key sizes rather than having just
14065 parameters (and 'speed' generating keys each time).
14069 * Speed up EVP routines.
14072 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
14073 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
14074 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
14075 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
14077 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
14078 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
14079 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
14082 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
14084 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
14088 * Added the OS2-EMX target.
14090 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
14092 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
14093 New functions to support `NCONF` routines in extension code.
14094 New function `CONF_set_nconf()`
14095 to allow functions which take an `NCONF` to also handle the old `LHASH`
14096 structure: this means that the old `CONF` compatible routines can be
14097 retained (in particular w.rt. extensions) without having to duplicate the
14098 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
14102 * Enhance the general user interface with mechanisms for inner control
14103 and with possibilities to have yes/no kind of prompts.
14107 * Change all calls to low-level digest routines in the library and
14108 applications to use EVP. Add missing calls to HMAC_cleanup() and
14109 don't assume HMAC_CTX can be copied using memcpy().
14111 *Verdon Walker <VWalker@novell.com>, Steve Henson*
14113 * Add the possibility to control engines through control names but with
14114 arbitrary arguments instead of just a string.
14115 Change the key loaders to take a UI_METHOD instead of a callback
14116 function pointer. NOTE: this breaks binary compatibility with earlier
14117 versions of OpenSSL [engine].
14118 Adapt the nCipher code for these new conditions and add a card insertion
14123 * Enhance the general user interface with mechanisms to better support
14124 dialog box interfaces, application-defined prompts, the possibility
14125 to use defaults (for example default passwords from somewhere else)
14126 and interrupts/cancellations.
14130 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14131 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14135 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14136 tidy up some unnecessarily weird code in 'sk_new()').
14138 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14140 * Change the key loading routines for ENGINEs to use the same kind
14141 callback (pem_password_cb) as all other routines that need this
14146 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14147 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14148 than this minimum value is recommended.
14152 * New random seeder for OpenVMS, using the system process statistics
14153 that are easily reachable.
14157 * Windows apparently can't transparently handle global
14158 variables defined in DLLs. Initialisations such as:
14160 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14162 won't compile. This is used by the any applications that need to
14163 declare their own ASN1 modules. This was fixed by adding the option
14164 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14165 needed for static libraries under Win32.
14169 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14170 setting of purpose and trust fields. New X509_STORE trust and
14171 purpose functions and tidy up setting in other SSL functions.
14175 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14176 structure. These are inherited by X509_STORE_CTX when it is
14177 initialised. This allows various defaults to be set in the
14178 X509_STORE structure (such as flags for CRL checking and custom
14179 purpose or trust settings) for functions which only use X509_STORE_CTX
14180 internally such as S/MIME.
14182 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14183 trust settings if they are not set in X509_STORE. This allows X509_STORE
14184 purposes and trust (in S/MIME for example) to override any set by default.
14186 Add command line options for CRL checking to smime, s_client and s_server
14191 * Initial CRL based revocation checking. If the CRL checking flag(s)
14192 are set then the CRL is looked up in the X509_STORE structure and
14193 its validity and signature checked, then if the certificate is found
14194 in the CRL the verify fails with a revoked error.
14196 Various new CRL related callbacks added to X509_STORE_CTX structure.
14198 Command line options added to 'verify' application to support this.
14200 This needs some additional work, such as being able to handle multiple
14201 CRLs with different times, extension based lookup (rather than just
14202 by subject name) and ultimately more complete V2 CRL extension
14207 * Add a general user interface API (crypto/ui/). This is designed
14208 to replace things like des_read_password and friends (backward
14209 compatibility functions using this new API are provided).
14210 The purpose is to remove prompting functions from the DES code
14211 section as well as provide for prompting through dialog boxes in
14212 a window system and the like.
14216 * Add "ex_data" support to ENGINE so implementations can add state at a
14217 per-structure level rather than having to store it globally.
14221 * Make it possible for ENGINE structures to be copied when retrieved by
14222 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14223 This causes the "original" ENGINE structure to act like a template,
14224 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14225 operational state can be localised to each ENGINE structure, despite the
14226 fact they all share the same "methods". New ENGINE structures returned in
14227 this case have no functional references and the return value is the single
14228 structural reference. This matches the single structural reference returned
14229 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14234 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14235 needs to match any other type at all we need to manually clear the
14240 * Changes to the "openssl engine" utility to include;
14241 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14242 about an ENGINE's available control commands.
14243 - executing control commands from command line arguments using the
14244 '-pre' and '-post' switches. '-post' is only used if '-t' is
14245 specified and the ENGINE is successfully initialised. The syntax for
14246 the individual commands are colon-separated, for example;
14247 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14251 * New dynamic control command support for ENGINEs. ENGINEs can now
14252 declare their own commands (numbers), names (strings), descriptions,
14253 and input types for run-time discovery by calling applications. A
14254 subset of these commands are implicitly classed as "executable"
14255 depending on their input type, and only these can be invoked through
14256 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14257 can be based on user input, config files, etc). The distinction is
14258 that "executable" commands cannot return anything other than a boolean
14259 result and can only support numeric or string input, whereas some
14260 discoverable commands may only be for direct use through
14261 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14262 pointers, or other custom uses. The "executable" commands are to
14263 support parameterisations of ENGINE behaviour that can be
14264 unambiguously defined by ENGINEs and used consistently across any
14265 OpenSSL-based application. Commands have been added to all the
14266 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14267 control over shared-library paths without source code alterations.
14271 * Changed all ENGINE implementations to dynamically allocate their
14272 ENGINEs rather than declaring them statically. Apart from this being
14273 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14274 this also allows the implementations to compile without using the
14275 internal engine_int.h header.
14279 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14280 'const' value. Any code that should be able to modify a RAND_METHOD
14281 should already have non-const pointers to it (ie. they should only
14282 modify their own ones).
14286 * Made a variety of little tweaks to the ENGINE code.
14287 - "atalla" and "ubsec" string definitions were moved from header files
14288 to C code. "nuron" string definitions were placed in variables
14289 rather than hard-coded - allowing parameterisation of these values
14290 later on via ctrl() commands.
14291 - Removed unused "#if 0"'d code.
14292 - Fixed engine list iteration code so it uses ENGINE_free() to release
14293 structural references.
14294 - Constified the RAND_METHOD element of ENGINE structures.
14295 - Constified various get/set functions as appropriate and added
14296 missing functions (including a catch-all ENGINE_cpy that duplicates
14297 all ENGINE values onto a new ENGINE except reference counts/state).
14298 - Removed NULL parameter checks in get/set functions. Setting a method
14299 or function to NULL is a way of cancelling out a previously set
14300 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14301 and doesn't justify the extra error symbols and code.
14302 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14303 flags from engine_int.h to engine.h.
14304 - Changed prototypes for ENGINE handler functions (init(), finish(),
14305 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14309 * Implement binary inversion algorithm for BN_mod_inverse in addition
14310 to the algorithm using long division. The binary algorithm can be
14311 used only if the modulus is odd. On 32-bit systems, it is faster
14312 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14313 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14314 up to 450 bits. In 64-bit environments, the binary algorithm
14315 appears to be advantageous for much longer moduli; here we use it
14316 for moduli up to 2048 bits.
14320 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14321 could not support the combine flag in choice fields.
14325 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14326 extensions from a certificate request to the certificate.
14330 * Allow multiple 'certopt' and 'nameopt' options to be separated
14331 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14332 file: this allows the display of the certificate about to be
14333 signed to be customised, to allow certain fields to be included
14334 or excluded and extension details. The old system didn't display
14335 multicharacter strings properly, omitted fields not in the policy
14336 and couldn't display additional details such as extensions.
14340 * Function EC_POINTs_mul for multiple scalar multiplication
14341 of an arbitrary number of elliptic curve points
14342 \sum scalars[i]*points[i],
14343 optionally including the generator defined for the EC_GROUP:
14344 scalar*generator + \sum scalars[i]*points[i].
14346 EC_POINT_mul is a simple wrapper function for the typical case
14347 that the point list has just one item (besides the optional
14352 * First EC_METHODs for curves over GF(p):
14354 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14355 operations and provides various method functions that can also
14356 operate with faster implementations of modular arithmetic.
14358 EC_GFp_mont_method() reuses most functions that are part of
14359 EC_GFp_simple_method, but uses Montgomery arithmetic.
14361 *Bodo Moeller; point addition and point doubling
14362 implementation directly derived from source code provided by
14363 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14365 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14366 crypto/ec/ec_lib.c):
14368 Curves are EC_GROUP objects (with an optional group generator)
14369 based on EC_METHODs that are built into the library.
14371 Points are EC_POINT objects based on EC_GROUP objects.
14373 Most of the framework would be able to handle curves over arbitrary
14374 finite fields, but as there are no obvious types for fields other
14375 than GF(p), some functions are limited to that for now.
14379 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14380 that the file contains a complete HTTP response.
14384 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14385 change the def and num file printf format specifier from "%-40sXXX"
14386 to "%-39s XXX". The latter will always guarantee a space after the
14387 field while the former will cause them to run together if the field
14388 is 40 of more characters long.
14392 * Constify the cipher and digest 'method' functions and structures
14393 and modify related functions to take constant EVP_MD and EVP_CIPHER
14398 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14399 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14403 * Modify `EVP_Digest*()` routines so they now return values. Although the
14404 internal software routines can never fail additional hardware versions
14409 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14411 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14412 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14415 ERR_R_NESTED_ASN1_ERROR
14417 ERR_R_MISSING_ASN1_EOS
14418 were 4 .. 9, conflicting with
14419 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14421 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14422 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14424 Add new error code 'ERR_R_INTERNAL_ERROR'.
14428 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14433 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14434 sets the subject name for a new request or supersedes the
14435 subject name in a given request. Formats that can be parsed are
14436 'CN=Some Name, OU=myOU, C=IT'
14438 'CN=Some Name/OU=myOU/C=IT'.
14440 Add options '-batch' and '-verbose' to 'openssl req'.
14442 *Massimiliano Pala <madwolf@hackmasters.net>*
14444 * Introduce the possibility to access global variables through
14445 functions on platform were that's the best way to handle exporting
14446 global variables in shared libraries. To enable this functionality,
14447 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14448 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14449 is normally done by Configure or something similar).
14451 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14452 in the source file (foo.c) like this:
14454 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14455 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14457 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14458 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14460 OPENSSL_DECLARE_GLOBAL(int,foo);
14461 #define foo OPENSSL_GLOBAL_REF(foo)
14462 OPENSSL_DECLARE_GLOBAL(double,bar);
14463 #define bar OPENSSL_GLOBAL_REF(bar)
14465 The #defines are very important, and therefore so is including the
14466 header file everywhere where the defined globals are used.
14468 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14469 of ASN.1 items, but that structure is a bit different.
14471 The largest change is in util/mkdef.pl which has been enhanced with
14472 better and easier to understand logic to choose which symbols should
14473 go into the Windows .def files as well as a number of fixes and code
14474 cleanup (among others, algorithm keywords are now sorted
14475 lexicographically to avoid constant rewrites).
14479 * In BN_div() keep a copy of the sign of 'num' before writing the
14480 result to 'rm' because if rm==num the value will be overwritten
14481 and produce the wrong result if 'num' is negative: this caused
14482 problems with BN_mod() and BN_nnmod().
14486 * Function OCSP_request_verify(). This checks the signature on an
14487 OCSP request and verifies the signer certificate. The signer
14488 certificate is just checked for a generic purpose and OCSP request
14493 * Add OCSP_check_validity() function to check the validity of OCSP
14494 responses. OCSP responses are prepared in real time and may only
14495 be a few seconds old. Simply checking that the current time lies
14496 between thisUpdate and nextUpdate max reject otherwise valid responses
14497 caused by either OCSP responder or client clock inaccuracy. Instead
14498 we allow thisUpdate and nextUpdate to fall within a certain period of
14499 the current time. The age of the response can also optionally be
14500 checked. Two new options -validity_period and -status_age added to
14505 * If signature or public key algorithm is unrecognized print out its
14506 OID rather that just UNKNOWN.
14510 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14511 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14512 ID to be generated from the issuer certificate alone which can then be
14513 passed to OCSP_id_issuer_cmp().
14517 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14518 ASN1 modules to export functions returning ASN1_ITEM pointers
14519 instead of the ASN1_ITEM structures themselves. This adds several
14520 new macros which allow the underlying ASN1 function/structure to
14521 be accessed transparently. As a result code should not use ASN1_ITEM
14522 references directly (such as &X509_it) but instead use the relevant
14523 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14524 use of the new ASN1 code on platforms where exporting structures
14525 is problematical (for example in shared libraries) but exporting
14526 functions returning pointers to structures is not.
14530 * Add support for overriding the generation of SSL/TLS session IDs.
14531 These callbacks can be registered either in an SSL_CTX or per SSL.
14532 The purpose of this is to allow applications to control, if they wish,
14533 the arbitrary values chosen for use as session IDs, particularly as it
14534 can be useful for session caching in multiple-server environments. A
14535 command-line switch for testing this (and any client code that wishes
14536 to use such a feature) has been added to "s_server".
14538 *Geoff Thorpe, Lutz Jaenicke*
14540 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14541 of the form `#if defined(...) || defined(...) || ...` and
14542 `#if !defined(...) && !defined(...) && ...`. This also avoids
14543 the growing number of special cases it was previously handling.
14547 * Make all configuration macros available for application by making
14548 sure they are available in opensslconf.h, by giving them names starting
14549 with `OPENSSL_` to avoid conflicts with other packages and by making
14550 sure e_os2.h will cover all platform-specific cases together with
14552 Additionally, it is now possible to define configuration/platform-
14553 specific names (called "system identities"). In the C code, these
14554 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14555 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14556 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14561 * New option -set_serial to 'req' and 'x509' this allows the serial
14562 number to use to be specified on the command line. Previously self
14563 signed certificates were hard coded with serial number 0 and the
14564 CA options of 'x509' had to use a serial number in a file which was
14569 * New options to 'ca' utility to support V2 CRL entry extensions.
14570 Currently CRL reason, invalidity date and hold instruction are
14571 supported. Add new CRL extensions to V3 code and some new objects.
14575 * New function EVP_CIPHER_CTX_set_padding() this is used to
14576 disable standard block padding (aka PKCS#5 padding) in the EVP
14577 API, which was previously mandatory. This means that the data is
14578 not padded in any way and so the total length much be a multiple
14579 of the block size, otherwise an error occurs.
14583 * Initial (incomplete) OCSP SSL support.
14587 * New function OCSP_parse_url(). This splits up a URL into its host,
14588 port and path components: primarily to parse OCSP URLs. New -url
14589 option to ocsp utility.
14593 * New nonce behavior. The return value of OCSP_check_nonce() now
14594 reflects the various checks performed. Applications can decide
14595 whether to tolerate certain situations such as an absent nonce
14596 in a response when one was present in a request: the ocsp application
14597 just prints out a warning. New function OCSP_add1_basic_nonce()
14598 this is to allow responders to include a nonce in a response even if
14599 the request is nonce-less.
14603 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14604 skipped when using openssl x509 multiple times on a single input file,
14605 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14609 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14610 set string type: to handle setting ASN1_TIME structures. Fix ca
14611 utility to correctly initialize revocation date of CRLs.
14615 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14616 the clients preferred ciphersuites and rather use its own preferences.
14617 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14618 Internet Explorer by ensuring unchanged hash method during stepup.
14619 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14623 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14624 to aes and add a new 'exist' option to print out symbols that don't
14629 * Additional options to ocsp utility to allow flags to be set and
14630 additional certificates supplied.
14634 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14635 OCSP client a number of certificate to only verify the response
14640 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14641 handle the new API. Currently only ECB, CBC modes supported. Add new
14644 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14645 Encryption Standard (AES) Ciphersuites for Transport Layer
14646 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14647 not enabled by default and were not part of the "ALL" ciphersuite
14648 alias because they were not yet official; they could be
14649 explicitly requested by specifying the "AESdraft" ciphersuite
14650 group alias. In the final release of OpenSSL 0.9.7, the group
14651 alias is called "AES" and is part of "ALL".)
14653 *Ben Laurie, Steve Henson, Bodo Moeller*
14655 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14656 request to response.
14660 * Functions for OCSP responders. OCSP_request_onereq_count(),
14661 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14662 extract information from a certificate request. OCSP_response_create()
14663 creates a response and optionally adds a basic response structure.
14664 OCSP_basic_add1_status() adds a complete single response to a basic
14665 response and returns the OCSP_SINGLERESP structure just added (to allow
14666 extensions to be included for example). OCSP_basic_add1_cert() adds a
14667 certificate to a basic response and OCSP_basic_sign() signs a basic
14668 response with various flags. New helper functions ASN1_TIME_check()
14669 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14670 (converts ASN1_TIME to GeneralizedTime).
14674 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14675 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14676 structure from a certificate. X509_pubkey_digest() digests the public_key
14677 contents: this is used in various key identifiers.
14681 * Make sk_sort() tolerate a NULL argument.
14683 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14685 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14686 passed by the function are trusted implicitly. If any of them signed the
14687 response then it is assumed to be valid and is not verified.
14691 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14692 to data. This was previously part of the PKCS7 ASN1 code. This
14693 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14694 *Steve Henson, reported by Kenneth R. Robinette
14695 <support@securenetterm.com>*
14697 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14698 routines: without these tracing memory leaks is very painful.
14699 Fix leaks in PKCS12 and PKCS7 routines.
14703 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14704 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14705 effectively meant GeneralizedTime would never be used. Now it
14706 is initialised to -1 but X509_time_adj() now has to check the value
14707 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14708 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14709 *Steve Henson, reported by Kenneth R. Robinette
14710 <support@securenetterm.com>*
14712 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14713 result in a zero length in the ASN1_INTEGER structure which was
14714 not consistent with the structure when d2i_ASN1_INTEGER() was used
14715 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14716 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14717 where it did not print out a minus for negative ASN1_INTEGER.
14721 * Add summary printout to ocsp utility. The various functions which
14722 convert status values to strings have been renamed to:
14723 OCSP_response_status_str(), OCSP_cert_status_str() and
14724 OCSP_crl_reason_str() and are no longer static. New options
14725 to verify nonce values and to disable verification. OCSP response
14726 printout format cleaned up.
14730 * Add additional OCSP certificate checks. These are those specified
14731 in RFC2560. This consists of two separate checks: the CA of the
14732 certificate being checked must either be the OCSP signer certificate
14733 or the issuer of the OCSP signer certificate. In the latter case the
14734 OCSP signer certificate must contain the OCSP signing extended key
14735 usage. This check is performed by attempting to match the OCSP
14736 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14737 in the OCSP_CERTID structures of the response.
14741 * Initial OCSP certificate verification added to OCSP_basic_verify()
14742 and related routines. This uses the standard OpenSSL certificate
14743 verify routines to perform initial checks (just CA validity) and
14744 to obtain the certificate chain. Then additional checks will be
14745 performed on the chain. Currently the root CA is checked to see
14746 if it is explicitly trusted for OCSP signing. This is used to set
14747 a root CA as a global signing root: that is any certificate that
14748 chains to that CA is an acceptable OCSP signing certificate.
14752 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14753 extensions from a separate configuration file.
14754 As when reading extensions from the main configuration file,
14755 the '-extensions ...' option may be used for specifying the
14758 *Massimiliano Pala <madwolf@comune.modena.it>*
14760 * New OCSP utility. Allows OCSP requests to be generated or
14761 read. The request can be sent to a responder and the output
14762 parsed, outputted or printed in text form. Not complete yet:
14763 still needs to check the OCSP response validity.
14767 * New subcommands for 'openssl ca':
14768 `openssl ca -status <serial>` prints the status of the cert with
14769 the given serial number (according to the index file).
14770 `openssl ca -updatedb` updates the expiry status of certificates
14773 *Massimiliano Pala <madwolf@comune.modena.it>*
14775 * New '-newreq-nodes' command option to CA.pl. This is like
14776 '-newreq', but calls 'openssl req' with the '-nodes' option
14777 so that the resulting key is not encrypted.
14779 *Damien Miller <djm@mindrot.org>*
14781 * New configuration for the GNU Hurd.
14783 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14785 * Initial code to implement OCSP basic response verify. This
14786 is currently incomplete. Currently just finds the signer's
14787 certificate and verifies the signature on the response.
14791 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14792 value of OPENSSLDIR. This is available via the new '-d' option
14793 to 'openssl version', and is also included in 'openssl version -a'.
14797 * Allowing defining memory allocation callbacks that will be given
14798 file name and line number information in additional arguments
14799 (a `const char*` and an int). The basic functionality remains, as
14800 well as the original possibility to just replace malloc(),
14801 realloc() and free() by functions that do not know about these
14802 additional arguments. To register and find out the current
14803 settings for extended allocation functions, the following
14804 functions are provided:
14806 CRYPTO_set_mem_ex_functions
14807 CRYPTO_set_locked_mem_ex_functions
14808 CRYPTO_get_mem_ex_functions
14809 CRYPTO_get_locked_mem_ex_functions
14811 These work the same way as CRYPTO_set_mem_functions and friends.
14812 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14813 extended allocation function is enabled.
14814 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14815 a conventional allocation function is enabled.
14817 *Richard Levitte, Bodo Moeller*
14819 * Finish off removing the remaining LHASH function pointer casts.
14820 There should no longer be any prototype-casting required when using
14821 the LHASH abstraction, and any casts that remain are "bugs". See
14822 the callback types and macros at the head of lhash.h for details
14823 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14827 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14828 If /dev/[u]random devices are not available or do not return enough
14829 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14831 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14832 /etc/entropy will be queried once each in this sequence, querying stops
14833 when enough entropy was collected without querying more sockets.
14837 * Change the Unix RAND_poll() variant to be able to poll several
14838 random devices, as specified by DEVRANDOM, until a sufficient amount
14839 of data has been collected. We spend at most 10 ms on each file
14840 (select timeout) and read in non-blocking mode. DEVRANDOM now
14841 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14842 (previously it was just the string "/dev/urandom"), so on typical
14843 platforms the 10 ms delay will never occur.
14844 Also separate out the Unix variant to its own file, rand_unix.c.
14845 For VMS, there's a currently-empty rand_vms.c.
14849 * Move OCSP client related routines to ocsp_cl.c. These
14850 provide utility functions which an application needing
14851 to issue a request to an OCSP responder and analyse the
14852 response will typically need: as opposed to those which an
14853 OCSP responder itself would need which will be added later.
14855 OCSP_request_sign() signs an OCSP request with an API similar
14856 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14857 response. OCSP_response_get1_basic() extracts basic response
14858 from response. OCSP_resp_find_status(): finds and extracts status
14859 information from an OCSP_CERTID structure (which will be created
14860 when the request structure is built). These are built from lower
14861 level functions which work on OCSP_SINGLERESP structures but
14862 won't normally be used unless the application wishes to examine
14863 extensions in the OCSP response for example.
14865 Replace nonce routines with a pair of functions.
14866 OCSP_request_add1_nonce() adds a nonce value and optionally
14867 generates a random value. OCSP_check_nonce() checks the
14868 validity of the nonce in an OCSP response.
14872 * Change function OCSP_request_add() to OCSP_request_add0_id().
14873 This doesn't copy the supplied OCSP_CERTID and avoids the
14874 need to free up the newly created id. Change return type
14875 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14876 This can then be used to add extensions to the request.
14877 Deleted OCSP_request_new(), since most of its functionality
14878 is now in OCSP_REQUEST_new() (and the case insensitive name
14879 clash) apart from the ability to set the request name which
14880 will be added elsewhere.
14884 * Update OCSP API. Remove obsolete extensions argument from
14885 various functions. Extensions are now handled using the new
14886 OCSP extension code. New simple OCSP HTTP function which
14887 can be used to send requests and parse the response.
14891 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14892 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14893 uses the special reorder version of SET OF to sort the attributes
14894 and reorder them to match the encoded order. This resolves a long
14895 standing problem: a verify on a PKCS7 structure just after signing
14896 it used to fail because the attribute order did not match the
14897 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14898 it uses the received order. This is necessary to tolerate some broken
14899 software that does not order SET OF. This is handled by encoding
14900 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14901 to produce the required SET OF.
14905 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14906 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14907 files to get correct declarations of the ASN.1 item variables.
14911 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14912 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14913 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14914 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14915 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14916 ASN1_ITEM and no wrapper functions.
14920 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14921 replace the old function pointer based I/O routines. Change most of
14922 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14926 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14927 lines, recognize more "algorithms" that can be deselected, and make
14928 it complain about algorithm deselection that isn't recognised.
14932 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14933 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14934 to use new functions. Add NO_ASN1_OLD which can be set to remove
14935 some old style ASN1 functions: this can be used to determine if old
14936 code will still work when these eventually go away.
14940 * New extension functions for OCSP structures, these follow the
14941 same conventions as certificates and CRLs.
14945 * New function X509V3_add1_i2d(). This automatically encodes and
14946 adds an extension. Its behaviour can be customised with various
14947 flags to append, replace or delete. Various wrappers added for
14948 certificates and CRLs.
14952 * Fix to avoid calling the underlying ASN1 print routine when
14953 an extension cannot be parsed. Correct a typo in the
14954 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14958 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14959 entries for variables.
14963 * Add functionality to `apps/openssl.c` for detecting locking
14964 problems: As the program is single-threaded, all we have
14965 to do is register a locking callback using an array for
14966 storing which locks are currently held by the program.
14970 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14971 SSL_get_ex_data_X509_STORE_idx(), which is used in
14972 ssl_verify_cert_chain() and thus can be called at any time
14973 during TLS/SSL handshakes so that thread-safety is essential.
14974 Unfortunately, the ex_data design is not at all suited
14975 for multi-threaded use, so it probably should be abolished.
14979 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14981 *Broadcom, tweaked and integrated by Geoff Thorpe*
14983 * Move common extension printing code to new function
14984 X509V3_print_extensions(). Reorganise OCSP print routines and
14985 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14989 * New function X509_signature_print() to remove duplication in some
14994 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14995 set (this was treated exactly the same as SET OF previously). This
14996 is used to reorder the STACK representing the structure to match the
14997 encoding. This will be used to get round a problem where a PKCS7
14998 structure which was signed could not be verified because the STACK
14999 order did not reflect the encoded order.
15003 * Reimplement the OCSP ASN1 module using the new code.
15007 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
15008 for its ASN1 operations. The old style function pointers still exist
15009 for now but they will eventually go away.
15013 * Merge in replacement ASN1 code from the ASN1 branch. This almost
15014 completely replaces the old ASN1 functionality with a table driven
15015 encoder and decoder which interprets an ASN1_ITEM structure describing
15016 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
15017 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
15018 has also been converted to the new form.
15022 * Change BN_mod_exp_recp so that negative moduli are tolerated
15023 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
15024 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
15025 for negative moduli.
15029 * Fix BN_uadd and BN_usub: Always return non-negative results instead
15030 of not touching the result's sign bit.
15034 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
15039 * Changed the LHASH code to use prototypes for callbacks, and created
15040 macros to declare and implement thin (optionally static) functions
15041 that provide type-safety and avoid function pointer casting for the
15042 type-specific callbacks.
15046 * Added Kerberos Cipher Suites to be used with TLS, as written in
15048 *Veers Staats <staatsvr@asc.hpc.mil>,
15049 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
15051 * Reformat the FAQ so the different questions and answers can be divided
15052 in sections depending on the subject.
15056 * Have the zlib compression code load ZLIB.DLL dynamically under
15061 * New function BN_mod_sqrt for computing square roots modulo a prime
15062 (using the probabilistic Tonelli-Shanks algorithm unless
15063 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
15064 be handled deterministically).
15066 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15068 * Make BN_mod_inverse faster by explicitly handling small quotients
15069 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
15070 512 bits], about 30% for larger ones [1024 or 2048 bits].)
15074 * New function BN_kronecker.
15078 * Fix BN_gcd so that it works on negative inputs; the result is
15079 positive unless both parameters are zero.
15080 Previously something reasonably close to an infinite loop was
15081 possible because numbers could be growing instead of shrinking
15082 in the implementation of Euclid's algorithm.
15086 * Fix BN_is_word() and BN_is_one() macros to take into account the
15087 sign of the number in question.
15089 Fix BN_is_word(a,w) to work correctly for w == 0.
15091 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
15092 because its test if the absolute value of 'a' equals 'w'.
15093 Note that BN_abs_is_word does *not* handle w == 0 reliably;
15094 it exists mostly for use in the implementations of BN_is_zero(),
15095 BN_is_one(), and BN_is_word().
15099 * New function BN_swap.
15103 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
15104 the exponentiation functions are more likely to produce reasonable
15105 results on negative inputs.
15109 * Change BN_mod_mul so that the result is always non-negative.
15110 Previously, it could be negative if one of the factors was negative;
15111 I don't think anyone really wanted that behaviour.
15115 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
15116 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
15117 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
15118 and add new functions:
15127 BN_mod_lshift1_quick
15129 BN_mod_lshift_quick
15131 These functions always generate non-negative results.
15133 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15134 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15136 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15137 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15138 be reduced modulo `m`.
15140 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15143 The following entry accidentally appeared in the CHANGES file
15144 distributed with OpenSSL 0.9.7. The modifications described in
15145 it do *not* apply to OpenSSL 0.9.7.
15147 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15148 was actually never needed) and in BN_mul(). The removal in BN_mul()
15149 required a small change in bn_mul_part_recursive() and the addition
15150 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15151 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15152 bn_sub_words() and bn_add_words() except they take arrays with
15158 * In 'openssl passwd', verify passwords read from the terminal
15159 unless the '-salt' option is used (which usually means that
15160 verification would just waste user's time since the resulting
15161 hash is going to be compared with some given password hash)
15162 or the new '-noverify' option is used.
15164 This is an incompatible change, but it does not affect
15165 non-interactive use of 'openssl passwd' (passwords on the command
15166 line, '-stdin' option, '-in ...' option) and thus should not
15167 cause any problems.
15171 * Remove all references to RSAref, since there's no more need for it.
15175 * Make DSO load along a path given through an environment variable
15176 (SHLIB_PATH) with shl_load().
15180 * Constify the ENGINE code as a result of BIGNUM constification.
15181 Also constify the RSA code and most things related to it. In a
15182 few places, most notable in the depth of the ASN.1 code, ugly
15183 casts back to non-const were required (to be solved at a later
15188 * Make it so the openssl application has all engines loaded by default.
15192 * Constify the BIGNUM routines a little more.
15196 * Add the following functions:
15198 ENGINE_load_cswift()
15200 ENGINE_load_atalla()
15201 ENGINE_load_nuron()
15202 ENGINE_load_builtin_engines()
15204 That way, an application can itself choose if external engines that
15205 are built-in in OpenSSL shall ever be used or not. The benefit is
15206 that applications won't have to be linked with libdl or other dso
15207 libraries unless it's really needed.
15209 Changed 'openssl engine' to load all engines on demand.
15210 Changed the engine header files to avoid the duplication of some
15211 declarations (they differed!).
15215 * 'openssl engine' can now list capabilities.
15219 * Better error reporting in 'openssl engine'.
15223 * Never call load_dh_param(NULL) in s_server.
15227 * Add engine application. It can currently list engines by name and
15228 identity, and test if they are actually available.
15232 * Improve RPM specification file by forcing symbolic linking and making
15233 sure the installed documentation is also owned by root.root.
15235 *Damien Miller <djm@mindrot.org>*
15237 * Give the OpenSSL applications more possibilities to make use of
15238 keys (public as well as private) handled by engines.
15242 * Add OCSP code that comes from CertCo.
15246 * Add VMS support for the Rijndael code.
15250 * Added untested support for Nuron crypto accelerator.
15254 * Add support for external cryptographic devices. This code was
15255 previously distributed separately as the "engine" branch.
15257 *Geoff Thorpe, Richard Levitte*
15259 * Rework the filename-translation in the DSO code. It is now possible to
15260 have far greater control over how a "name" is turned into a filename
15261 depending on the operating environment and any oddities about the
15262 different shared library filenames on each system.
15266 * Support threads on FreeBSD-elf in Configure.
15270 * Fix for SHA1 assembly problem with MASM: it produces
15271 warnings about corrupt line number information when assembling
15272 with debugging information. This is caused by the overlapping
15275 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15278 NCONF_get_number() has no error checking at all. As a replacement,
15279 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15280 promoted strongly. The old NCONF_get_number is kept around for
15281 binary backward compatibility.
15282 Make it possible for methods to load from something other than a BIO,
15283 by providing a function pointer that is given a name instead of a BIO.
15284 For example, this could be used to load configuration data from an
15289 * Fix for non blocking accept BIOs. Added new I/O special reason
15290 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15291 with non blocking I/O was not possible because no retry code was
15292 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15297 * Added the beginnings of Rijndael support.
15301 * Fix for bug in DirectoryString mask setting. Add support for
15302 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15303 to allow certificate printing to more controllable, additional
15304 'certopt' option to 'x509' to allow new printing options to be
15309 * Clean old EAY MD5 hack from e_os.h.
15313 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15315 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15316 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15318 *Joe Orton, Steve Henson*
15320 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15322 * Fix additional bug revealed by the NISCC test suite:
15324 Stop bug triggering large recursion when presented with
15325 certain ASN.1 tags ([CVE-2003-0851])
15329 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15331 * Fix various bugs revealed by running the NISCC test suite:
15333 Stop out of bounds reads in the ASN1 code when presented with
15334 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15336 If verify callback ignores invalid public key errors don't try to check
15337 certificate signature with the NULL public key.
15341 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15342 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15347 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15348 extra data after the compression methods not only for TLS 1.0
15349 but also for SSL 3.0 (as required by the specification).
15351 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15353 * Change X509_certificate_type() to mark the key as exported/exportable
15354 when it's 512 *bits* long, not 512 bytes.
15358 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15360 * Countermeasure against the Klima-Pokorny-Rosa extension of
15361 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15362 a protocol version number mismatch like a decryption error
15363 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15367 * Turn on RSA blinding by default in the default implementation
15368 to avoid a timing attack. Applications that don't want it can call
15369 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15370 They would be ill-advised to do so in most cases.
15372 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15374 * Change RSA blinding code so that it works when the PRNG is not
15375 seeded (in this case, the secret RSA exponent is abused as
15376 an unpredictable seed -- if it is not unpredictable, there
15377 is no point in blinding anyway). Make RSA blinding thread-safe
15378 by remembering the creator's thread ID in rsa->blinding and
15379 having all other threads use local one-time blinding factors
15380 (this requires more computation than sharing rsa->blinding, but
15381 avoids excessive locking; and if an RSA object is not shared
15382 between threads, blinding will still be very fast).
15386 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15388 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15389 via timing by performing a MAC computation even if incorrect
15390 block cipher padding has been found. This is a countermeasure
15391 against active attacks where the attacker has to distinguish
15392 between bad padding and a MAC verification error. ([CVE-2003-0078])
15394 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15395 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15396 Martin Vuagnoux (EPFL, Ilion)*
15398 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15400 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15401 memory from its contents. This is done with a counter that will
15402 place alternating values in each byte. This can be used to solve
15403 two issues: 1) the removal of calls to memset() by highly optimizing
15404 compilers, and 2) cleansing with other values than 0, since those can
15405 be read through on certain media, for example a swap space on disk.
15409 * Bugfix: client side session caching did not work with external caching,
15410 because the session->cipher setting was not restored when reloading
15411 from the external cache. This problem was masked, when
15412 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15413 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15417 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15418 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15420 *Zeev Lieber <zeev-l@yahoo.com>*
15422 * Undo an undocumented change introduced in 0.9.6e which caused
15423 repeated calls to OpenSSL_add_all_ciphers() and
15424 OpenSSL_add_all_digests() to be ignored, even after calling
15429 * Change the default configuration reader to deal with last line not
15430 being properly terminated.
15434 * Change X509_NAME_cmp() so it applies the special rules on handling
15435 DN values that are of type PrintableString, as well as RDNs of type
15436 emailAddress where the value has the type ia5String.
15438 *stefank@valicert.com via Richard Levitte*
15440 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15441 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15442 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15443 the bitwise-OR of the two for use by the majority of applications
15444 wanting this behaviour, and update the docs. The documented
15445 behaviour and actual behaviour were inconsistent and had been
15446 changing anyway, so this is more a bug-fix than a behavioural
15449 *Geoff Thorpe, diagnosed by Nadav Har'El*
15451 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15452 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15456 * Fix initialization code race conditions in
15457 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15458 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15459 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15460 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15461 ssl2_get_cipher_by_char(),
15462 ssl3_get_cipher_by_char().
15464 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15466 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15467 the cached sessions are flushed, as the remove_cb() might use ex_data
15468 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15469 (see [openssl.org #212]).
15471 *Geoff Thorpe, Lutz Jaenicke*
15473 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15474 length, instead of the encoding length to d2i_ASN1_OBJECT.
15478 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15480 * [In 0.9.6g-engine release:]
15481 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15483 *Lynn Gazis <lgazis@rainbow.com>*
15485 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15487 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15488 and get fix the header length calculation.
15489 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15490 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15492 * Use proper error handling instead of 'assertions' in buffer
15493 overflow checks added in 0.9.6e. This prevents DoS (the
15494 assertions could call abort()).
15496 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15498 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15500 * Add various sanity checks to asn1_get_length() to reject
15501 the ASN1 length bytes if they exceed sizeof(long), will appear
15502 negative or the content length exceeds the length of the
15505 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15507 * Fix cipher selection routines: ciphers without encryption had no flags
15508 for the cipher strength set and where therefore not handled correctly
15509 by the selection routines (PR #130).
15513 * Fix EVP_dsa_sha macro.
15518 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15519 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15520 that was added in OpenSSL 0.9.6d.
15522 As the countermeasure turned out to be incompatible with some
15523 broken SSL implementations, the new option is part of SSL_OP_ALL.
15524 SSL_OP_ALL is usually employed when compatibility with weird SSL
15525 implementations is desired (e.g. '-bugs' option to 's_client' and
15526 's_server'), so the new option is automatically set in many
15531 * Changes in security patch:
15533 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15534 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15535 Air Force Materiel Command, USAF, under agreement number
15538 * Add various sanity checks to asn1_get_length() to reject
15539 the ASN1 length bytes if they exceed sizeof(long), will appear
15540 negative or the content length exceeds the length of the
15541 supplied buffer. ([CVE-2002-0659])
15543 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15545 * Assertions for various potential buffer overflows, not known to
15546 happen in practice.
15548 *Ben Laurie (CHATS)*
15550 * Various temporary buffers to hold ASCII versions of integers were
15551 too small for 64 bit platforms. ([CVE-2002-0655])
15552 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15554 * Remote buffer overflow in SSL3 protocol - an attacker could
15555 supply an oversized session ID to a client. ([CVE-2002-0656])
15557 *Ben Laurie (CHATS)*
15559 * Remote buffer overflow in SSL2 protocol - an attacker could
15560 supply an oversized client master key. ([CVE-2002-0656])
15562 *Ben Laurie (CHATS)*
15564 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15566 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15567 encoded as NULL) with id-dsa-with-sha1.
15569 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15571 * Check various `X509_...()` return values in `apps/req.c`.
15573 *Nils Larsch <nla@trustcenter.de>*
15575 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15576 an end-of-file condition would erroneously be flagged, when the CRLF
15577 was just at the end of a processed block. The bug was discovered when
15578 processing data through a buffering memory BIO handing the data to a
15579 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15580 <ptsekov@syntrex.com> and Nedelcho Stanev.
15584 * Implement a countermeasure against a vulnerability recently found
15585 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15586 before application data chunks to avoid the use of known IVs
15587 with data potentially chosen by the attacker.
15591 * Fix length checks in ssl3_get_client_hello().
15595 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15596 to prevent ssl3_read_internal() from incorrectly assuming that
15597 ssl3_read_bytes() found application data while handshake
15598 processing was enabled when in fact s->s3->in_read_app_data was
15599 merely automatically cleared during the initial handshake.
15601 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15603 * Fix object definitions for Private and Enterprise: they were not
15604 recognized in their shortname (=lowercase) representation. Extend
15605 obj_dat.pl to issue an error when using undefined keywords instead
15606 of silently ignoring the problem (Svenning Sorensen
15607 <sss@sss.dnsalias.net>).
15611 * Fix DH_generate_parameters() so that it works for 'non-standard'
15612 generators, i.e. generators other than 2 and 5. (Previously, the
15613 code did not properly initialise the 'add' and 'rem' values to
15614 BN_generate_prime().)
15616 In the new general case, we do not insist that 'generator' is
15617 actually a primitive root: This requirement is rather pointless;
15618 a generator of the order-q subgroup is just as good, if not
15623 * Map new X509 verification errors to alerts. Discovered and submitted by
15624 Tom Wu <tom@arcot.com>.
15628 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15629 returning non-zero before the data has been completely received
15630 when using non-blocking I/O.
15632 *Bodo Moeller; problem pointed out by John Hughes*
15634 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15636 *Ben Laurie, Lutz Jaenicke*
15638 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15639 Yoram Zahavi <YoramZ@gilian.com>).
15643 * Add information about CygWin 1.3 and on, and preserve proper
15644 configuration for the versions before that.
15646 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15648 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15649 check whether we deal with a copy of a session and do not delete from
15650 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15651 <izhar@checkpoint.com>.
15655 * Do not store session data into the internal session cache, if it
15656 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15657 flag is set). Proposed by Aslam <aslam@funk.com>.
15661 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15666 * [In 0.9.6d-engine release:]
15667 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15669 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15671 * Add the configuration target linux-s390x.
15673 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15675 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15676 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15677 variable as an indication that a ClientHello message has been
15678 received. As the flag value will be lost between multiple
15679 invocations of ssl3_accept when using non-blocking I/O, the
15680 function may not be aware that a handshake has actually taken
15681 place, thus preventing a new session from being added to the
15684 To avoid this problem, we now set s->new_session to 2 instead of
15685 using a local variable.
15687 *Lutz Jaenicke, Bodo Moeller*
15689 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15690 if the SSL_R_LENGTH_MISMATCH error is detected.
15692 *Geoff Thorpe, Bodo Moeller*
15694 * New 'shared_ldflag' column in Configure platform table.
15698 * Fix EVP_CIPHER_mode macro.
15700 *"Dan S. Camper" <dan@bti.net>*
15702 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15703 type, we must throw them away by setting rr->length to 0.
15705 *D P Chang <dpc@qualys.com>*
15707 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15709 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15710 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15711 worked incorrectly for those cases where range = `10..._2` and
15712 `3*range` is two bits longer than range.)
15716 * Only add signing time to PKCS7 structures if it is not already
15721 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15722 OBJ_ld_ce should be OBJ_id_ce.
15723 Also some ip-pda OIDs in crypto/objects/objects.txt were
15724 incorrect (cf. RFC 3039).
15726 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15728 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15729 returns early because it has nothing to do.
15731 *Andy Schneider <andy.schneider@bjss.co.uk>*
15733 * [In 0.9.6c-engine release:]
15734 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15736 *Andy Schneider <andy.schneider@bjss.co.uk>*
15738 * [In 0.9.6c-engine release:]
15739 Add support for Cryptographic Appliance's keyserver technology.
15740 (Use engine 'keyclient')
15742 *Cryptographic Appliances and Geoff Thorpe*
15744 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15745 is called via tools/c89.sh because arguments have to be
15746 rearranged (all '-L' options must appear before the first object
15749 *Richard Shapiro <rshapiro@abinitio.com>*
15751 * [In 0.9.6c-engine release:]
15752 Add support for Broadcom crypto accelerator cards, backported
15755 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15757 * [In 0.9.6c-engine release:]
15758 Add support for SureWare crypto accelerator cards from
15759 Baltimore Technologies. (Use engine 'sureware')
15761 *Baltimore Technologies and Mark Cox*
15763 * [In 0.9.6c-engine release:]
15764 Add support for crypto accelerator cards from Accelerated
15765 Encryption Processing, www.aep.ie. (Use engine 'aep')
15767 *AEP Inc. and Mark Cox*
15769 * Add a configuration entry for gcc on UnixWare.
15771 *Gary Benson <gbenson@redhat.com>*
15773 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15774 messages are stored in a single piece (fixed-length part and
15775 variable-length part combined) and fix various bugs found on the way.
15779 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15780 instead. BIO_gethostbyname() does not know what timeouts are
15781 appropriate, so entries would stay in cache even when they have
15783 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15785 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15786 faced with a pathologically small ClientHello fragment that does
15787 not contain client_version: Instead of aborting with an error,
15788 simply choose the highest available protocol version (i.e.,
15789 TLS 1.0 unless it is disabled). In practice, ClientHello
15790 messages are never sent like this, but this change gives us
15791 strictly correct behaviour at least for TLS.
15795 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15796 never resets s->method to s->ctx->method when called from within
15797 one of the SSL handshake functions.
15799 *Bodo Moeller; problem pointed out by Niko Baric*
15801 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15802 (sent using the client's version number) if client_version is
15803 smaller than the protocol version in use. Also change
15804 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15805 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15806 the client will at least see that alert.
15810 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15815 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15816 client receives HelloRequest while in a handshake.
15818 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15820 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15821 should end in 'break', not 'goto end' which circumvents various
15822 cleanups done in state SSL_ST_OK. But session related stuff
15823 must be disabled for SSL_ST_OK in the case that we just sent a
15826 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15827 before just sending a HelloRequest.
15829 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15831 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15832 reveal whether illegal block cipher padding was found or a MAC
15833 verification error occurred. (Neither SSLerr() codes nor alerts
15834 are directly visible to potential attackers, but the information
15835 may leak via logfiles.)
15837 Similar changes are not required for the SSL 2.0 implementation
15838 because the number of padding bytes is sent in clear for SSL 2.0,
15839 and the extra bytes are just ignored. However ssl/s2_pkt.c
15840 failed to verify that the purported number of padding bytes is in
15845 * Add OpenUNIX-8 support including shared libraries
15846 (Boyd Lynn Gerber <gerberb@zenez.com>).
15850 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15851 'wristwatch attack' using huge encoding parameters (cf.
15852 James H. Manger's CRYPTO 2001 paper). Note that the
15853 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15854 encoding parameters and hence was not vulnerable.
15858 * BN_sqr() bug fix.
15860 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15862 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15863 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15864 followed by modular reduction.
15866 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15868 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15869 equivalent based on BN_pseudo_rand() instead of BN_rand().
15873 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15874 This function was broken, as the check for a new client hello message
15875 to handle SGC did not allow these large messages.
15876 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15880 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15884 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15885 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15889 * Rework the configuration and shared library support for Tru64 Unix.
15890 The configuration part makes use of modern compiler features and
15891 still retains old compiler behavior for those that run older versions
15892 of the OS. The shared library support part includes a variant that
15893 uses the RPATH feature, and is available through the special
15894 configuration target "alpha-cc-rpath", which will never be selected
15897 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15899 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15900 with the same message size as in ssl3_get_certificate_request().
15901 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15902 messages might inadvertently be reject as too long.
15904 *Petr Lampa <lampa@fee.vutbr.cz>*
15906 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15910 * Modified SSL library such that the verify_callback that has been set
15911 specifically for an SSL object with SSL_set_verify() is actually being
15912 used. Before the change, a verify_callback set with this function was
15913 ignored and the verify_callback() set in the SSL_CTX at the time of
15914 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15915 to allow the necessary settings.
15919 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15920 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15921 done automatically (in contradiction to the requirements of the C
15922 standard). This made problems when used from OpenSSH.
15926 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15927 dh->length and always used
15929 BN_rand_range(priv_key, dh->p).
15931 BN_rand_range() is not necessary for Diffie-Hellman, and this
15932 specific range makes Diffie-Hellman unnecessarily inefficient if
15933 dh->length (recommended exponent length) is much smaller than the
15934 length of dh->p. We could use BN_rand_range() if the order of
15935 the subgroup was stored in the DH structure, but we only have
15940 BN_rand(priv_key, l, ...)
15942 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15949 RSA_eay_public_encrypt
15950 RSA_eay_private_decrypt
15951 RSA_eay_private_encrypt (signing)
15952 RSA_eay_public_decrypt (signature verification)
15954 (default implementations for RSA_public_encrypt,
15955 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15956 always reject numbers >= n.
15960 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15961 to synchronize access to 'locking_thread'. This is necessary on
15962 systems where access to 'locking_thread' (an 'unsigned long'
15963 variable) is not atomic.
15967 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15968 *before* setting the 'crypto_lock_rand' flag. The previous code had
15969 a race condition if 0 is a valid thread ID.
15971 *Travis Vitek <vitek@roguewave.com>*
15973 * Add support for shared libraries under Irix.
15975 *Albert Chin-A-Young <china@thewrittenword.com>*
15977 * Add configuration option to build on Linux on both big-endian and
15978 little-endian MIPS.
15980 *Ralf Baechle <ralf@uni-koblenz.de>*
15982 * Add the possibility to create shared libraries on HP-UX.
15986 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15988 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15989 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15990 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15991 PRNG state recovery was possible based on the output of
15992 one PRNG request appropriately sized to gain knowledge on
15993 'md' followed by enough consecutive 1-byte PRNG requests
15994 to traverse all of 'state'.
15996 1. When updating 'md_local' (the current thread's copy of 'md')
15997 during PRNG output generation, hash all of the previous
15998 'md_local' value, not just the half used for PRNG output.
16000 2. Make the number of bytes from 'state' included into the hash
16001 independent from the number of PRNG bytes requested.
16003 The first measure alone would be sufficient to avoid
16004 Markku-Juhani's attack. (Actually it had never occurred
16005 to me that the half of 'md_local' used for chaining was the
16006 half from which PRNG output bytes were taken -- I had always
16007 assumed that the secret half would be used.) The second
16008 measure makes sure that additional data from 'state' is never
16009 mixed into 'md_local' in small portions; this heuristically
16010 further strengthens the PRNG.
16014 * Fix crypto/bn/asm/mips3.s.
16018 * When only the key is given to "enc", the IV is undefined. Print out
16019 an error message in this case.
16023 * Handle special case when X509_NAME is empty in X509 printing routines.
16027 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
16028 positive and less than q.
16032 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
16033 used: it isn't thread safe and the add_lock_callback should handle
16036 *Paul Rose <Paul.Rose@bridge.com>*
16038 * Verify that incoming data obeys the block size in
16039 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
16045 *Ulf Möller, Bodo Möller*
16047 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
16048 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
16049 when fixing the server behaviour for backwards-compatible 'client
16050 hello' messages. (Note that the attack is impractical against
16051 SSL 3.0 and TLS 1.0 anyway because length and version checking
16052 means that the probability of guessing a valid ciphertext is
16053 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
16056 Before 0.9.5, the countermeasure (hide the error by generating a
16057 random 'decryption result') did not work properly because
16058 ERR_clear_error() was missing, meaning that SSL_get_error() would
16059 detect the supposedly ignored error.
16061 Both problems are now fixed.
16065 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
16066 (previously it was 1024).
16070 * Fix for compatibility mode trust settings: ignore trust settings
16071 unless some valid trust or reject settings are present.
16075 * Fix for blowfish EVP: its a variable length cipher.
16079 * Fix various bugs related to DSA S/MIME verification. Handle missing
16080 parameters in DSA public key structures and return an error in the
16081 DSA routines if parameters are absent.
16085 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
16086 in the current directory if neither $RANDFILE nor $HOME was set.
16087 RAND_file_name() in 0.9.6a returned NULL in this case. This has
16088 caused some confusion to Windows users who haven't defined $HOME.
16089 Thus RAND_file_name() is changed again: e_os.h can define a
16090 DEFAULT_HOME, which will be used if $HOME is not set.
16091 For Windows, we use "C:"; on other platforms, we still require
16092 environment variables.
16094 * Move 'if (!initialized) RAND_poll()' into regions protected by
16095 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
16096 having multiple threads call RAND_poll() concurrently.
16100 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
16101 combination of a flag and a thread ID variable.
16102 Otherwise while one thread is in ssleay_rand_bytes (which sets the
16103 flag), *other* threads can enter ssleay_add_bytes without obeying
16104 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
16105 that they do not hold after the first thread unsets add_do_not_lock).
16109 * Change bctest again: '-x' expressions are not available in all
16110 versions of 'test'.
16114 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
16116 * Fix a couple of memory leaks in PKCS7_dataDecode()
16118 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
16120 * Change Configure and Makefiles to provide EXE_EXT, which will contain
16121 the default extension for executables, if any. Also, make the perl
16122 scripts that use symlink() to test if it really exists and use "cp"
16123 if it doesn't. All this made OpenSSL compilable and installable in
16128 * Fix for asn1_GetSequence() for indefinite length constructed data.
16129 If SEQUENCE is length is indefinite just set c->slen to the total
16130 amount of data available.
16132 *Steve Henson, reported by shige@FreeBSD.org*
16134 *This change does not apply to 0.9.7.*
16136 * Change bctest to avoid here-documents inside command substitution
16137 (workaround for FreeBSD /bin/sh bug).
16138 For compatibility with Ultrix, avoid shell functions (introduced
16139 in the bctest version that searches along $PATH).
16143 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16144 with des_encrypt() defined on some operating systems, like Solaris
16149 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16150 On the Importance of Eliminating Errors in Cryptographic
16151 Computations, J. Cryptology 14 (2001) 2, 101-119,
16152 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16156 * MIPS assembler BIGNUM division bug fix.
16160 * Disabled incorrect Alpha assembler code.
16164 * Fix PKCS#7 decode routines so they correctly update the length
16165 after reading an EOC for the EXPLICIT tag.
16169 *This change does not apply to 0.9.7.*
16171 * Fix bug in PKCS#12 key generation routines. This was triggered
16172 if a 3DES key was generated with a 0 initial byte. Include
16173 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16174 (but broken) behaviour.
16178 * Enhance bctest to search for a working bc along $PATH and print
16181 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16183 * Fix memory leaks in err.c: free err_data string if necessary;
16184 don't write to the wrong index in ERR_set_error_data.
16188 * Implement ssl23_peek (analogous to ssl23_read), which previously
16193 * Replace rdtsc with `_emit` statements for VC++ version 5.
16195 *Jeremy Cooper <jeremy@baymoo.org>*
16197 * Make it possible to reuse SSLv2 sessions.
16201 * In copy_email() check for >= 0 as a return value for
16202 X509_NAME_get_index_by_NID() since 0 is a valid index.
16204 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16206 * Avoid coredump with unsupported or invalid public keys by checking if
16207 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16208 PKCS7_verify() fails with non detached data.
16212 * Don't use getenv in library functions when run as setuid/setgid.
16213 New function OPENSSL_issetugid().
16217 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16218 due to incorrect handling of multi-threading:
16220 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16222 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16224 3. Count how many times MemCheck_off() has been called so that
16225 nested use can be treated correctly. This also avoids
16226 inband-signalling in the previous code (which relied on the
16227 assumption that thread ID 0 is impossible).
16231 * Add "-rand" option also to s_client and s_server.
16235 * Fix CPU detection on Irix 6.x.
16236 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16237 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16239 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16244 *This change does not apply to 0.9.7.*
16246 * Use the cached encoding of an X509_NAME structure rather than
16247 copying it. This is apparently the reason for the libsafe "errors"
16248 but the code is actually correct.
16252 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16253 Bleichenbacher's DSA attack.
16254 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16255 to be set and top=0 forces the highest bit to be set; top=-1 is new
16256 and leaves the highest bit random.
16258 *Ulf Moeller, Bodo Moeller*
16260 * In the `NCONF_...`-based implementations for `CONF_...` queries
16261 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16262 a temporary CONF structure with the data component set to NULL
16263 (which gives segmentation faults in lh_retrieve).
16264 Instead, use NULL for the CONF pointer in CONF_get_string and
16265 CONF_get_number (which may use environment variables) and directly
16266 return NULL from CONF_get_section.
16270 * Fix potential buffer overrun for EBCDIC.
16274 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16275 keyUsage if basicConstraints absent for a CA.
16279 * Make SMIME_write_PKCS7() write mail header values with a format that
16280 is more generally accepted (no spaces before the semicolon), since
16281 some programs can't parse those values properly otherwise. Also make
16282 sure BIO's that break lines after each write do not create invalid
16287 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16288 macros previously used would not encode an empty SEQUENCE OF
16289 and break the signature.
16293 *This change does not apply to 0.9.7.*
16295 * Zero the premaster secret after deriving the master secret in
16300 * Add some EVP_add_digest_alias registrations (as found in
16301 OpenSSL_add_all_digests()) to SSL_library_init()
16302 aka OpenSSL_add_ssl_algorithms(). This provides improved
16303 compatibility with peers using X.509 certificates
16304 with unconventional AlgorithmIdentifier OIDs.
16308 * Fix for Irix with NO_ASM.
16310 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16312 * ./config script fixes.
16314 *Ulf Moeller, Richard Levitte*
16316 * Fix 'openssl passwd -1'.
16320 * Change PKCS12_key_gen_asc() so it can cope with non null
16321 terminated strings whose length is passed in the passlen
16322 parameter, for example from PEM callbacks. This was done
16323 by adding an extra length parameter to asc2uni().
16325 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16327 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16328 call failed, free the DSA structure.
16332 * Fix to uni2asc() to cope with zero length Unicode strings.
16333 These are present in some PKCS#12 files.
16337 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16338 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16339 when writing a 32767 byte record.
16341 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16343 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16344 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16346 (RSA objects have a reference count access to which is protected
16347 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16348 so they are meant to be shared between threads.)
16349 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16350 "Reddie, Steven" <Steven.Reddie@ca.com>*
16352 * Fix a deadlock in CRYPTO_mem_leaks().
16356 * Use better test patterns in bntest.
16360 * rand_win.c fix for Borland C.
16364 * BN_rshift bugfix for n == 0.
16368 * Add a 'bctest' script that checks for some known 'bc' bugs
16369 so that 'make test' does not abort just because 'bc' is broken.
16373 * Store verify_result within SSL_SESSION also for client side to
16374 avoid potential security hole. (Re-used sessions on the client side
16375 always resulted in verify_result==X509_V_OK, not using the original
16376 result of the server certificate verification.)
16380 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16381 SSL3_RT_APPLICATION_DATA, return 0.
16382 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16387 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16388 releases, have been re-implemented by renaming the previous
16389 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16390 and ssl3_read_internal, respectively, and adding 'peek' parameters
16391 to them. The new ssl[23]_{read,peek} functions are calls to
16392 ssl[23]_read_internal with the 'peek' flag set appropriately.
16393 A 'peek' parameter has also been added to ssl3_read_bytes, which
16394 does the actual work for ssl3_read_internal.
16398 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16399 the method-specific "init()" handler. Also clean up ex_data after
16400 calling the method-specific "finish()" handler. Previously, this was
16401 happening the other way round.
16405 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16406 The previous value, 12, was not always sufficient for BN_mod_exp().
16410 * Make sure that shared libraries get the internal name engine with
16411 the full version number and not just 0. This should mark the
16412 shared libraries as not backward compatible. Of course, this should
16413 be changed again when we can guarantee backward binary compatibility.
16417 * Fix typo in get_cert_by_subject() in by_dir.c
16419 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16421 * Rework the system to generate shared libraries:
16423 - Make note of the expected extension for the shared libraries and
16424 if there is a need for symbolic links from for example libcrypto.so.0
16425 to libcrypto.so.0.9.7. There is extended info in Configure for
16428 - Make as few rebuilds of the shared libraries as possible.
16430 - Still avoid linking the OpenSSL programs with the shared libraries.
16432 - When installing, install the shared libraries separately from the
16437 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16439 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16440 and not in SSL_clear because the latter is also used by the
16441 accept/connect functions; previously, the settings made by
16442 SSL_set_read_ahead would be lost during the handshake.
16444 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16446 * Correct util/mkdef.pl to be selective about disabled algorithms.
16447 Previously, it would create entries for disabled algorithms no
16452 * Added several new manual pages for SSL_* function.
16456 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16458 * In ssl23_get_client_hello, generate an error message when faced
16459 with an initial SSL 3.0/TLS record that is too small to contain the
16460 first two bytes of the ClientHello message, i.e. client_version.
16461 (Note that this is a pathologic case that probably has never happened
16462 in real life.) The previous approach was to use the version number
16463 from the record header as a substitute; but our protocol choice
16464 should not depend on that one because it is not authenticated
16465 by the Finished messages.
16469 * More robust randomness gathering functions for Windows.
16471 *Jeffrey Altman <jaltman@columbia.edu>*
16473 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16474 not set then we don't setup the error code for issuer check errors
16475 to avoid possibly overwriting other errors which the callback does
16476 handle. If an application does set the flag then we assume it knows
16477 what it is doing and can handle the new informational codes
16482 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16483 a general "ANY" type, as such it should be able to decode anything
16484 including tagged types. However it didn't check the class so it would
16485 wrongly interpret tagged types in the same way as their universal
16486 counterpart and unknown types were just rejected. Changed so that the
16487 tagged and unknown types are handled in the same way as a SEQUENCE:
16488 that is the encoding is stored intact. There is also a new type
16489 "V_ASN1_OTHER" which is used when the class is not universal, in this
16490 case we have no idea what the actual type is so we just lump them all
16495 * On VMS, stdout may very well lead to a file that is written to
16496 in a record-oriented fashion. That means that every write() will
16497 write a separate record, which will be read separately by the
16498 programs trying to read from it. This can be very confusing.
16500 The solution is to put a BIO filter in the way that will buffer
16501 text until a linefeed is reached, and then write everything a
16502 line at a time, so every record written will be an actual line,
16503 not chunks of lines and not (usually doesn't happen, but I've
16504 seen it once) several lines in one record. BIO_f_linebuffer() is
16507 Currently, it's a VMS-only method, because that's where it has
16508 been tested well enough.
16512 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16513 it can return incorrect results.
16514 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16515 but it was in 0.9.6-beta[12].)
16519 * Disable the check for content being present when verifying detached
16520 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16521 include zero length content when signing messages.
16525 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16526 BIO_ctrl (for BIO pairs).
16530 * Add DSO method for VMS.
16534 * Bug fix: Montgomery multiplication could produce results with the
16539 * Add RPM specification openssl.spec and modify it to build three
16540 packages. The default package contains applications, application
16541 documentation and run-time libraries. The devel package contains
16542 include files, static libraries and function documentation. The
16543 doc package contains the contents of the doc directory. The original
16544 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16548 * Add a large number of documentation files for many SSL routines.
16550 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16552 * Add a configuration entry for Sony News 4.
16554 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16556 * Don't set the two most significant bits to one when generating a
16557 random number < q in the DSA library.
16561 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16562 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16563 the underlying transport is blocking) if a handshake took place.
16564 (The default behaviour is needed by applications such as s_client
16565 and s_server that use select() to determine when to use SSL_read;
16566 but for applications that know in advance when to expect data, it
16567 just makes things more complicated.)
16571 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16576 * Add a few more EBCDIC conditionals that make `req` and `x509`
16577 work better on such systems.
16579 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16581 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16582 Update PKCS12_parse() so it copies the friendlyName and the
16583 keyid to the certificates aux info.
16587 * Fix bug in PKCS7_verify() which caused an infinite loop
16588 if there was more than one signature.
16590 *Sven Uszpelkat <su@celocom.de>*
16592 * Major change in util/mkdef.pl to include extra information
16593 about each symbol, as well as presenting variables as well
16594 as functions. This change means that there's n more need
16595 to rebuild the .num files when some algorithms are excluded.
16599 * Allow the verify time to be set by an application,
16600 rather than always using the current time.
16604 * Phase 2 verify code reorganisation. The certificate
16605 verify code now looks up an issuer certificate by a
16606 number of criteria: subject name, authority key id
16607 and key usage. It also verifies self signed certificates
16608 by the same criteria. The main comparison function is
16609 X509_check_issued() which performs these checks.
16611 Lot of changes were necessary in order to support this
16612 without completely rewriting the lookup code.
16614 Authority and subject key identifier are now cached.
16616 The LHASH 'certs' is X509_STORE has now been replaced
16617 by a STACK_OF(X509_OBJECT). This is mainly because an
16618 LHASH can't store or retrieve multiple objects with
16619 the same hash value.
16621 As a result various functions (which were all internal
16622 use only) have changed to handle the new X509_STORE
16623 structure. This will break anything that messed round
16624 with X509_STORE internally.
16626 The functions X509_STORE_add_cert() now checks for an
16627 exact match, rather than just subject name.
16629 The X509_STORE API doesn't directly support the retrieval
16630 of multiple certificates matching a given criteria, however
16631 this can be worked round by performing a lookup first
16632 (which will fill the cache with candidate certificates)
16633 and then examining the cache for matches. This is probably
16634 the best we can do without throwing out X509_LOOKUP
16635 entirely (maybe later...).
16637 The X509_VERIFY_CTX structure has been enhanced considerably.
16639 All certificate lookup operations now go via a get_issuer()
16640 callback. Although this currently uses an X509_STORE it
16641 can be replaced by custom lookups. This is a simple way
16642 to bypass the X509_STORE hackery necessary to make this
16643 work and makes it possible to use more efficient techniques
16644 in future. A very simple version which uses a simple
16645 STACK for its trusted certificate store is also provided
16646 using X509_STORE_CTX_trusted_stack().
16648 The verify_cb() and verify() callbacks now have equivalents
16649 in the X509_STORE_CTX structure.
16651 X509_STORE_CTX also has a 'flags' field which can be used
16652 to customise the verify behaviour.
16656 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16657 excludes S/MIME capabilities.
16661 * When a certificate request is read in keep a copy of the
16662 original encoding of the signed data and use it when outputting
16663 again. Signatures then use the original encoding rather than
16664 a decoded, encoded version which may cause problems if the
16665 request is improperly encoded.
16669 * For consistency with other BIO_puts implementations, call
16670 buffer_write(b, ...) directly in buffer_puts instead of calling
16673 In BIO_puts, increment b->num_write as in BIO_write.
16675 *Peter.Sylvester@EdelWeb.fr*
16677 * Fix BN_mul_word for the case where the word is 0. (We have to use
16678 BN_zero, we may not return a BIGNUM with an array consisting of
16679 words set to zero.)
16683 * Avoid calling abort() from within the library when problems are
16684 detected, except if preprocessor symbols have been defined
16685 (such as REF_CHECK, BN_DEBUG etc.).
16689 * New openssl application 'rsautl'. This utility can be
16690 used for low-level RSA operations. DER public key
16691 BIO/fp routines also added.
16695 * New Configure entry and patches for compiling on QNX 4.
16697 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16699 * A demo state-machine implementation was sponsored by
16700 Nuron (<http://www.nuron.com/>) and is now available in
16701 demos/state_machine.
16705 * New options added to the 'dgst' utility for signature
16706 generation and verification.
16710 * Unrecognized PKCS#7 content types are now handled via a
16711 catch all ASN1_TYPE structure. This allows unsupported
16712 types to be stored as a "blob" and an application can
16713 encode and decode it manually.
16717 * Fix various signed/unsigned issues to make a_strex.c
16718 compile under VC++.
16720 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16722 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16723 length if passed a buffer. ASN1_INTEGER_to_BN failed
16724 if passed a NULL BN and its argument was negative.
16726 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16728 * Modification to PKCS#7 encoding routines to output definite
16729 length encoding. Since currently the whole structures are in
16730 memory there's not real point in using indefinite length
16731 constructed encoding. However if OpenSSL is compiled with
16732 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16736 * Added BIO_vprintf() and BIO_vsnprintf().
16740 * Added more prefixes to parse for in the strings written
16741 through a logging bio, to cover all the levels that are available
16742 through syslog. The prefixes are now:
16744 PANIC, EMERG, EMR => LOG_EMERG
16745 ALERT, ALR => LOG_ALERT
16746 CRIT, CRI => LOG_CRIT
16747 ERROR, ERR => LOG_ERR
16748 WARNING, WARN, WAR => LOG_WARNING
16749 NOTICE, NOTE, NOT => LOG_NOTICE
16750 INFO, INF => LOG_INFO
16751 DEBUG, DBG => LOG_DEBUG
16753 and as before, if none of those prefixes are present at the
16754 beginning of the string, LOG_ERR is chosen.
16756 On Win32, the `LOG_*` levels are mapped according to this:
16758 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16759 LOG_WARNING => EVENTLOG_WARNING_TYPE
16760 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16764 * Made it possible to reconfigure with just the configuration
16765 argument "reconf" or "reconfigure". The command line arguments
16766 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16767 and are retrieved from there when reconfiguring.
16773 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16775 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16779 * The obj_dat.pl script was messing up the sorting of object
16780 names. The reason was that it compared the quoted version
16781 of strings as a result "OCSP" > "OCSP Signing" because
16782 " > SPACE. Changed script to store unquoted versions of
16783 names and add quotes on output. It was also omitting some
16784 names from the lookup table if they were given a default
16785 value (that is if SN is missing it is given the same
16786 value as LN and vice versa), these are now added on the
16787 grounds that if an object has a name we should be able to
16788 look it up. Finally added warning output when duplicate
16789 short or long names are found.
16793 * Changes needed for Tandem NSK.
16795 *Scott Uroff <scott@xypro.com>*
16797 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16798 RSA_padding_check_SSLv23(), special padding was never detected
16799 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16800 version rollback attacks was not effective.
16802 In s23_clnt.c, don't use special rollback-attack detection padding
16803 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16804 client; similarly, in s23_srvr.c, don't do the rollback check if
16805 SSL 2.0 is the only protocol enabled in the server.
16809 * Make it possible to get hexdumps of unprintable data with 'openssl
16810 asn1parse'. By implication, the functions ASN1_parse_dump() and
16811 BIO_dump_indent() are added.
16815 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16816 these print out strings and name structures based on various
16817 flags including RFC2253 support and proper handling of
16818 multibyte characters. Added options to the 'x509' utility
16819 to allow the various flags to be set.
16823 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16824 Also change the functions X509_cmp_current_time() and
16825 X509_gmtime_adj() work with an ASN1_TIME structure,
16826 this will enable certificates using GeneralizedTime in validity
16827 dates to be checked.
16831 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16832 negative public key encodings) on by default,
16833 NO_NEG_PUBKEY_BUG can be set to disable it.
16837 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16838 content octets. An i2c_ASN1_OBJECT is unnecessary because
16839 the encoding can be trivially obtained from the structure.
16843 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16844 not read locks (`CRYPTO_r_[un]lock`).
16848 * A first attempt at creating official support for shared
16849 libraries through configuration. I've kept it so the
16850 default is static libraries only, and the OpenSSL programs
16851 are always statically linked for now, but there are
16852 preparations for dynamic linking in place.
16853 This has been tested on Linux and Tru64.
16857 * Randomness polling function for Win9x, as described in:
16858 Peter Gutmann, Software Generation of Practically Strong
16863 * Fix so PRNG is seeded in req if using an already existing
16868 * New options to smime application. -inform and -outform
16869 allow alternative formats for the S/MIME message including
16870 PEM and DER. The -content option allows the content to be
16871 specified separately. This should allow things like Netscape
16872 form signing output easier to verify.
16876 * Fix the ASN1 encoding of tags using the 'long form'.
16880 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16881 STRING types. These convert content octets to and from the
16882 underlying type. The actual tag and length octets are
16883 already assumed to have been read in and checked. These
16884 are needed because all other string types have virtually
16885 identical handling apart from the tag. By having versions
16886 of the ASN1 functions that just operate on content octets
16887 IMPLICIT tagging can be handled properly. It also allows
16888 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16889 and ASN1_INTEGER are identical apart from the tag.
16893 * Change the handling of OID objects as follows:
16895 - New object identifiers are inserted in objects.txt, following
16896 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16897 - objects.pl is used to process obj_mac.num and create a new
16899 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16902 This is currently kind of a hack, and the perl code in objects.pl
16903 isn't very elegant, but it works as I intended. The simplest way
16904 to check that it worked correctly is to look in obj_dat.h and
16905 check the array nid_objs and make sure the objects haven't moved
16906 around (this is important!). Additions are OK, as well as
16907 consistent name changes.
16911 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16915 * Addition of the command line parameter '-rand file' to 'openssl req'.
16916 The given file adds to whatever has already been seeded into the
16917 random pool through the RANDFILE configuration file option or
16918 environment variable, or the default random state file.
16922 * mkstack.pl now sorts each macro group into lexical order.
16923 Previously the output order depended on the order the files
16924 appeared in the directory, resulting in needless rewriting
16929 * Patches to make OpenSSL compile under Win32 again. Mostly
16930 work arounds for the VC++ problem that it treats func() as
16931 func(void). Also stripped out the parts of mkdef.pl that
16932 added extra typesafe functions: these no longer exist.
16936 * Reorganisation of the stack code. The macros are now all
16937 collected in safestack.h . Each macro is defined in terms of
16938 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16939 DEBUG_SAFESTACK is now handled in terms of function casts,
16940 this has the advantage of retaining type safety without the
16941 use of additional functions. If DEBUG_SAFESTACK is not defined
16942 then the non typesafe macros are used instead. Also modified the
16943 mkstack.pl script to handle the new form. Needs testing to see
16944 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16945 the default if no major problems. Similar behaviour for ASN1_SET_OF
16946 and PKCS12_STACK_OF.
16950 * When some versions of IIS use the 'NET' form of private key the
16951 key derivation algorithm is different. Normally MD5(password) is
16952 used as a 128 bit RC4 key. In the modified case
16953 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16954 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16955 as the old Netscape_RSA functions except they have an additional
16956 'sgckey' parameter which uses the modified algorithm. Also added
16957 an -sgckey command line option to the rsa utility. Thanks to
16958 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16959 algorithm to openssl-dev.
16963 * The evp_local.h macros were using 'c.##kname' which resulted in
16964 invalid expansion on some systems (SCO 5.0.5 for example).
16965 Corrected to 'c.kname'.
16967 *Phillip Porch <root@theporch.com>*
16969 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16970 a STACK of email addresses from a certificate or request, these look
16971 in the subject name and the subject alternative name extensions and
16972 omit any duplicate addresses.
16976 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16977 This makes DSA verification about 2 % faster.
16981 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16982 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16983 plus overhead for 1024 bit moduli).
16984 This makes exponentiations about 0.5 % faster for 1024 bit
16985 exponents (as measured by "openssl speed rsa2048").
16989 * Rename memory handling macros to avoid conflicts with other
16991 Malloc => OPENSSL_malloc
16992 Malloc_locked => OPENSSL_malloc_locked
16993 Realloc => OPENSSL_realloc
16994 Free => OPENSSL_free
16998 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16999 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
17003 * CygWin32 support.
17005 *John Jarvie <jjarvie@newsguy.com>*
17007 * The type-safe stack code has been rejigged. It is now only compiled
17008 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
17009 by default all type-specific stack functions are "#define"d back to
17010 standard stack functions. This results in more streamlined output
17011 but retains the type-safety checking possibilities of the original
17016 * The STACK code has been cleaned up, and certain type declarations
17017 that didn't make a lot of sense have been brought in line. This has
17018 also involved a cleanup of sorts in safestack.h to more correctly
17019 map type-safe stack functions onto their plain stack counterparts.
17020 This work has also resulted in a variety of "const"ifications of
17021 lots of the code, especially `_cmp` operations which should normally
17022 be prototyped with "const" parameters anyway.
17026 * When generating bytes for the first time in md_rand.c, 'stir the pool'
17027 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
17028 (The PRNG state consists of two parts, the large pool 'state' and 'md',
17029 where all of 'md' is used each time the PRNG is used, but 'state'
17030 is used only indexed by a cyclic counter. As entropy may not be
17031 well distributed from the beginning, 'md' is important as a
17032 chaining variable. However, the output function chains only half
17033 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
17034 all of 'md', and seeding with STATE_SIZE dummy bytes will result
17035 in all of 'state' being rewritten, with the new values depending
17036 on virtually all of 'md'. This overcomes the 80 bit limitation.)
17040 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
17041 the handshake is continued after ssl_verify_cert_chain();
17042 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
17043 can lead to 'unexplainable' connection aborts later.
17045 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
17047 * Major EVP API cipher revision.
17048 Add hooks for extra EVP features. This allows various cipher
17049 parameters to be set in the EVP interface. Support added for variable
17050 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
17051 setting of RC2 and RC5 parameters.
17053 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
17056 Remove lots of duplicated code from the EVP library. For example *every*
17057 cipher init() function handles the 'iv' in the same way according to the
17058 cipher mode. They also all do nothing if the 'key' parameter is NULL and
17059 for CFB and OFB modes they zero ctx->num.
17061 New functionality allows removal of S/MIME code RC2 hack.
17063 Most of the routines have the same form and so can be declared in terms
17066 By shifting this to the top level EVP_CipherInit() it can be removed from
17067 all individual ciphers. If the cipher wants to handle IVs or keys
17068 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
17071 Change lots of functions like EVP_EncryptUpdate() to now return a
17072 value: although software versions of the algorithms cannot fail
17073 any installed hardware versions can.
17077 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
17078 this option is set, tolerate broken clients that send the negotiated
17079 protocol version number instead of the requested protocol version
17084 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
17085 i.e. non-zero for export ciphersuites, zero otherwise.
17086 Previous versions had this flag inverted, inconsistent with
17087 rsa_tmp_cb (..._TMP_RSA_CB).
17089 *Bodo Moeller; problem reported by Amit Chopra*
17091 * Add missing DSA library text string. Work around for some IIS
17092 key files with invalid SEQUENCE encoding.
17096 * Add a document (doc/standards.txt) that list all kinds of standards
17097 and so on that are implemented in OpenSSL.
17101 * Enhance c_rehash script. Old version would mishandle certificates
17102 with the same subject name hash and wouldn't handle CRLs at all.
17103 Added -fingerprint option to crl utility, to support new c_rehash
17108 * Eliminate non-ANSI declarations in crypto.h and stack.h.
17112 * Fix for SSL server purpose checking. Server checking was
17113 rejecting certificates which had extended key usage present
17114 but no ssl client purpose.
17116 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
17118 * Make PKCS#12 code work with no password. The PKCS#12 spec
17119 is a little unclear about how a blank password is handled.
17120 Since the password in encoded as a BMPString with terminating
17121 double NULL a zero length password would end up as just the
17122 double NULL. However no password at all is different and is
17123 handled differently in the PKCS#12 key generation code. NS
17124 treats a blank password as zero length. MSIE treats it as no
17125 password on export: but it will try both on import. We now do
17126 the same: PKCS12_parse() tries zero length and no password if
17127 the password is set to "" or NULL (NULL is now a valid password:
17128 it wasn't before) as does the pkcs12 application.
17132 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17133 perror when PEM_read_bio_X509_REQ fails, the error message must
17134 be obtained from the error queue.
17138 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17139 it in ERR_remove_state if appropriate, and change ERR_get_state
17140 accordingly to avoid race conditions (this is necessary because
17141 thread_hash is no longer constant once set).
17145 * Bugfix for linux-elf makefile.one.
17149 * RSA_get_default_method() will now cause a default
17150 RSA_METHOD to be chosen if one doesn't exist already.
17151 Previously this was only set during a call to RSA_new()
17152 or RSA_new_method(NULL) meaning it was possible for
17153 RSA_get_default_method() to return NULL.
17157 * Added native name translation to the existing DSO code
17158 that will convert (if the flag to do so is set) filenames
17159 that are sufficiently small and have no path information
17160 into a canonical native form. Eg. "blah" converted to
17161 "libblah.so" or "blah.dll" etc.
17165 * New function ERR_error_string_n(e, buf, len) which is like
17166 ERR_error_string(e, buf), but writes at most 'len' bytes
17167 including the 0 terminator. For ERR_error_string_n, 'buf'
17170 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17172 * CONF library reworked to become more general. A new CONF
17173 configuration file reader "class" is implemented as well as a
17174 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17175 old `CONF_*` functions are still there, but are reimplemented to
17176 work in terms of the new functions. Also, a set of functions
17177 to handle the internal storage of the configuration data is
17178 provided to make it easier to write new configuration file
17179 reader "classes" (I can definitely see something reading a
17180 configuration file in XML format, for example), called `_CONF_*`,
17181 or "the configuration storage API"...
17183 The new configuration file reading functions are:
17185 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17186 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17188 NCONF_default, NCONF_WIN32
17190 NCONF_dump_fp, NCONF_dump_bio
17192 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17193 NCONF_new creates a new CONF object. This works in the same way
17194 as other interfaces in OpenSSL, like the BIO interface.
17195 `NCONF_dump_*` dump the internal storage of the configuration file,
17196 which is useful for debugging. All other functions take the same
17197 arguments as the old `CONF_*` functions with the exception of the
17198 first that must be a `CONF *` instead of a `LHASH *`.
17200 To make it easier to use the new classes with the old `CONF_*` functions,
17201 the function CONF_set_default_method is provided.
17205 * Add '-tls1' option to 'openssl ciphers', which was already
17206 mentioned in the documentation but had not been implemented.
17207 (This option is not yet really useful because even the additional
17208 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17212 * Initial DSO code added into libcrypto for letting OpenSSL (and
17213 OpenSSL-based applications) load shared libraries and bind to
17214 them in a portable way.
17216 *Geoff Thorpe, with contributions from Richard Levitte*
17218 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17220 * Make sure _lrotl and _lrotr are only used with MSVC.
17222 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17223 (the default implementation of RAND_status).
17225 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17226 to '-clrext' (= clear extensions), as intended and documented.
17227 *Bodo Moeller; inconsistency pointed out by Michael Attili
17228 <attili@amaxo.com>*
17230 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17231 was larger than the MD block size.
17233 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17235 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17236 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17237 using the passed key: if the passed key was a private key the result
17238 of X509_print(), for example, would be to print out all the private key
17243 * des_quad_cksum() byte order bug fix.
17244 *Ulf Möller, using the problem description in krb4-0.9.7, where
17245 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17247 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17250 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17252 * For easily testing in shell scripts whether some command
17253 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17254 returns with exit code 0 iff no command of the given name is available.
17255 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17256 the output goes to stdout and nothing is printed to stderr.
17257 Additional arguments are always ignored.
17259 Since for each cipher there is a command of the same name,
17260 the 'no-cipher' compilation switches can be tested this way.
17262 ('openssl no-XXX' is not able to detect pseudo-commands such
17263 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17267 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17271 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17272 is set; it will be thrown away anyway because each handshake creates
17274 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17275 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17276 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17277 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17281 * New s_client option -ign_eof: EOF at stdin is ignored, and
17282 'Q' and 'R' lose their special meanings (quit/renegotiate).
17283 This is part of what -quiet does; unlike -quiet, -ign_eof
17284 does not suppress any output.
17288 * Add compatibility options to the purpose and trust code. The
17289 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17290 accepts a certificate or CA, this was the previous behaviour,
17291 with all the associated security issues.
17293 X509_TRUST_COMPAT is the old trust behaviour: only and
17294 automatically trust self signed roots in certificate store. A
17295 new trust setting X509_TRUST_DEFAULT is used to specify that
17296 a purpose has no associated trust setting and it should instead
17297 use the value in the default purpose.
17301 * Fix the PKCS#8 DSA private key code so it decodes keys again
17302 and fix a memory leak.
17306 * In util/mkerr.pl (which implements 'make errors'), preserve
17307 reason strings from the previous version of the .c file, as
17308 the default to have only downcase letters (and digits) in
17309 automatically generated reasons codes is not always appropriate.
17313 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17314 using strerror. Previously, ERR_reason_error_string() returned
17315 library names as reason strings for SYSerr; but SYSerr is a special
17316 case where small numbers are errno values, not library numbers.
17320 * Add '-dsaparam' option to 'openssl dhparam' application. This
17321 converts DSA parameters into DH parameters. (When creating parameters,
17322 DSA_generate_parameters is used.)
17326 * Include 'length' (recommended exponent length) in C code generated
17327 by 'openssl dhparam -C'.
17331 * The second argument to set_label in perlasm was already being used
17332 so couldn't be used as a "file scope" flag. Moved to third argument
17337 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17338 instead of RAND_bytes for encryption IVs and salts.
17342 * Include RAND_status() into RAND_METHOD instead of implementing
17343 it only for md_rand.c Otherwise replacing the PRNG by calling
17344 RAND_set_rand_method would be impossible.
17348 * Don't let DSA_generate_key() enter an infinite loop if the random
17349 number generation fails.
17353 * New 'rand' application for creating pseudo-random output.
17357 * Added configuration support for Linux/IA64
17359 *Rolf Haberrecker <rolf@suse.de>*
17361 * Assembler module support for Mingw32.
17365 * Shared library support for HPUX (in shlib/).
17367 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17369 * Shared library support for Solaris gcc.
17371 *Lutz Behnke <behnke@trustcenter.de>*
17373 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17375 * PKCS7_encrypt() was adding text MIME headers twice because they
17376 were added manually and by SMIME_crlf_copy().
17380 * In bntest.c don't call BN_rand with zero bits argument.
17382 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17384 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17385 case was implemented. This caused BN_div_recp() to fail occasionally.
17389 * Add an optional second argument to the set_label() in the perl
17390 assembly language builder. If this argument exists and is set
17391 to 1 it signals that the assembler should use a symbol whose
17392 scope is the entire file, not just the current function. This
17393 is needed with MASM which uses the format label:: for this scope.
17395 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17397 * Change the ASN1 types so they are typedefs by default. Before
17398 almost all types were #define'd to ASN1_STRING which was causing
17399 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17404 * Change names of new functions to the new get1/get0 naming
17405 convention: After 'get1', the caller owns a reference count
17406 and has to call `..._free`; 'get0' returns a pointer to some
17407 data structure without incrementing reference counters.
17408 (Some of the existing 'get' functions increment a reference
17409 counter, some don't.)
17410 Similarly, 'set1' and 'add1' functions increase reference
17411 counters or duplicate objects.
17415 * Allow for the possibility of temp RSA key generation failure:
17416 the code used to assume it always worked and crashed on failure.
17420 * Fix potential buffer overrun problem in BIO_printf().
17421 *Ulf Möller, using public domain code by Patrick Powell; problem
17422 pointed out by David Sacerdote <das33@cornell.edu>*
17424 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17425 RAND_egd() and RAND_status(). In the command line application,
17426 the EGD socket can be specified like a seed file using RANDFILE
17431 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17432 Some CAs (e.g. Verisign) distribute certificates in this form.
17436 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17437 list to exclude them. This means that no special compilation option
17438 is needed to use anonymous DH: it just needs to be included in the
17443 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17444 EVP_MD_type. The old functionality is available in a new macro called
17445 EVP_MD_md(). Change code that uses it and update docs.
17449 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17450 where the `void *` argument is replaced by a function pointer argument.
17451 Previously `void *` was abused to point to functions, which works on
17452 many platforms, but is not correct. As these functions are usually
17453 called by macros defined in OpenSSL header files, most source code
17454 should work without changes.
17458 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17459 sections with information on -D... compiler switches used for
17460 compiling the library so that applications can see them. To enable
17461 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17462 must be defined. E.g.,
17463 #define OPENSSL_ALGORITHM_DEFINES
17464 #include <openssl/opensslconf.h>
17465 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17467 *Richard Levitte, Ulf and Bodo Möller*
17469 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17474 * Change the 'other' type in certificate aux info to a STACK_OF
17475 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17476 the required ASN1 format: arbitrary types determined by an OID.
17480 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17481 argument to 'req'. This is not because the function is newer or
17482 better than others it just uses the work 'NEW' in the certificate
17483 request header lines. Some software needs this.
17487 * Reorganise password command line arguments: now passwords can be
17488 obtained from various sources. Delete the PEM_cb function and make
17489 it the default behaviour: i.e. if the callback is NULL and the
17490 usrdata argument is not NULL interpret it as a null terminated pass
17491 phrase. If usrdata and the callback are NULL then the pass phrase
17492 is prompted for as usual.
17496 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17497 the support is automatically enabled. The resulting binaries will
17498 autodetect the card and use it if present.
17500 *Ben Laurie and Compaq Inc.*
17502 * Work around for Netscape hang bug. This sends certificate request
17503 and server done in one record. Since this is perfectly legal in the
17504 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17505 the bugs/SSLv3 entry for more info.
17509 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17513 * Add -rand argument to smime and pkcs12 applications and read/write
17518 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17522 * Add command line password options to the remaining applications.
17526 * Bug fix for BN_div_recp() for numerators with an even number of
17531 * More tests in bntest.c, and changed test_bn output.
17535 * ./config recognizes MacOS X now.
17539 * Bug fix for BN_div() when the first words of num and divisor are
17540 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17544 * Add support for various broken PKCS#8 formats, and command line
17545 options to produce them.
17549 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17550 get temporary BIGNUMs from a BN_CTX.
17554 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17559 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17560 include a #define from the old name to the new. The original intent
17561 was that statically linked binaries could for example just call
17562 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17563 link with digests. This never worked because SSLeay_add_all_digests()
17564 and SSLeay_add_all_ciphers() were in the same source file so calling
17565 one would link with the other. They are now in separate source files.
17569 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17573 * Use a less unusual form of the Miller-Rabin primality test (it used
17574 a binary algorithm for exponentiation integrated into the Miller-Rabin
17575 loop, our standard modexp algorithms are faster).
17579 * Support for the EBCDIC character set completed.
17581 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17583 * Source code cleanups: use const where appropriate, eliminate casts,
17584 use `void *` instead of `char *` in lhash.
17588 * Bugfix: ssl3_send_server_key_exchange was not restartable
17589 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17590 this the server could overwrite ephemeral keys that the client
17595 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17596 using 50 iterations of the Rabin-Miller test.
17598 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17599 iterations of the Rabin-Miller test as required by the appendix
17600 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17601 As BN_is_prime_fasttest includes trial division, DSA parameter
17602 generation becomes much faster.
17604 This implies a change for the callback functions in DSA_is_prime
17605 and DSA_generate_parameters: The callback function is called once
17606 for each positive witness in the Rabin-Miller test, not just
17607 occasionally in the inner loop; and the parameters to the
17608 callback function now provide an iteration count for the outer
17609 loop rather than for the current invocation of the inner loop.
17610 DSA_generate_parameters additionally can call the callback
17611 function with an 'iteration count' of -1, meaning that a
17612 candidate has passed the trial division test (when q is generated
17613 from an application-provided seed, trial division is skipped).
17617 * New function BN_is_prime_fasttest that optionally does trial
17618 division before starting the Rabin-Miller test and has
17619 an additional BN_CTX * argument (whereas BN_is_prime always
17620 has to allocate at least one BN_CTX).
17621 'callback(1, -1, cb_arg)' is called when a number has passed the
17622 trial division stage.
17626 * Fix for bug in CRL encoding. The validity dates weren't being handled
17631 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17635 * New function BN_pseudo_rand().
17639 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17640 bignum version of BN_from_montgomery() with the working code from
17641 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17646 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17647 made it impossible to use the same SSL_SESSION data structure in
17648 SSL2 clients in multiple threads.
17652 * The return value of RAND_load_file() no longer counts bytes obtained
17653 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17654 to seed the PRNG (previously an explicit byte count was required).
17656 *Ulf Möller, Bodo Möller*
17658 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17659 used `char *` instead of `void *` and had casts all over the place.
17663 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17667 * Retain source code compatibility for BN_prime_checks macro:
17668 BN_is_prime(..., BN_prime_checks, ...) now uses
17669 BN_prime_checks_for_size to determine the appropriate number of
17670 Rabin-Miller iterations.
17674 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17675 DH_CHECK_P_NOT_SAFE_PRIME.
17676 (Check if this is true? OpenPGP calls them "strong".)
17680 * Merge the functionality of "dh" and "gendh" programs into a new program
17681 "dhparam". The old programs are retained for now but will handle DH keys
17682 (instead of parameters) in future.
17686 * Make the ciphers, s_server and s_client programs check the return values
17687 when a new cipher list is set.
17691 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17692 ciphers. Before when the 56bit ciphers were enabled the sorting was
17695 The syntax for the cipher sorting has been extended to support sorting by
17696 cipher-strength (using the strength_bits hard coded in the tables).
17697 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17699 Fix a bug in the cipher-command parser: when supplying a cipher command
17700 string with an "undefined" symbol (neither command nor alphanumeric
17701 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17702 an error is flagged.
17704 Due to the strength-sorting extension, the code of the
17705 ssl_create_cipher_list() function was completely rearranged. I hope that
17706 the readability was also increased :-)
17708 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17710 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17711 for the first serial number and places 2 in the serial number file. This
17712 avoids problems when the root CA is created with serial number zero and
17713 the first user certificate has the same issuer name and serial number
17718 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17719 the new code. Add documentation for this stuff.
17723 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17724 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17725 structures and behave in an analogous way to the X509v3 functions:
17726 they shouldn't be called directly but wrapper functions should be used
17729 So we also now have some wrapper functions that call the X509at functions
17730 when passed certificate requests. (TO DO: similar things can be done with
17731 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17732 things. Some of these need some d2i or i2d and print functionality
17733 because they handle more complex structures.)
17737 * Add missing #ifndefs that caused missing symbols when building libssl
17738 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17739 NO_RSA in `ssl/s2*.c`.
17741 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17743 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17744 has a return value which indicates the quality of the random data
17745 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17746 error queue. New function RAND_pseudo_bytes() generates output that is
17747 guaranteed to be unique but not unpredictable. RAND_add is like
17748 RAND_seed, but takes an extra argument for an entropy estimate
17749 (RAND_seed always assumes full entropy).
17753 * Do more iterations of Rabin-Miller probable prime test (specifically,
17754 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17755 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17756 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17757 false-positive rate of at most 2^-80 for random input.
17761 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17765 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17766 in the 0.9.5 release), this returns the chain
17767 from an X509_CTX structure with a dup of the stack and all
17768 the X509 reference counts upped: so the stack will exist
17769 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17772 Also make SSL_SESSION_print() print out the verify return
17777 * Add manpage for the pkcs12 command. Also change the default
17778 behaviour so MAC iteration counts are used unless the new
17779 -nomaciter option is used. This improves file security and
17780 only older versions of MSIE (4.0 for example) need it.
17784 * Honor the no-xxx Configure options when creating .DEF files.
17788 * Add PKCS#10 attributes to field table: challengePassword,
17789 unstructuredName and unstructuredAddress. These are taken from
17790 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17791 international characters are used.
17793 More changes to X509_ATTRIBUTE code: allow the setting of types
17794 based on strings. Remove the 'loc' parameter when adding
17795 attributes because these will be a SET OF encoding which is sorted
17800 * Initial changes to the 'req' utility to allow request generation
17801 automation. This will allow an application to just generate a template
17802 file containing all the field values and have req construct the
17805 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17806 used all over the place including certificate requests and PKCS#7
17807 structures. They are currently handled manually where necessary with
17808 some primitive wrappers for PKCS#7. The new functions behave in a
17809 manner analogous to the X509 extension functions: they allow
17810 attributes to be looked up by NID and added.
17812 Later something similar to the X509V3 code would be desirable to
17813 automatically handle the encoding, decoding and printing of the
17814 more complex types. The string types like challengePassword can
17815 be handled by the string table functions.
17817 Also modified the multi byte string table handling. Now there is
17818 a 'global mask' which masks out certain types. The table itself
17819 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17820 is useful when for example there is only one permissible type
17821 (as in countryName) and using the mask might result in no valid
17826 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17827 SSL_get_peer_finished to allow applications to obtain the latest
17828 Finished messages sent to the peer or expected from the peer,
17829 respectively. (SSL_get_peer_finished is usually the Finished message
17830 actually received from the peer, otherwise the protocol will be aborted.)
17832 As the Finished message are message digests of the complete handshake
17833 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17834 be used for external authentication procedures when the authentication
17835 provided by SSL/TLS is not desired or is not enough.
17839 * Enhanced support for Alpha Linux is added. Now ./config checks if
17840 the host supports BWX extension and if Compaq C is present on the
17841 $PATH. Just exploiting of the BWX extension results in 20-30%
17842 performance kick for some algorithms, e.g. DES and RC4 to mention
17843 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17848 * Add support for MS "fast SGC". This is arguably a violation of the
17849 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17850 weak crypto and after checking the certificate is SGC a second one
17851 with strong crypto. MS SGC stops the first handshake after receiving
17852 the server certificate message and sends a second client hello. Since
17853 a server will typically do all the time consuming operations before
17854 expecting any further messages from the client (server key exchange
17855 is the most expensive) there is little difference between the two.
17857 To get OpenSSL to support MS SGC we have to permit a second client
17858 hello message after we have sent server done. In addition we have to
17859 reset the MAC if we do get this second client hello.
17863 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17864 if a DER encoded private key is RSA or DSA traditional format. Changed
17865 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17866 format DER encoded private key. Newer code should use PKCS#8 format which
17867 has the key type encoded in the ASN1 structure. Added DER private key
17868 support to pkcs8 application.
17872 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17873 ciphersuites has been selected (as required by the SSL 3/TLS 1
17874 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17875 is set, we interpret this as a request to violate the specification
17876 (the worst that can happen is a handshake failure, and 'correct'
17877 behaviour would result in a handshake failure anyway).
17881 * In SSL_CTX_add_session, take into account that there might be multiple
17882 SSL_SESSION structures with the same session ID (e.g. when two threads
17883 concurrently obtain them from an external cache).
17884 The internal cache can handle only one SSL_SESSION with a given ID,
17885 so if there's a conflict, we now throw out the old one to achieve
17890 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17891 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17892 some routines that use cipher OIDs: some ciphers do not have OIDs
17893 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17898 * Simplify the trust setting structure and code. Now we just have
17899 two sequences of OIDs for trusted and rejected settings. These will
17900 typically have values the same as the extended key usage extension
17901 and any application specific purposes.
17903 The trust checking code now has a default behaviour: it will just
17904 check for an object with the same NID as the passed id. Functions can
17905 be provided to override either the default behaviour or the behaviour
17906 for a given id. SSL client, server and email already have functions
17907 in place for compatibility: they check the NID and also return "trusted"
17908 if the certificate is self signed.
17912 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17913 traditional format into an EVP_PKEY structure.
17917 * Add a password callback function PEM_cb() which either prompts for
17918 a password if usr_data is NULL or otherwise assumes it is a null
17919 terminated password. Allow passwords to be passed on command line
17920 environment or config files in a few more utilities.
17924 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17925 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17926 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17927 Update documentation.
17931 * Support for ASN1 "NULL" type. This could be handled before by using
17932 ASN1_TYPE but there wasn't any function that would try to read a NULL
17933 and produce an error if it couldn't. For compatibility we also have
17934 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17935 don't allocate anything because they don't need to.
17939 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17942 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17944 * Rebuild of the memory allocation routines used by OpenSSL code and
17945 possibly others as well. The purpose is to make an interface that
17946 provide hooks so anyone can build a separate set of allocation and
17947 deallocation routines to be used by OpenSSL, for example memory
17948 pool implementations, or something else, which was previously hard
17949 since Malloc(), Realloc() and Free() were defined as macros having
17950 the values malloc, realloc and free, respectively (except for Win32
17951 compilations). The same is provided for memory debugging code.
17952 OpenSSL already comes with functionality to find memory leaks, but
17953 this gives people a chance to debug other memory problems.
17955 With these changes, a new set of functions and macros have appeared:
17957 CRYPTO_set_mem_debug_functions() [F]
17958 CRYPTO_get_mem_debug_functions() [F]
17959 CRYPTO_dbg_set_options() [F]
17960 CRYPTO_dbg_get_options() [F]
17961 CRYPTO_malloc_debug_init() [M]
17963 The memory debug functions are NULL by default, unless the library
17964 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17965 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17966 gives the standard debugging functions that come with OpenSSL) or
17967 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17968 provided by the library user) must be used. When the standard
17969 debugging functions are used, CRYPTO_dbg_set_options can be used to
17970 request additional information:
17971 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17972 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17974 Also, things like CRYPTO_set_mem_functions will always give the
17975 expected result (the new set of functions is used for allocation
17976 and deallocation) at all times, regardless of platform and compiler
17979 To finish it up, some functions that were never use in any other
17980 way than through macros have a new API and new semantic:
17982 CRYPTO_dbg_malloc()
17983 CRYPTO_dbg_realloc()
17986 All macros of value have retained their old syntax.
17988 *Richard Levitte and Bodo Moeller*
17990 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17991 ordering of SMIMECapabilities wasn't in "strength order" and there
17992 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17997 * Some ASN1 types with illegal zero length encoding (INTEGER,
17998 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
18000 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
18002 * Merge in my S/MIME library for OpenSSL. This provides a simple
18003 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
18004 functionality to handle multipart/signed properly) and a utility
18005 called 'smime' to call all this stuff. This is based on code I
18006 originally wrote for Celo who have kindly allowed it to be
18007 included in OpenSSL.
18011 * Add variants des_set_key_checked and des_set_key_unchecked of
18012 des_set_key (aka des_key_sched). Global variable des_check_key
18013 decides which of these is called by des_set_key; this way
18014 des_check_key behaves as it always did, but applications and
18015 the library itself, which was buggy for des_check_key == 1,
18016 have a cleaner way to pick the version they need.
18020 * New function PKCS12_newpass() which changes the password of a
18025 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
18026 dynamic mix. In both cases the ids can be used as an index into the
18027 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
18028 functions so they accept a list of the field values and the
18029 application doesn't need to directly manipulate the X509_TRUST
18034 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
18039 * Modify the way the V3 extension code looks up extensions. This now
18040 works in a similar way to the object code: we have some "standard"
18041 extensions in a static table which is searched with OBJ_bsearch()
18042 and the application can add dynamic ones if needed. The file
18043 crypto/x509v3/ext_dat.h now has the info: this file needs to be
18044 updated whenever a new extension is added to the core code and kept
18045 in ext_nid order. There is a simple program 'tabtest.c' which checks
18046 this. New extensions are not added too often so this file can readily
18047 be maintained manually.
18049 There are two big advantages in doing things this way. The extensions
18050 can be looked up immediately and no longer need to be "added" using
18051 X509V3_add_standard_extensions(): this function now does nothing.
18052 Side note: I get *lots* of email saying the extension code doesn't
18053 work because people forget to call this function.
18054 Also no dynamic allocation is done unless new extensions are added:
18055 so if we don't add custom extensions there is no need to call
18056 X509V3_EXT_cleanup().
18060 * Modify enc utility's salting as follows: make salting the default. Add a
18061 magic header, so unsalted files fail gracefully instead of just decrypting
18062 to garbage. This is because not salting is a big security hole, so people
18063 should be discouraged from doing it.
18067 * Fixes and enhancements to the 'x509' utility. It allowed a message
18068 digest to be passed on the command line but it only used this
18069 parameter when signing a certificate. Modified so all relevant
18070 operations are affected by the digest parameter including the
18071 -fingerprint and -x509toreq options. Also -x509toreq choked if a
18072 DSA key was used because it didn't fix the digest.
18076 * Initial certificate chain verify code. Currently tests the untrusted
18077 certificates for consistency with the verify purpose (which is set
18078 when the X509_STORE_CTX structure is set up) and checks the pathlength.
18080 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
18081 this is because it will reject chains with invalid extensions whereas
18082 every previous version of OpenSSL and SSLeay made no checks at all.
18084 Trust code: checks the root CA for the relevant trust settings. Trust
18085 settings have an initial value consistent with the verify purpose: e.g.
18086 if the verify purpose is for SSL client use it expects the CA to be
18087 trusted for SSL client use. However the default value can be changed to
18088 permit custom trust settings: one example of this would be to only trust
18089 certificates from a specific "secure" set of CAs.
18091 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
18092 which should be used for version portability: especially since the
18093 verify structure is likely to change more often now.
18095 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
18096 to set them. If not set then assume SSL clients will verify SSL servers
18099 Two new options to the verify program: -untrusted allows a set of
18100 untrusted certificates to be passed in and -purpose which sets the
18101 intended purpose of the certificate. If a purpose is set then the
18102 new chain verify code is used to check extension consistency.
18106 * Support for the authority information access extension.
18110 * Modify RSA and DSA PEM read routines to transparently handle
18111 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
18112 public keys in a format compatible with certificate
18113 SubjectPublicKeyInfo structures. Unfortunately there were already
18114 functions called *_PublicKey_* which used various odd formats so
18115 these are retained for compatibility: however the DSA variants were
18116 never in a public release so they have been deleted. Changed dsa/rsa
18117 utilities to handle the new format: note no releases ever handled public
18118 keys so we should be OK.
18120 The primary motivation for this change is to avoid the same fiasco
18121 that dogs private keys: there are several incompatible private key
18122 formats some of which are standard and some OpenSSL specific and
18123 require various evil hacks to allow partial transparent handling and
18124 even then it doesn't work with DER formats. Given the option anything
18125 other than PKCS#8 should be dumped: but the other formats have to
18126 stay in the name of compatibility.
18128 With public keys and the benefit of hindsight one standard format
18129 is used which works with EVP_PKEY, RSA or DSA structures: though
18130 it clearly returns an error if you try to read the wrong kind of key.
18132 Added a -pubkey option to the 'x509' utility to output the public key.
18133 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18134 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18135 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18136 that do the same as the `EVP_PKEY_assign_*()` except they up the
18137 reference count of the added key (they don't "swallow" the
18142 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18143 CRLs would fail if the file contained no certificates or no CRLs:
18144 added a new function to read in both types and return the number
18145 read: this means that if none are read it will be an error. The
18146 DER versions of the certificate and CRL reader would always fail
18147 because it isn't possible to mix certificates and CRLs in DER format
18148 without choking one or the other routine. Changed this to just read
18149 a certificate: this is the best we can do. Also modified the code
18150 in `apps/verify.c` to take notice of return codes: it was previously
18151 attempting to read in certificates from NULL pointers and ignoring
18152 any errors: this is one reason why the cert and CRL reader seemed
18153 to work. It doesn't check return codes from the default certificate
18154 routines: these may well fail if the certificates aren't installed.
18158 * Code to support otherName option in GeneralName.
18162 * First update to verify code. Change the verify utility
18163 so it warns if it is passed a self signed certificate:
18164 for consistency with the normal behaviour. X509_verify
18165 has been modified to it will now verify a self signed
18166 certificate if *exactly* the same certificate appears
18167 in the store: it was previously impossible to trust a
18168 single self signed certificate. This means that:
18169 openssl verify ss.pem
18170 now gives a warning about a self signed certificate but
18171 openssl verify -CAfile ss.pem ss.pem
18176 * For servers, store verify_result in SSL_SESSION data structure
18177 (and add it to external session representation).
18178 This is needed when client certificate verifications fails,
18179 but an application-provided verification callback (set by
18180 SSL_CTX_set_cert_verify_callback) allows accepting the session
18181 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18182 but returns 1): When the session is reused, we have to set
18183 ssl->verify_result to the appropriate error code to avoid
18186 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18188 * Fix a bug in the new PKCS#7 code: it didn't consider the
18189 case in PKCS7_dataInit() where the signed PKCS7 structure
18190 didn't contain any existing data because it was being created.
18192 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18194 * Add a salt to the key derivation routines in enc.c. This
18195 forms the first 8 bytes of the encrypted file. Also add a
18196 -S option to allow a salt to be input on the command line.
18200 * New function X509_cmp(). Oddly enough there wasn't a function
18201 to compare two certificates. We do this by working out the SHA1
18202 hash and comparing that. X509_cmp() will be needed by the trust
18207 * SSL_get1_session() is like SSL_get_session(), but increments
18208 the reference count in the SSL_SESSION returned.
18210 *Geoff Thorpe <geoff@eu.c2.net>*
18212 * Fix for 'req': it was adding a null to request attributes.
18213 Also change the X509_LOOKUP and X509_INFO code to handle
18214 certificate auxiliary information.
18218 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18223 * Add the possibility to add extra information to the memory leak
18224 detecting output, to form tracebacks, showing from where each
18225 allocation was originated: CRYPTO_push_info("constant string") adds
18226 the string plus current file name and line number to a per-thread
18227 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18228 is like calling CYRPTO_pop_info() until the stack is empty.
18229 Also updated memory leak detection code to be multi-thread-safe.
18233 * Add options -text and -noout to pkcs7 utility and delete the
18234 encryption options which never did anything. Update docs.
18238 * Add options to some of the utilities to allow the pass phrase
18239 to be included on either the command line (not recommended on
18240 OSes like Unix) or read from the environment. Update the
18241 manpages and fix a few bugs.
18245 * Add a few manpages for some of the openssl commands.
18249 * Fix the -revoke option in ca. It was freeing up memory twice,
18250 leaking and not finding already revoked certificates.
18254 * Extensive changes to support certificate auxiliary information.
18255 This involves the use of X509_CERT_AUX structure and X509_AUX
18256 functions. An X509_AUX function such as PEM_read_X509_AUX()
18257 can still read in a certificate file in the usual way but it
18258 will also read in any additional "auxiliary information". By
18259 doing things this way a fair degree of compatibility can be
18260 retained: existing certificates can have this information added
18261 using the new 'x509' options.
18263 Current auxiliary information includes an "alias" and some trust
18264 settings. The trust settings will ultimately be used in enhanced
18265 certificate chain verification routines: currently a certificate
18266 can only be trusted if it is self signed and then it is trusted
18271 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18272 The problem was that one of the replacement routines had not been working
18273 since SSLeay releases. For now the offending routine has been replaced
18274 with non-optimised assembler. Even so, this now gives around 95%
18275 performance improvement for 1024 bit RSA signs.
18279 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18280 handling. Most clients have the effective key size in bits equal to
18281 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18282 A few however don't do this and instead use the size of the decrypted key
18283 to determine the RC2 key length and the AlgorithmIdentifier to determine
18284 the effective key length. In this case the effective key length can still
18285 be 40 bits but the key length can be 168 bits for example. This is fixed
18286 by manually forcing an RC2 key into the EVP_PKEY structure because the
18287 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18288 the key length and effective key length are equal.
18292 * Add a bunch of functions that should simplify the creation of
18293 X509_NAME structures. Now you should be able to do:
18294 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18295 and have it automatically work out the correct field type and fill in
18296 the structures. The more adventurous can try:
18297 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18298 and it will (hopefully) work out the correct multibyte encoding.
18302 * Change the 'req' utility to use the new field handling and multibyte
18303 copy routines. Before the DN field creation was handled in an ad hoc
18304 way in req, ca, and x509 which was rather broken and didn't support
18305 BMPStrings or UTF8Strings. Since some software doesn't implement
18306 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18307 using the dirstring_type option. See the new comment in the default
18308 openssl.cnf for more info.
18312 * Make crypto/rand/md_rand.c more robust:
18313 - Assure unique random numbers after fork().
18314 - Make sure that concurrent threads access the global counter and
18315 md serializably so that we never lose entropy in them
18316 or use exactly the same state in multiple threads.
18317 Access to the large state is not always serializable because
18318 the additional locking could be a performance killer, and
18319 md should be large enough anyway.
18323 * New file `apps/app_rand.c` with commonly needed functionality
18324 for handling the random seed file.
18326 Use the random seed file in some applications that previously did not:
18328 dsaparam -genkey (which also ignored its '-rand' option),
18331 x509 (when signing).
18332 Except on systems with /dev/urandom, it is crucial to have a random
18333 seed file at least for key creation, DSA signing, and for DH exchanges;
18334 for RSA signatures we could do without one.
18336 gendh and gendsa (unlike genrsa) used to read only the first byte
18337 of each file listed in the '-rand' option. The function as previously
18338 found in genrsa is now in app_rand.c and is used by all programs
18339 that support '-rand'.
18343 * In RAND_write_file, use mode 0600 for creating files;
18344 don't just chmod when it may be too late.
18348 * Report an error from X509_STORE_load_locations
18349 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18353 * New function ASN1_mbstring_copy() this copies a string in either
18354 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18355 into an ASN1_STRING type. A mask of permissible types is passed
18356 and it chooses the "minimal" type to use or an error if not type
18361 * Add function equivalents to the various macros in asn1.h. The old
18362 macros are retained with an `M_` prefix. Code inside the library can
18363 use the `M_` macros. External code (including the openssl utility)
18364 should *NOT* in order to be "shared library friendly".
18368 * Add various functions that can check a certificate's extensions
18369 to see if it usable for various purposes such as SSL client,
18370 server or S/MIME and CAs of these types. This is currently
18371 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18372 verification. Also added a -purpose flag to x509 utility to
18373 print out all the purposes.
18377 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18382 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18383 for, obtain and decode and extension and obtain its critical flag.
18384 This allows all the necessary extension code to be handled in a
18385 single function call.
18389 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18390 platforms. See crypto/rc4/rc4_enc.c for further details.
18394 * New -noout option to asn1parse. This causes no output to be produced
18395 its main use is when combined with -strparse and -out to extract data
18396 from a file (which may not be in ASN.1 format).
18400 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18401 when producing the local key id.
18403 *Richard Levitte <levitte@stacken.kth.se>*
18405 * New option -dhparam in s_server. This allows a DH parameter file to be
18406 stated explicitly. If it is not stated then it tries the first server
18407 certificate file. The previous behaviour hard coded the filename
18412 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18413 a public key to be input or output. For example:
18414 openssl rsa -in key.pem -pubout -out pubkey.pem
18415 Also added necessary DSA public key functions to handle this.
18419 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18420 in the message. This was handled by allowing
18421 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18423 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18425 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18426 to the end of the strings whereas this didn't. This would cause problems
18427 if strings read with d2i_ASN1_bytes() were later modified.
18429 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18431 * Fix for base64 decode bug. When a base64 bio reads only one line of
18432 data and it contains EOF it will end up returning an error. This is
18433 caused by input 46 bytes long. The cause is due to the way base64
18434 BIOs find the start of base64 encoded data. They do this by trying a
18435 trial decode on each line until they find one that works. When they
18436 do a flag is set and it starts again knowing it can pass all the
18437 data directly through the decoder. Unfortunately it doesn't reset
18438 the context it uses. This means that if EOF is reached an attempt
18439 is made to pass two EOFs through the context and this causes the
18440 resulting error. This can also cause other problems as well. As is
18441 usual with these problems it takes *ages* to find and the fix is
18442 trivial: move one line.
18444 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18446 * Ugly workaround to get s_client and s_server working under Windows. The
18447 old code wouldn't work because it needed to select() on sockets and the
18448 tty (for keypresses and to see if data could be written). Win32 only
18449 supports select() on sockets so we select() with a 1s timeout on the
18450 sockets and then see if any characters are waiting to be read, if none
18451 are present then we retry, we also assume we can always write data to
18452 the tty. This isn't nice because the code then blocks until we've
18453 received a complete line of data and it is effectively polling the
18454 keyboard at 1s intervals: however it's quite a bit better than not
18455 working at all :-) A dedicated Windows application might handle this
18456 with an event loop for example.
18460 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18461 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18462 will be called when RSA_sign() and RSA_verify() are used. This is useful
18463 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18464 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18465 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18466 This necessitated the support of an extra signature type NID_md5_sha1
18467 for SSL signatures and modifications to the SSL library to use it instead
18468 of calling RSA_public_decrypt() and RSA_private_encrypt().
18472 * Add new -verify -CAfile and -CApath options to the crl program, these
18473 will lookup a CRL issuers certificate and verify the signature in a
18474 similar way to the verify program. Tidy up the crl program so it
18475 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18476 less strict. It will now permit CRL extensions even if it is not
18477 a V2 CRL: this will allow it to tolerate some broken CRLs.
18481 * Initialize all non-automatic variables each time one of the openssl
18482 sub-programs is started (this is necessary as they may be started
18483 multiple times from the "OpenSSL>" prompt).
18485 *Lennart Bang, Bodo Moeller*
18487 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18488 removing all other RSA functionality (this is what NO_RSA does). This
18489 is so (for example) those in the US can disable those operations covered
18490 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18495 * Non-copying interface to BIO pairs.
18496 (still largely untested)
18500 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18501 ASCII string. This was handled independently in various places before.
18505 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18506 UTF8 strings a character at a time.
18510 * Use client_version from client hello to select the protocol
18511 (s23_srvr.c) and for RSA client key exchange verification
18512 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18516 * Add various utility functions to handle SPKACs, these were previously
18517 handled by poking round in the structure internals. Added new function
18518 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18519 print, verify and generate SPKACs. Based on an original idea from
18520 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18524 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18528 * Allow the config file extension section to be overwritten on the
18529 command line. Based on an original idea from Massimiliano Pala
18530 <madwolf@comune.modena.it>. The new option is called -extensions
18531 and can be applied to ca, req and x509. Also -reqexts to override
18532 the request extensions in req and -crlexts to override the crl extensions
18537 * Add new feature to the SPKAC handling in ca. Now you can include
18538 the same field multiple times by preceding it by "XXXX." for example:
18541 this is the same syntax as used in the req config file.
18545 * Allow certificate extensions to be added to certificate requests. These
18546 are specified in a 'req_extensions' option of the req section of the
18547 config file. They can be printed out with the -text option to req but
18548 are otherwise ignored at present.
18552 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18553 data read consists of only the final block it would not decrypted because
18554 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18555 A misplaced 'break' also meant the decrypted final block might not be
18556 copied until the next read.
18560 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18561 a few extra parameters to the DH structure: these will be useful if
18562 for example we want the value of 'q' or implement X9.42 DH.
18566 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18567 provides hooks that allow the default DSA functions or functions on a
18568 "per key" basis to be replaced. This allows hardware acceleration and
18569 hardware key storage to be handled without major modification to the
18570 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18571 associated functions.
18575 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18576 as "read only": it can't be written to and the buffer it points to will
18577 not be freed. Reading from a read only BIO is much more efficient than
18578 a normal memory BIO. This was added because there are several times when
18579 an area of memory needs to be read from a BIO. The previous method was
18580 to create a memory BIO and write the data to it, this results in two
18581 copies of the data and an O(n^2) reading algorithm. There is a new
18582 function BIO_new_mem_buf() which creates a read only memory BIO from
18583 an area of memory. Also modified the PKCS#7 routines to use read only
18588 * Bugfix: ssl23_get_client_hello did not work properly when called in
18589 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18590 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18591 but a retry condition occurred while trying to read the rest.
18595 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18596 NID_pkcs7_encrypted by default: this was wrong since this should almost
18597 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18598 the encrypted data type: this is a more sensible place to put it and it
18599 allows the PKCS#12 code to be tidied up that duplicated this
18604 * Changed obj_dat.pl script so it takes its input and output files on
18605 the command line. This should avoid shell escape redirection problems
18610 * Initial support for certificate extension requests, these are included
18611 in things like Xenroll certificate requests. Included functions to allow
18612 extensions to be obtained and added.
18616 * -crlf option to s_client and s_server for sending newlines as
18617 CRLF (as required by many protocols).
18621 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18623 * Install libRSAglue.a when OpenSSL is built with RSAref.
18625 *Ralf S. Engelschall*
18627 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18629 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18631 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18636 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18637 DH parameters/keys (q is lost during that conversion, but the resulting
18638 DH parameters contain its length).
18640 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18641 much faster than DH_generate_parameters (which creates parameters
18642 where `p = 2*q + 1`), and also the smaller q makes DH computations
18643 much more efficient (160-bit exponentiation instead of 1024-bit
18644 exponentiation); so this provides a convenient way to support DHE
18645 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18646 utter importance to use
18647 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18649 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18650 when such DH parameters are used, because otherwise small subgroup
18651 attacks may become possible!
18655 * Avoid memory leak in i2d_DHparams.
18659 * Allow the -k option to be used more than once in the enc program:
18660 this allows the same encrypted message to be read by multiple recipients.
18664 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18665 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18666 it will always use the numerical form of the OID, even if it has a short
18671 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18672 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18673 otherwise bn_mod_exp was called. In the case of hardware keys for example
18674 no private key components need be present and it might store extra data
18675 in the RSA structure, which cannot be accessed from bn_mod_exp.
18676 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18677 private key operations.
18681 * Added support for SPARC Linux.
18685 * pem_password_cb function type incompatibly changed from
18686 typedef int pem_password_cb(char *buf, int size, int rwflag);
18688 ....(char *buf, int size, int rwflag, void *userdata);
18689 so that applications can pass data to their callbacks:
18690 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18691 additional void * argument, which is just handed through whenever
18692 the password callback is called.
18694 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18696 New function SSL_CTX_set_default_passwd_cb_userdata.
18698 Compatibility note: As many C implementations push function arguments
18699 onto the stack in reverse order, the new library version is likely to
18700 interoperate with programs that have been compiled with the old
18701 pem_password_cb definition (PEM_whatever takes some data that
18702 happens to be on the stack as its last argument, and the callback
18703 just ignores this garbage); but there is no guarantee whatsoever that
18706 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18707 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18708 problems not only on Windows, but also on some Unix platforms.
18709 To avoid problematic command lines, these definitions are now in an
18710 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18711 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18715 * MIPS III/IV assembler module is reimplemented.
18719 * More DES library cleanups: remove references to srand/rand and
18720 delete an unused file.
18724 * Add support for the free Netwide assembler (NASM) under Win32,
18725 since not many people have MASM (ml) and it can be hard to obtain.
18726 This is currently experimental but it seems to work OK and pass all
18727 the tests. Check out INSTALL.W32 for info.
18731 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18732 without temporary keys kept an extra copy of the server key,
18733 and connections with temporary keys did not free everything in case
18738 * New function RSA_check_key and new openssl rsa option -check
18739 for verifying the consistency of RSA keys.
18741 *Ulf Moeller, Bodo Moeller*
18743 * Various changes to make Win32 compile work:
18744 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18745 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18746 comparison" warnings.
18747 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18751 * Add a debugging option to PKCS#5 v2 key generation function: when
18752 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18753 derived keys are printed to stderr.
18757 * Copy the flags in ASN1_STRING_dup().
18759 *Roman E. Pavlov <pre@mo.msk.ru>*
18761 * The x509 application mishandled signing requests containing DSA
18762 keys when the signing key was also DSA and the parameters didn't match.
18764 It was supposed to omit the parameters when they matched the signing key:
18765 the verifying software was then supposed to automatically use the CA's
18766 parameters if they were absent from the end user certificate.
18768 Omitting parameters is no longer recommended. The test was also
18769 the wrong way round! This was probably due to unusual behaviour in
18770 EVP_cmp_parameters() which returns 1 if the parameters match.
18771 This meant that parameters were omitted when they *didn't* match and
18772 the certificate was useless. Certificates signed with 'ca' didn't have
18775 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18777 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18778 The interface is as follows:
18779 Applications can use
18780 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18781 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18782 "off" is now the default.
18783 The library internally uses
18784 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18785 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18786 to disable memory-checking temporarily.
18788 Some inconsistent states that previously were possible (and were
18789 even the default) are now avoided.
18791 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18792 with each memory chunk allocated; this is occasionally more helpful
18793 than just having a counter.
18795 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18797 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18802 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18803 which largely parallels "options", but is for changing API behaviour,
18804 whereas "options" are about protocol behaviour.
18805 Initial "mode" flags are:
18807 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18808 a single record has been written.
18809 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18810 retries use the same buffer location.
18811 (But all of the contents must be
18816 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18819 * Fix problems with no-hmac etc.
18821 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18823 * New functions RSA_get_default_method(), RSA_set_method() and
18824 RSA_get_method(). These allows replacement of RSA_METHODs without having
18825 to mess around with the internals of an RSA structure.
18829 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18830 Also really enable memory leak checks in openssl.c and in some
18833 *Chad C. Mulligan, Bodo Moeller*
18835 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18836 up the length of negative integers. This has now been simplified to just
18837 store the length when it is first determined and use it later, rather
18838 than trying to keep track of where data is copied and updating it to
18840 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18842 * Add a new function PKCS7_signatureVerify. This allows the verification
18843 of a PKCS#7 signature but with the signing certificate passed to the
18844 function itself. This contrasts with PKCS7_dataVerify which assumes the
18845 certificate is present in the PKCS#7 structure. This isn't always the
18846 case: certificates can be omitted from a PKCS#7 structure and be
18847 distributed by "out of band" means (such as a certificate database).
18851 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18852 function prototypes in pem.h, also change util/mkdef.pl to add the
18853 necessary function names.
18857 * mk1mf.pl (used by Windows builds) did not properly read the
18858 options set by Configure in the top level Makefile, and Configure
18859 was not even able to write more than one option correctly.
18860 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18864 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18865 file to be loaded from a BIO or FILE pointer. The BIO version will
18866 for example allow memory BIOs to contain config info.
18870 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18871 Whoever hopes to achieve shared-library compatibility across versions
18872 must use this, not the compile-time macro.
18873 (Exercise 0.9.4: Which is the minimum library version required by
18875 Note: All this applies only to multi-threaded programs, others don't
18880 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18881 through a BIO pair triggered the default case, i.e.
18882 SSLerr(...,SSL_R_UNKNOWN_STATE).
18886 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18887 can use the SSL library even if none of the specific BIOs is
18892 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18893 for the encoded length.
18895 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18897 * Add initial documentation of the X509V3 functions.
18901 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18902 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18903 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18904 secure PKCS#8 private key format with a high iteration count.
18908 * Fix determination of Perl interpreter: A perl or perl5
18909 *directory* in $PATH was also accepted as the interpreter.
18911 *Ralf S. Engelschall*
18913 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18914 wrong with it but it was very old and did things like calling
18915 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18916 unusual formatting.
18920 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18921 to use the new extension code.
18925 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18926 with macros. This should make it easier to change their form, add extra
18927 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18932 * Add to configuration table a new entry that can specify an alternative
18933 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18934 according to Mark Crispin <MRC@Panda.COM>.
18938 * DES CBC did not update the IV. Weird.
18942 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18943 Changing the behaviour of the former might break existing programs --
18944 where IV updating is needed, des_ncbc_encrypt can be used.
18947 * When bntest is run from "make test" it drives bc to check its
18948 calculations, as well as internally checking them. If an internal check
18949 fails, it needs to cause bc to give a non-zero result or make test carries
18950 on without noticing the failure. Fixed.
18954 * DES library cleanups.
18958 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18959 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18960 ciphers. NOTE: although the key derivation function has been verified
18961 against some published test vectors it has not been extensively tested
18962 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18967 * Instead of "mkdir -p", which is not fully portable, use new
18968 Perl script "util/mkdir-p.pl".
18972 * Rewrite the way password based encryption (PBE) is handled. It used to
18973 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18974 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18975 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18976 the 'parameter' field of the AlgorithmIdentifier is passed to the
18977 underlying key generation function so it must do its own ASN1 parsing.
18978 This has also changed the EVP_PBE_CipherInit() function which now has a
18979 'parameter' argument instead of literal salt and iteration count values
18980 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18984 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18985 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18986 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18987 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18988 value was just used as a "magic string" and not used directly its
18989 value doesn't matter.
18993 * Introduce some semblance of const correctness to BN. Shame C doesn't
18998 * "linux-sparc64" configuration (ultrapenguin).
19000 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
19001 "linux-sparc" configuration.
19003 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
19005 * config now generates no-xxx options for missing ciphers.
19009 * Support the EBCDIC character set (work in progress).
19010 File ebcdic.c not yet included because it has a different license.
19012 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19014 * Support BS2000/OSD-POSIX.
19016 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19018 * Make callbacks for key generation use `void *` instead of `char *`.
19022 * Make S/MIME samples compile (not yet tested).
19026 * Additional typesafe stacks.
19030 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
19034 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
19036 * New configuration variant "sco5-gcc".
19038 * Updated some demos.
19040 *Sean O Riordain, Wade Scholine*
19042 * Add missing BIO_free at exit of pkcs12 application.
19046 * Fix memory leak in conf.c.
19050 * Updates for Win32 to assembler version of MD5.
19054 * Set #! path to perl in `apps/der_chop` to where we found it
19055 instead of using a fixed path.
19059 * SHA library changes for irix64-mips4-cc.
19063 * Improvements for VMS support.
19067 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
19069 * Bignum library bug fix. IRIX 6 passes "make test" now!
19070 This also avoids the problems with SC4.2 and unpatched SC5.
19072 *Andy Polyakov <appro@fy.chalmers.se>*
19074 * New functions sk_num, sk_value and sk_set to replace the previous macros.
19075 These are required because of the typesafe stack would otherwise break
19076 existing code. If old code used a structure member which used to be STACK
19077 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
19078 sk_num or sk_value it would produce an error because the num, data members
19079 are not present in STACK_OF. Now it just produces a warning. sk_set
19080 replaces the old method of assigning a value to sk_value
19081 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
19082 that does this will no longer work (and should use sk_set instead) but
19083 this could be regarded as a "questionable" behaviour anyway.
19087 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
19088 correctly handle encrypted S/MIME data.
19092 * Change type of various DES function arguments from des_cblock
19093 (which means, in function argument declarations, pointer to char)
19094 to des_cblock * (meaning pointer to array with 8 char elements),
19095 which allows the compiler to do more typechecking; it was like
19096 that back in SSLeay, but with lots of ugly casts.
19098 Introduce new type const_des_cblock.
19102 * Reorganise the PKCS#7 library and get rid of some of the more obvious
19103 problems: find RecipientInfo structure that matches recipient certificate
19104 and initialise the ASN1 structures properly based on passed cipher.
19108 * Belatedly make the BN tests actually check the results.
19112 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
19113 to and from BNs: it was completely broken. New compilation option
19114 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
19115 key elements as negative integers.
19119 * Reorganize and speed up MD5.
19121 *Andy Polyakov <appro@fy.chalmers.se>*
19125 *Richard Levitte <richard@levitte.org>*
19127 * New option -out to asn1parse to allow the parsed structure to be
19128 output to a file. This is most useful when combined with the -strparse
19129 option to examine the output of things like OCTET STRINGS.
19133 * Make SSL library a little more fool-proof by not requiring any longer
19134 that `SSL_set_{accept,connect}_state` be called before
19135 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19136 in many applications because usually everything *appeared* to work as
19137 intended anyway -- now it really works as intended).
19141 * Move openssl.cnf out of lib/.
19145 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19146 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19147 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19149 *Ralf S. Engelschall*
19151 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19152 handle PKCS#7 enveloped data properly.
19154 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19156 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19157 copying pointers. The cert_st handling is changed by this in
19158 various ways (and thus what used to be known as ctx->default_cert
19159 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19160 any longer when s->cert does not give us what we need).
19161 ssl_cert_instantiate becomes obsolete by this change.
19162 As soon as we've got the new code right (possibly it already is?),
19163 we have solved a couple of bugs of the earlier code where s->cert
19164 was used as if it could not have been shared with other SSL structures.
19166 Note that using the SSL API in certain dirty ways now will result
19167 in different behaviour than observed with earlier library versions:
19168 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19169 does not influence s as it used to.
19171 In order to clean up things more thoroughly, inside SSL_SESSION
19172 we don't use CERT any longer, but a new structure SESS_CERT
19173 that holds per-session data (if available); currently, this is
19174 the peer's certificate chain and, for clients, the server's certificate
19175 and temporary key. CERT holds only those values that can have
19176 meaningful defaults in an SSL_CTX.
19180 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19181 from the internal representation. Various PKCS#7 fixes: remove some
19182 evil casts and set the enc_dig_alg field properly based on the signing
19187 * Allow PKCS#12 password to be set from the command line or the
19188 environment. Let 'ca' get its config file name from the environment
19189 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19194 * Allow certificate policies extension to use an IA5STRING for the
19195 organization field. This is contrary to the PKIX definition but
19196 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19201 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19202 without disallowing inline assembler and the like for non-pedantic builds.
19206 * Support Borland C++ builder.
19208 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19214 * SHA-1 cleanups and performance enhancements.
19216 *Andy Polyakov <appro@fy.chalmers.se>*
19218 * Sparc v8plus assembler for the bignum library.
19220 *Andy Polyakov <appro@fy.chalmers.se>*
19222 * Accept any -xxx and +xxx compiler options in Configure.
19226 * Update HPUX configuration.
19230 * Add missing `sk_<type>_unshift()` function to safestack.h
19232 *Ralf S. Engelschall*
19234 * New function SSL_CTX_use_certificate_chain_file that sets the
19235 "extra_cert"s in addition to the certificate. (This makes sense
19236 only for "PEM" format files, as chains as a whole are not
19241 * Support verify_depth from the SSL API.
19242 x509_vfy.c had what can be considered an off-by-one-error:
19243 Its depth (which was not part of the external interface)
19244 was actually counting the number of certificates in a chain;
19245 now it really counts the depth.
19249 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19250 instead of X509err, which often resulted in confusing error
19251 messages since the error codes are not globally unique
19252 (e.g. an alleged error in ssl3_accept when a certificate
19253 didn't match the private key).
19255 * New function SSL_CTX_set_session_id_context that allows to set a default
19256 value (so that you don't need SSL_set_session_id_context for each
19257 connection using the SSL_CTX).
19261 * OAEP decoding bug fix.
19265 * Support INSTALL_PREFIX for package builders, as proposed by
19270 * New Configure options "threads" and "no-threads". For systems
19271 where the proper compiler options are known (currently Solaris
19272 and Linux), "threads" is the default.
19276 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19280 * Install various scripts to $(OPENSSLDIR)/misc, not to
19281 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19282 such as /usr/local/bin.
19286 * "make linux-shared" to build shared libraries.
19288 *Niels Poppe <niels@netbox.org>*
19290 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19294 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19295 extension adding in x509 utility.
19299 * Remove NOPROTO sections and error code comments.
19303 * Partial rewrite of the DEF file generator to now parse the ANSI
19308 * New Configure options --prefix=DIR and --openssldir=DIR.
19312 * Complete rewrite of the error code script(s). It is all now handled
19313 by one script at the top level which handles error code gathering,
19314 header rewriting and C source file generation. It should be much better
19315 than the old method: it now uses a modified version of Ulf's parser to
19316 read the ANSI prototypes in all header files (thus the old K&R definitions
19317 aren't needed for error creation any more) and do a better job of
19318 translating function codes into names. The old 'ASN1 error code embedded
19319 in a comment' is no longer necessary and it doesn't use .err files which
19320 have now been deleted. Also the error code call doesn't have to appear all
19321 on one line (which resulted in some large lines...).
19325 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19329 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19330 0 (which usually indicates a closed connection), but continue reading.
19334 * Fix some race conditions.
19338 * Add support for CRL distribution points extension. Add Certificate
19339 Policies and CRL distribution points documentation.
19343 * Move the autogenerated header file parts to crypto/opensslconf.h.
19347 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19348 8 of keying material. Merlin has also confirmed interop with this fix
19349 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19351 *Merlin Hughes <merlin@baltimore.ie>*
19353 * Fix lots of warnings.
19355 *Richard Levitte <levitte@stacken.kth.se>*
19357 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19358 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19360 *Richard Levitte <levitte@stacken.kth.se>*
19362 * Fix problems with sizeof(long) == 8.
19364 *Andy Polyakov <appro@fy.chalmers.se>*
19366 * Change functions to ANSI C.
19370 * Fix typos in error codes.
19372 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19374 * Remove defunct assembler files from Configure.
19378 * SPARC v8 assembler BIGNUM implementation.
19380 *Andy Polyakov <appro@fy.chalmers.se>*
19382 * Support for Certificate Policies extension: both print and set.
19383 Various additions to support the r2i method this uses.
19387 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19388 return a const string when you are expecting an allocated buffer.
19392 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19393 types DirectoryString and DisplayText.
19397 * Add code to allow r2i extensions to access the configuration database,
19398 add an LHASH database driver and add several ctx helper functions.
19402 * Fix an evil bug in bn_expand2() which caused various BN functions to
19403 fail when they extended the size of a BIGNUM.
19407 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19408 support typesafe stack.
19412 * Fix typo in SSL_[gs]et_options().
19414 *Nils Frostberg <nils@medcom.se>*
19416 * Delete various functions and files that belonged to the (now obsolete)
19417 old X509V3 handling code.
19421 * New Configure option "rsaref".
19425 * Don't auto-generate pem.h.
19429 * Introduce type-safe ASN.1 SETs.
19433 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19435 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19437 * Introduce type-safe STACKs. This will almost certainly break lots of code
19438 that links with OpenSSL (well at least cause lots of warnings), but fear
19439 not: the conversion is trivial, and it eliminates loads of evil casts. A
19440 few STACKed things have been converted already. Feel free to convert more.
19441 In the fullness of time, I'll do away with the STACK type altogether.
19445 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19446 specified in `<certfile>` by updating the entry in the index.txt file.
19447 This way one no longer has to edit the index.txt file manually for
19448 revoking a certificate. The -revoke option does the gory details now.
19450 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19452 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19453 `-text` option at all and this way the `-noout -text` combination was
19454 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19456 *Ralf S. Engelschall*
19458 * Make sure a corresponding plain text error message exists for the
19459 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19460 verify callback function determined that a certificate was revoked.
19462 *Ralf S. Engelschall*
19464 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19465 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19466 all available ciphers including rc5, which was forgotten until now.
19467 In order to let the testing shell script know which algorithms
19468 are available, a new (up to now undocumented) command
19469 `openssl list-cipher-commands` is used.
19473 * Bugfix: s_client occasionally would sleep in select() when
19474 it should have checked SSL_pending() first.
19478 * New functions DSA_do_sign and DSA_do_verify to provide access to
19479 the raw DSA values prior to ASN.1 encoding.
19483 * Tweaks to Configure
19485 *Niels Poppe <niels@netbox.org>*
19487 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19492 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19496 * New config option to avoid instructions that are illegal on the 80386.
19497 The default code is faster, but requires at least a 486.
19501 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19502 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19503 same as SSL2_VERSION anyway.
19507 * New "-showcerts" option for s_client.
19511 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19512 application. Various cleanups and fixes.
19516 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19517 modify error routines to work internally. Add error codes and PBE init
19518 to library startup routines.
19522 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19523 packing functions to asn1 and evp. Changed function names and error
19524 codes along the way.
19528 * PKCS12 integration: and so it begins... First of several patches to
19529 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19530 objects to objects.h
19534 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19535 and display support for Thawte strong extranet extension.
19539 * Add LinuxPPC support.
19541 *Jeff Dubrule <igor@pobox.org>*
19543 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19544 bn_div_words in alpha.s.
19546 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19548 * Make sure the RSA OAEP test is skipped under -DRSAref because
19549 OAEP isn't supported when OpenSSL is built with RSAref.
19551 *Ulf Moeller <ulf@fitug.de>*
19553 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19554 so they no longer are missing under -DNOPROTO.
19556 *Soren S. Jorvang <soren@t.dk>*
19558 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19560 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19561 doesn't work when the session is reused. Coming soon!
19565 * Fix a security hole, that allows sessions to be reused in the wrong
19566 context thus bypassing client cert protection! All software that uses
19567 client certs and session caches in multiple contexts NEEDS PATCHING to
19568 allow session reuse! A fuller solution is in the works.
19570 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19572 * Some more source tree cleanups (removed obsolete files
19573 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19574 permission on "config" script to be executable) and a fix for the INSTALL
19577 *Ulf Moeller <ulf@fitug.de>*
19579 * Remove some legacy and erroneous uses of malloc, free instead of
19582 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19584 * Make rsa_oaep_test return non-zero on error.
19586 *Ulf Moeller <ulf@fitug.de>*
19588 * Add support for native Solaris shared libraries. Configure
19589 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19590 if someone would make that last step automatic.
19592 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19594 * ctx_size was not built with the right compiler during "make links". Fixed.
19598 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19599 except NULL ciphers". This means the default cipher list will no longer
19600 enable NULL ciphers. They need to be specifically enabled e.g. with
19601 the string "DEFAULT:eNULL".
19605 * Fix to RSA private encryption routines: if p < q then it would
19606 occasionally produce an invalid result. This will only happen with
19607 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19611 * Be less restrictive and allow also `perl util/perlpath.pl
19612 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19613 because this way one can also use an interpreter named `perl5` (which is
19614 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19615 installed as `perl`).
19617 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19619 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19621 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19623 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19624 advapi32.lib to Win32 build and change the pem test comparison
19625 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19626 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19627 and crypto/des/ede_cbcm_enc.c.
19631 * DES quad checksum was broken on big-endian architectures. Fixed.
19635 * Comment out two functions in bio.h that aren't implemented. Fix up the
19636 Win32 test batch file so it (might) work again. The Win32 test batch file
19637 is horrible: I feel ill....
19641 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19642 in e_os.h. Audit of header files to check ANSI and non ANSI
19643 sections: 10 functions were absent from non ANSI section and not exported
19644 from Windows DLLs. Fixed up libeay.num for new functions.
19648 * Make `openssl version` output lines consistent.
19650 *Ralf S. Engelschall*
19652 * Fix Win32 symbol export lists for BIO functions: Added
19653 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19654 to ms/libeay{16,32}.def.
19656 *Ralf S. Engelschall*
19658 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19659 fine under Unix and passes some trivial tests I've now added. But the
19660 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19661 added to make sure no one expects that this stuff really works in the
19662 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19663 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19666 *Ralf S. Engelschall*
19668 * Fix the generation of two part addresses in perl.
19670 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19672 * Add config entry for Linux on MIPS.
19674 *John Tobey <jtobey@channel1.com>*
19676 * Make links whenever Configure is run, unless we are on Windoze.
19680 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19681 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19686 * Add a useful kludge to allow package maintainers to specify compiler and
19687 other platforms details on the command line without having to patch the
19688 Configure script every time: One now can use
19689 `perl Configure <id>:<details>`,
19690 i.e. platform ids are allowed to have details appended
19691 to them (separated by colons). This is treated as there would be a static
19692 pre-configured entry in Configure's %table under key `<id>` with value
19693 `<details>` and `perl Configure <id>` is called. So, when you want to
19694 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19695 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19696 now, which overrides the FreeBSD-elf entry on-the-fly.
19698 *Ralf S. Engelschall*
19700 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19704 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19705 on the `perl Configure ...` command line. This way one can compile
19706 OpenSSL libraries with Position Independent Code (PIC) which is needed
19707 for linking it into DSOs.
19709 *Ralf S. Engelschall*
19711 * Remarkably, export ciphers were totally broken and no-one had noticed!
19716 * Cleaned up the LICENSE document: The official contact for any license
19717 questions now is the OpenSSL core team under openssl-core@openssl.org.
19718 And add a paragraph about the dual-license situation to make sure people
19719 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19720 to the OpenSSL toolkit.
19722 *Ralf S. Engelschall*
19724 * General source tree makefile cleanups: Made `making xxx in yyy...`
19725 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19726 Additionally cleaned up the `make links` target: Remove unnecessary
19727 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19728 to speed processing and no longer clutter the display with confusing
19729 stuff. Instead only the actually done links are displayed.
19731 *Ralf S. Engelschall*
19733 * Permit null encryption ciphersuites, used for authentication only. It used
19734 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19735 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19740 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19741 signed attributes when verifying signatures (this would break them),
19742 the detached data encoding was wrong and public keys obtained using
19743 X509_get_pubkey() weren't freed.
19747 * Add text documentation for the BUFFER functions. Also added a work around
19748 to a Win95 console bug. This was triggered by the password read stuff: the
19749 last character typed gets carried over to the next fread(). If you were
19750 generating a new cert request using 'req' for example then the last
19751 character of the passphrase would be CR which would then enter the first
19756 * Added the new 'Includes OpenSSL Cryptography Software' button as
19757 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19758 button and can be used by applications based on OpenSSL to show the
19759 relationship to the OpenSSL project.
19761 *Ralf S. Engelschall*
19763 * Remove confusing variables in function signatures in files
19764 ssl/ssl_lib.c and ssl/ssl.h.
19766 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19768 * Don't install bss_file.c under PREFIX/include/
19770 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19772 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19773 functions that return function pointers and has support for NT specific
19774 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19775 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19776 unsigned to signed types: this was killing the Win32 compile.
19780 * Add new certificate file to stack functions,
19781 SSL_add_dir_cert_subjects_to_stack() and
19782 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19783 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19784 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19785 This means that Apache-SSL and similar packages don't have to mess around
19786 to add as many CAs as they want to the preferred list.
19790 * Experiment with doxygen documentation. Currently only partially applied to
19792 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19793 openssl.doxy as the configuration file.
19797 * Get rid of remaining C++-style comments which strict C compilers hate.
19799 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19801 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19802 compiled in by default: it has problems with large keys.
19806 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19807 DH private keys and/or callback functions which directly correspond to
19808 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19809 is needed for applications which have to configure certificates on a
19810 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19812 For the RSA certificate situation is makes no difference, but
19813 for the DSA certificate situation this fixes the "no shared cipher"
19814 problem where the OpenSSL cipher selection procedure failed because the
19815 temporary keys were not overtaken from the context and the API provided
19816 no way to reconfigure them.
19817 The new functions now let applications reconfigure the stuff and they
19818 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19819 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19820 non-public-API function ssl_cert_instantiate() is used as a helper
19821 function and also to reduce code redundancy inside ssl_rsa.c.
19823 *Ralf S. Engelschall*
19825 * Move s_server -dcert and -dkey options out of the undocumented feature
19826 area because they are useful for the DSA situation and should be
19827 recognized by the users.
19829 *Ralf S. Engelschall*
19831 * Fix the cipher decision scheme for export ciphers: the export bits are
19832 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19833 SSL_EXP_MASK. So, the original variable has to be used instead of the
19834 already masked variable.
19836 *Richard Levitte <levitte@stacken.kth.se>*
19838 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19840 *Richard Levitte <levitte@stacken.kth.se>*
19842 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19843 from `int` to `unsigned int` because it is a length and initialized by
19844 EVP_DigestFinal() which expects an `unsigned int *`.
19846 *Richard Levitte <levitte@stacken.kth.se>*
19848 * Don't hard-code path to Perl interpreter on shebang line of Configure
19849 script. Instead use the usual Shell->Perl transition trick.
19851 *Ralf S. Engelschall*
19853 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19854 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19855 -noout -modulus` as it's already the case for `openssl rsa -noout
19856 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19857 currently the public key is printed (a decision which was already done by
19858 `openssl dsa -modulus` in the past) which serves a similar purpose.
19859 Additionally the NO_RSA no longer completely removes the whole -modulus
19860 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19863 *Ralf S. Engelschall*
19865 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19866 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19868 *Arne Ansper <arne@ats.cyber.ee>*
19870 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19871 to be added. Now both 'req' and 'ca' can use new objects defined in the
19876 * Add cool BIO that does syslog (or event log on NT).
19878 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19880 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19881 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19882 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19883 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19887 * Add preliminary config info for new extension code.
19891 * Make RSA_NO_PADDING really use no padding.
19893 *Ulf Moeller <ulf@fitug.de>*
19895 * Generate errors when private/public key check is done.
19899 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19900 for some CRL extensions and new objects added.
19904 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19905 key usage extension and fuller support for authority key id.
19909 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19910 padding method for RSA, which is recommended for new applications in PKCS
19911 #1 v2.0 (RFC 2437, October 1998).
19912 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19913 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19914 against Bleichbacher's attack on RSA.
19915 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19918 * Updates to the new SSL compression code
19920 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19922 * Fix so that the version number in the master secret, when passed
19923 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19924 (because the server will not accept higher), that the version number
19925 is 0x03,0x01, not 0x03,0x00
19927 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19929 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19930 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19931 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19935 * Support for RAW extensions where an arbitrary extension can be
19936 created by including its DER encoding. See `apps/openssl.cnf` for
19941 * Make sure latest Perl versions don't interpret some generated C array
19942 code as Perl array code in the crypto/err/err_genc.pl script.
19944 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19946 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19947 not many people have the assembler. Various Win32 compilation fixes and
19948 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19949 build instructions.
19953 * Modify configure script 'Configure' to automatically create crypto/date.h
19954 file under Win32 and also build pem.h from pem.org. New script
19955 util/mkfiles.pl to create the MINFO file on environments that can't do a
19956 'make files': perl util/mkfiles.pl >MINFO should work.
19960 * Major rework of DES function declarations, in the pursuit of correctness
19961 and purity. As a result, many evil casts evaporated, and some weirdness,
19962 too. You may find this causes warnings in your code. Zapping your evil
19963 casts will probably fix them. Mostly.
19967 * Fix for a typo in asn1.h. Bug fix to object creation script
19968 obj_dat.pl. It considered a zero in an object definition to mean
19969 "end of object": none of the objects in objects.h have any zeros
19970 so it wasn't spotted.
19972 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19974 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19975 Masking (CBCM). In the absence of test vectors, the best I have been able
19976 to do is check that the decrypt undoes the encrypt, so far. Send me test
19977 vectors if you have them.
19981 * Correct calculation of key length for export ciphers (too much space was
19982 allocated for null ciphers). This has not been tested!
19986 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19987 message is now correct (it understands "crypto" and "ssl" on its
19988 command line). There is also now an "update" option. This will update
19989 the util/ssleay.num and util/libeay.num files with any new functions.
19991 perl util/mkdef.pl crypto ssl update
19992 it will update them.
19996 * Overhauled the Perl interface:
19997 - ported BN stuff to OpenSSL's different BN library
19998 - made the perl/ source tree CVS-aware
19999 - renamed the package from SSLeay to OpenSSL (the files still contain
20000 their history because I've copied them in the repository)
20001 - removed obsolete files (the test scripts will be replaced
20002 by better Test::Harness variants in the future)
20004 *Ralf S. Engelschall*
20006 * First cut for a very conservative source tree cleanup:
20007 1. merge various obsolete readme texts into doc/ssleay.txt
20008 where we collect the old documents and readme texts.
20009 2. remove the first part of files where I'm already sure that we no
20010 longer need them because of three reasons: either they are just temporary
20011 files which were left by Eric or they are preserved original files where
20012 I've verified that the diff is also available in the CVS via "cvs diff
20013 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
20014 the crypto/md/ stuff).
20016 *Ralf S. Engelschall*
20018 * More extension code. Incomplete support for subject and issuer alt
20019 name, issuer and authority key id. Change the i2v function parameters
20020 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
20021 what that's for :-) Fix to ASN1 macro which messed up
20022 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
20026 * Preliminary support for ENUMERATED type. This is largely copied from the
20031 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
20033 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20035 * Make sure `make rehash` target really finds the `openssl` program.
20037 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
20039 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
20040 like to hear about it if this slows down other processors.
20044 * Add CygWin32 platform information to Configure script.
20046 *Alan Batie <batie@aahz.jf.intel.com>*
20048 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
20050 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
20052 * New program nseq to manipulate netscape certificate sequences
20056 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
20061 * Fixes to BN code. Previously the default was to define BN_RECURSION
20062 but the BN code had some problems that would cause failures when
20063 doing certificate verification and some other functions.
20065 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20067 * Add ASN1 and PEM code to support netscape certificate sequences.
20071 * Add ASN1 and PEM code to support netscape certificate sequences.
20075 * Add several PKIX and private extended key usage OIDs.
20079 * Modify the 'ca' program to handle the new extension code. Modify
20080 openssl.cnf for new extension format, add comments.
20084 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
20085 and add a sample to openssl.cnf so req -x509 now adds appropriate
20090 * Continued X509 V3 changes. Add to other makefiles, integrate with the
20091 error code, add initial support to X509_print() and x509 application.
20095 * Takes a deep breath and start adding X509 V3 extension support code. Add
20096 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
20097 stuff is currently isolated and isn't even compiled yet.
20101 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
20102 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
20103 Removed the versions check from X509 routines when loading extensions:
20104 this allows certain broken certificates that don't set the version
20105 properly to be processed.
20109 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
20110 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
20111 can still be regenerated with "make depend".
20115 * Spelling mistake in C version of CAST-128.
20117 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
20119 * Changes to the error generation code. The perl script err-code.pl
20120 now reads in the old error codes and retains the old numbers, only
20121 adding new ones if necessary. It also only changes the .err files if new
20122 codes are added. The makefiles have been modified to only insert errors
20123 when needed (to avoid needlessly modifying header files). This is done
20124 by only inserting errors if the .err file is newer than the auto generated
20125 C file. To rebuild all the error codes from scratch (the old behaviour)
20126 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
20127 or delete all the .err files.
20131 * CAST-128 was incorrectly implemented for short keys. The C version has
20132 been fixed, but is untested. The assembler versions are also fixed, but
20133 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20134 to regenerate it if needed.
20135 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20136 Hagino <itojun@kame.net>*
20138 * File was opened incorrectly in randfile.c.
20140 *Ulf Möller <ulf@fitug.de>*
20142 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20143 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20144 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20145 al: it's just almost always a UTCTime. Note this patch adds new error
20146 codes so do a "make errors" if there are problems.
20150 * Correct Linux 1 recognition in config.
20152 *Ulf Möller <ulf@fitug.de>*
20154 * Remove pointless MD5 hash when using DSA keys in ca.
20156 *Anonymous <nobody@replay.com>*
20158 * Generate an error if given an empty string as a cert directory. Also
20159 generate an error if handed NULL (previously returned 0 to indicate an
20160 error, but didn't set one).
20162 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20164 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20168 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20169 parameters. This was causing a warning which killed off the Win32 compile.
20173 * Remove C++ style comments from crypto/bn/bn_local.h.
20175 *Neil Costigan <neil.costigan@celocom.com>*
20177 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20178 based on a text string, looking up short and long names and finally
20179 "dot" format. The "dot" format stuff didn't work. Added new function
20180 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20181 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20182 OID is not part of the table.
20186 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20187 X509_LOOKUP_by_alias().
20191 * Sort openssl functions by name.
20195 * Get the `gendsa` command working and add it to the `list` command. Remove
20196 encryption from sample DSA keys (in case anyone is interested the password
20201 * Make *all* `*_free` functions accept a NULL pointer.
20203 *Frans Heymans <fheymans@isaserver.be>*
20205 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20208 *Anonymous <nobody@replay.com>*
20210 * s_server should send the CAfile as acceptable CAs, not its own cert.
20212 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20214 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20216 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20218 * Temp key "for export" tests were wrong in s3_srvr.c.
20220 *Anonymous <nobody@replay.com>*
20222 * Add prototype for temp key callback functions
20223 SSL_CTX_set_tmp_{rsa,dh}_callback().
20227 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20228 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20232 * X509_name_add_entry() freed the wrong thing after an error.
20234 *Arne Ansper <arne@ats.cyber.ee>*
20236 * rsa_eay.c would attempt to free a NULL context.
20238 *Arne Ansper <arne@ats.cyber.ee>*
20240 * BIO_s_socket() had a broken should_retry() on Windoze.
20242 *Arne Ansper <arne@ats.cyber.ee>*
20244 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20246 *Arne Ansper <arne@ats.cyber.ee>*
20248 * Make sure the already existing X509_STORE->depth variable is initialized
20249 in X509_STORE_new(), but document the fact that this variable is still
20250 unused in the certificate verification process.
20252 *Ralf S. Engelschall*
20254 * Fix the various library and `apps/` files to free up pkeys obtained from
20255 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20259 * Fix reference counting in X509_PUBKEY_get(). This makes
20260 demos/maurice/example2.c work, amongst others, probably.
20262 *Steve Henson and Ben Laurie*
20264 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20265 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20266 are no longer created. This way we have a single and consistent command
20267 line interface `openssl <command>`, similar to `cvs <command>`.
20269 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20271 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20272 BIT STRING wrapper always have zero unused bits.
20276 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20280 * Make the top-level INSTALL documentation easier to understand.
20284 * Makefiles updated to exit if an error occurs in a sub-directory
20285 make (including if user presses ^C) [Paul Sutton]
20287 * Make Montgomery context stuff explicit in RSA data structure.
20291 * Fix build order of pem and err to allow for generated pem.h.
20295 * Fix renumbering bug in X509_NAME_delete_entry().
20299 * Enhanced the err-ins.pl script so it makes the error library number
20300 global and can add a library name. This is needed for external ASN1 and
20301 other error libraries.
20305 * Fixed sk_insert which never worked properly.
20309 * Fix ASN1 macros so they can handle indefinite length constructed
20310 EXPLICIT tags. Some non standard certificates use these: they can now
20315 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20316 into a single doc/ssleay.txt bundle. This way the information is still
20317 preserved but no longer messes up this directory. Now it's new room for
20318 the new set of documentation files.
20320 *Ralf S. Engelschall*
20322 * SETs were incorrectly DER encoded. This was a major pain, because they
20323 shared code with SEQUENCEs, which aren't coded the same. This means that
20324 almost everything to do with SETs or SEQUENCEs has either changed name or
20325 number of arguments.
20327 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20329 * Fix test data to work with the above.
20333 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20334 was already fixed by Eric for 0.9.1 it seems.
20336 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20338 * Autodetect FreeBSD3.
20342 * Fix various bugs in Configure. This affects the following platforms:
20346 unixware-2.0-pentium
20351 * Eliminate generated files from CVS. Reorder tests to regenerate files
20352 before they are needed.
20356 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20360 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20362 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20363 changed SSLeay to OpenSSL in version strings.
20365 *Ralf S. Engelschall*
20367 * Some fixups to the top-level documents.
20371 * Fixed the nasty bug where rsaref.h was not found under compile-time
20372 because the symlink to include/ was missing.
20374 *Ralf S. Engelschall*
20376 * Incorporated the popular no-RSA/DSA-only patches
20377 which allow to compile an RSA-free SSLeay.
20379 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20381 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20382 when "ssleay" is still not found.
20384 *Ralf S. Engelschall*
20386 * Added more platforms to Configure: Cray T3E, HPUX 11,
20388 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20390 * Updated the README file.
20392 *Ralf S. Engelschall*
20394 * Added various .cvsignore files in the CVS repository subdirs
20395 to make a "cvs update" really silent.
20397 *Ralf S. Engelschall*
20399 * Recompiled the error-definition header files and added
20400 missing symbols to the Win32 linker tables.
20402 *Ralf S. Engelschall*
20404 * Cleaned up the top-level documents;
20405 o new files: CHANGES and LICENSE
20406 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20407 o merged COPYRIGHT into LICENSE
20408 o removed obsolete TODO file
20409 o renamed MICROSOFT to INSTALL.W32
20411 *Ralf S. Engelschall*
20413 * Removed dummy files from the 0.9.1b source tree:
20414 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20415 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20416 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20417 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20418 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20420 *Ralf S. Engelschall*
20422 * Added various platform portability fixes.
20426 * The Genesis of the OpenSSL rpject:
20427 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20428 Young and Tim J. Hudson created while they were working for C2Net until
20431 *The OpenSSL Project*
20433 ### Changes between 0.9.0b and 0.9.1b [not released]
20435 * Updated a few CA certificates under certs/
20439 * Changed some BIGNUM api stuff.
20443 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20444 DGUX x86, Linux Alpha, etc.
20448 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20449 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20454 * Add -strparse option to asn1pars program which parses nested
20457 *Dr Stephen Henson <shenson@bigfoot.com>*
20459 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20463 * DSA fix for "ca" program.
20467 * Added "-genkey" option to "dsaparam" program.
20471 * Added RIPE MD160 (rmd160) message digest.
20475 * Added -a (all) option to "ssleay version" command.
20479 * Added PLATFORM define which is the id given to Configure.
20483 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20487 * Extended the ASN.1 parser routines.
20491 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20495 * Added a BN_CTX to the BN library.
20499 * Fixed the weak key values in DES library
20503 * Changed API in EVP library for cipher aliases.
20507 * Added support for RC2/64bit cipher.
20511 * Converted the lhash library to the crypto/mem.c functions.
20515 * Added more recognized ASN.1 object ids.
20519 * Added more RSA padding checks for SSL/TLS.
20523 * Added BIO proxy/filter functionality.
20527 * Added extra_certs to SSL_CTX which can be used
20528 send extra CA certificates to the client in the CA cert chain sending
20529 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20533 * Now Fortezza is denied in the authentication phase because
20534 this is key exchange mechanism is not supported by SSLeay at all.
20538 * Additional PKCS1 checks.
20542 * Support the string "TLSv1" for all TLS v1 ciphers.
20546 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20547 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20551 * Fixed a few memory leaks.
20555 * Fixed various code and comment typos.
20559 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20560 bytes sent in the client random.
20562 *Edward Bishop <ebishop@spyglass.com>*
20566 [CVE-2024-2511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-2511
20567 [CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727
20568 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20569 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20570 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20571 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20572 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20573 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20574 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20575 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20576 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20577 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20578 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20579 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20580 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20581 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20582 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20583 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20584 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20585 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20586 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20587 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20588 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20589 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20590 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20591 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20592 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20593 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20594 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20595 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20596 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20597 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20598 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20599 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20600 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20601 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20602 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20603 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20604 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20605 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20606 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20607 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20608 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20609 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20610 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20611 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20612 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20613 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20614 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20615 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20616 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20617 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20618 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20619 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20620 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20621 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20622 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20623 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20624 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20625 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20626 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20627 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20628 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20629 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20630 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20631 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20632 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20633 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20634 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20635 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20636 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20637 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20638 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20639 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20640 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20641 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20642 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20643 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20644 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20645 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20646 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20647 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20648 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20649 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20650 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20651 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20652 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20653 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20654 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20655 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20656 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20657 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20658 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20659 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20660 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20661 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20662 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20663 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20664 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20665 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20666 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20667 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20668 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20669 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20670 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20671 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20672 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20673 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20674 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20675 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20676 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20677 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20678 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20679 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20680 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20681 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20682 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20683 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20684 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20685 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20686 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20687 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20688 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20689 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20690 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20691 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20692 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20693 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20694 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20695 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20696 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20697 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20698 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20699 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20700 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20701 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20702 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20703 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20704 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20705 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20706 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20707 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20708 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20709 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20710 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20711 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20712 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20713 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20714 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20715 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20716 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20717 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20718 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20719 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20720 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20721 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20722 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20723 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20724 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20725 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20726 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20727 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20728 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20729 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20730 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20731 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20732 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20733 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20734 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20735 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20736 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20737 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20738 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20739 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20740 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20741 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20742 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20743 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20744 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20745 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20746 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20747 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20748 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20749 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20750 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20751 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20752 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20753 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20754 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655