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 3.4](#openssl-34)
16 - [OpenSSL 3.3](#openssl-33)
17 - [OpenSSL 3.2](#openssl-32)
18 - [OpenSSL 3.1](#openssl-31)
19 - [OpenSSL 3.0](#openssl-30)
20 - [OpenSSL 1.1.1](#openssl-111)
21 - [OpenSSL 1.1.0](#openssl-110)
22 - [OpenSSL 1.0.2](#openssl-102)
23 - [OpenSSL 1.0.1](#openssl-101)
24 - [OpenSSL 1.0.0](#openssl-100)
25 - [OpenSSL 0.9.x](#openssl-09x)
30 ### Changes between 3.3 and 3.4 [xx XXX xxxx]
32 * Added options `-not_before` and `-not_after` for explicit setting
33 start and end dates of certificates created with the `req` and `x509`
34 apps. Added the same options also to `ca` app as alias for
35 `-startdate` and `-enddate` options.
42 ### Changes between 3.2 and 3.3 [xx XXX xxxx]
44 * The `-verify` option to the `openssl crl` and `openssl req` will make
45 the program exit with 1 on failure.
49 * The BIO_get_new_index() function can only be called 127 times before it
50 reaches its upper bound of BIO_TYPE_MASK. It will now correctly return an
51 error of -1 once it is exhausted. Users may need to reserve using this
52 function for cases where BIO_find_type() is required. Either BIO_TYPE_NONE
53 or BIO_get_new_index() can be used to supply a type to BIO_meth_new().
57 * Added API functions SSL_SESSION_get_time_ex(), SSL_SESSION_set_time_ex()
58 using time_t which is Y2038 safe on 32 bit systems when 64 bit time
59 is enabled (e.g via setting glibc macro _TIME_BITS=64).
63 * The d2i_ASN1_GENERALIZEDTIME(), d2i_ASN1_UTCTIME(), ASN1_TIME_check(), and
64 related functions have been augmented to check for a minimum length of
65 the input string, in accordance with ITU-T X.690 section 11.7 and 11.8.
69 * Unknown entries in TLS SignatureAlgorithms, ClientSignatureAlgorithms
70 config options and the respective calls to SSL[_CTX]_set1_sigalgs() and
71 SSL[_CTX]_set1_client_sigalgs() that start with `?` character are
72 ignored and the configuration will still be used.
74 Similarly unknown entries that start with `?` character in a TLS
75 Groups config option or set with SSL[_CTX]_set1_groups_list() are ignored
76 and the configuration will still be used.
78 In both cases if the resulting list is empty, an error is returned.
82 * The EVP_PKEY_fromdata function has been augmented to allow for the derivation
83 of CRT (Chinese Remainder Theorem) parameters when requested. See the
84 OSSL_PKEY_PARAM_RSA_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
88 * The activate and soft_load configuration settings for providers in
89 openssl.cnf have been updated to require a value of [1|yes|true|on]
90 (in lower or UPPER case) to enable the setting. Conversely a value
91 of [0|no|false|off] will disable the setting. All other values, or the
92 omission of a value for these settings will result in an error.
96 * Added `-set_issuer` and `-set_subject` options to `openssl x509` to
97 override the Issuer and Subject when creating a certificate. The `-subj`
98 option now is an alias for `-set_subject`.
100 *Job Snijders, George Michaelson*
102 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
103 if called with a NULL stack argument.
107 * In `openssl speed`, changed the default hash function used with `hmac` from
112 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
113 - `certProfile` request message header and respective `-profile` CLI option
114 - support for delayed delivery of all types of response messages
118 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
119 be less hard coded in the build file templates, and to allow easier
120 addition of more exporters. With that, an exporter for CMake is also
125 * The BLAKE2s hash algorithm matches BLAKE2b's support
126 for configurable output length.
128 *Ahelenia Ziemiańska*
130 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
131 server to prefer session resumption using PSK-only key exchange over PSK
132 with DHE, if both are available.
134 *Markus Minichmayr, Tapkey GmbH*
136 * New API `SSL_write_ex2`, which can be used to send an end-of-stream (FIN)
137 condition in an optimised way when using QUIC.
141 * New atexit configuration switch, which controls whether the OPENSSL_cleanup
142 is registered when libcrypto is unloaded. This is turned off on NonStop
143 configurations because of loader differences on that platform compared to
148 * Support for qlog for tracing QUIC connections has been added.
150 The qlog output from OpenSSL currently uses a pre-standard draft version of
151 qlog. The output from OpenSSL will change in incompatible ways in future
152 releases, and is not subject to any format stability or compatibility
153 guarantees at this time. This functionality can be
154 disabled with the build-time option `no-unstable-qlog`. See the
155 openssl-qlog(7) manpage for details.
159 * Added APIs to allow configuring the negotiated idle timeout for QUIC
160 connections, and to allow determining the number of additional streams
161 that can currently be created for a QUIC connection.
165 * Added APIs to allow disabling implicit QUIC event processing for
166 QUIC SSL objects, allowing applications to control when event handling
167 occurs. Refer to the SSL_get_value_uint(3) manpage for details.
171 * Limited support for polling of QUIC connection and stream objects in a
172 non-blocking manner. Refer to the SSL_poll(3) manpage for details.
176 * Added APIs to allow querying the size and utilisation of a QUIC stream's
177 write buffer. Refer to the SSL_get_value_uint(3) manpage for details.
181 * New limit on HTTP response headers is introduced to HTTP client. The
182 default limit is set to 256 header lines. If limit is exceeded the
183 response processing stops with error HTTP_R_RESPONSE_TOO_MANY_HDRLINES.
184 Application may call OSSL_HTTP_REQ_CTX_set_max_response_hdr_lines(3)
185 to change the default. Setting the value to 0 disables the limit.
187 *Alexandr Nedvedicky*
189 * Applied AES-GCM unroll8 optimisation to Microsoft Azure Cobalt 100
193 * Added X509_STORE_get1_objects to avoid issues with the existing
194 X509_STORE_get0_objects API in multi-threaded applications. Refer to the
195 documentation for details.
199 * Added assembly implementation for md5 on loongarch64
203 * Optimized AES-CTR for ARM Neoverse V1 and V2
207 * Enable AES and SHA3 optimisations on Applie Silicon M3-based MacOS systems
212 * Added a new EVP_DigestSqueeze() API. This allows SHAKE to squeeze multiple
213 times with different output sizes.
215 *Shane Lontis, Holger Dengler*
217 * Various optimizations for cryptographic routines using RISC-V vector crypto
220 *Christoph Müllner, Charalampos Mitrodimas, Ard Biesheuvel, Phoebe Chen,
223 * Accept longer context for TLS 1.2 exporters
225 While RFC 5705 implies that the maximum length of a context for exporters is
226 65535 bytes as the length is embedded in uint16, the previous implementation
227 enforced a much smaller limit, which is less than 1024 bytes. This
228 restriction has been removed.
235 ### Changes between 3.2.1 and 3.2.2 [xx XXX xxxx]
237 * Fixed an issue where some non-default TLS server configurations can cause
238 unbounded memory growth when processing TLSv1.3 sessions. An attacker may
239 exploit certain server configurations to trigger unbounded memory growth that
240 would lead to a Denial of Service
242 This problem can occur in TLSv1.3 if the non-default SSL_OP_NO_TICKET option
243 is being used (but not if early_data is also configured and the default
244 anti-replay protection is in use). In this case, under certain conditions,
245 the session cache can get into an incorrect state and it will fail to flush
246 properly as it fills. The session cache will continue to grow in an unbounded
247 manner. A malicious client could deliberately create the scenario for this
248 failure to force a Denial of Service. It may also happen by accident in
255 * Fixed bug where SSL_export_keying_material() could not be used with QUIC
256 connections. (#23560)
260 ### Changes between 3.2.0 and 3.2.1 [30 Jan 2024]
262 * A file in PKCS12 format can contain certificates and keys and may come from
263 an untrusted source. The PKCS12 specification allows certain fields to be
264 NULL, but OpenSSL did not correctly check for this case. A fix has been
265 applied to prevent a NULL pointer dereference that results in OpenSSL
266 crashing. If an application processes PKCS12 files from an untrusted source
267 using the OpenSSL APIs then that application will be vulnerable to this
268 issue prior to this fix.
270 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
271 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
272 and PKCS12_newpass().
274 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
275 function is related to writing data we do not consider it security
282 * When function EVP_PKEY_public_check() is called on RSA public keys,
283 a computation is done to confirm that the RSA modulus, n, is composite.
284 For valid RSA keys, n is a product of two or more large primes and this
285 computation completes quickly. However, if n is an overly large prime,
286 then this computation would take a long time.
288 An application that calls EVP_PKEY_public_check() and supplies an RSA key
289 obtained from an untrusted source could be vulnerable to a Denial of Service
292 The function EVP_PKEY_public_check() is not called from other OpenSSL
293 functions however it is called from the OpenSSL pkey command line
294 application. For that reason that application is also vulnerable if used
295 with the "-pubin" and "-check" options on untrusted data.
297 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
298 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
304 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
305 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
310 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
311 for PowerPC CPUs saves the contents of vector registers in different
312 order than they are restored. Thus the contents of some of these vector
313 registers is corrupted when returning to the caller. The vulnerable code is
314 used only on newer PowerPC processors supporting the PowerISA 2.07
317 The consequences of this kind of internal application state corruption can
318 be various - from no consequences, if the calling application does not
319 depend on the contents of non-volatile XMM registers at all, to the worst
320 consequences, where the attacker could get complete control of the
321 application process. However unless the compiler uses the vector registers
322 for storing pointers, the most likely consequence, if any, would be an
323 incorrect result of some application dependent calculations or a crash
324 leading to a denial of service.
330 * Disable building QUIC server utility when OpenSSL is configured with
335 ### Changes between 3.1 and 3.2.0 [23 Nov 2023]
337 * Fix excessive time spent in DH check / generation with large Q parameter
340 Applications that use the functions DH_generate_key() to generate an
341 X9.42 DH key may experience long delays. Likewise, applications that use
342 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
343 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
344 Where the key or parameters that are being checked have been obtained from
345 an untrusted source this may lead to a Denial of Service.
351 * The BLAKE2b hash algorithm supports a configurable output length
352 by setting the "size" parameter.
354 *Čestmír Kalina and Tomáš Mráz*
356 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
360 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
361 and the corresponding provider-storemgmt API function
362 OSSL_FUNC_store_delete().
366 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
367 a passphrase callback when opening a store.
371 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
372 from 8 bytes to 16 bytes.
373 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
374 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
375 requires a salt length of 128 bits. This affects OpenSSL command line
376 applications such as "genrsa" and "pkcs8" and API's such as
377 PEM_write_bio_PrivateKey() that are reliant on the default value.
378 The additional commandline option 'saltlen' has been added to the
379 OpenSSL command line applications for "pkcs8" and "enc" to allow the
380 salt length to be set to a non default value.
384 * Changed the default value of the `ess_cert_id_alg` configuration
385 option which is used to calculate the TSA's public key certificate
386 identifier. The default algorithm is updated to be sha256 instead
389 *Małgorzata Olszówka*
391 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
392 table for point multiplication of the base point, which increases the size of
393 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
394 been added to disable the precomputed table.
398 * Added client side support for QUIC
400 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
402 * Added multiple tutorials on the OpenSSL library and in particular
403 on writing various clients (using TLS and QUIC protocols) with libssl.
407 * Added secp384r1 implementation using Solinas' reduction to improve
408 speed of the NIST P-384 elliptic curve. To enable the implementation
409 the build option `enable-ec_nistp_64_gcc_128` must be used.
413 * Improved RFC7468 compliance of the asn1parse command.
415 *Matthias St. Pierre*
417 * Added SHA256/192 algorithm support.
421 * Added support for securely getting root CA certificate update in
426 * Improved contention on global write locks by using more read locks where
431 * Improved performance of OSSL_PARAM lookups in performance critical
436 * Added the SSL_get0_group_name() function to provide access to the
437 name of the group used for the TLS key exchange.
441 * Provide a new configure option `no-http` that can be used to disable the
442 HTTP support. Provide new configure options `no-apps` and `no-docs` to
443 disable building the openssl command line application and the documentation.
447 * Provide a new configure option `no-ecx` that can be used to disable the
448 X25519, X448, and EdDSA support.
452 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
453 the EVP_KDF_CTX_set_params() function they are now concatenated not just
454 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
458 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
459 the provider context as a parameter.
463 * TLS round-trip time calculation was added by a Brigham Young University
464 Capstone team partnering with Sandia National Laboratories. A new function
465 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
470 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
471 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
472 option. Use of the "advanced" s_client command command via the "-adv" option
477 * Added an "advanced" command mode to s_client. Use this with the "-adv"
478 option. The old "basic" command mode recognises certain letters that must
479 always appear at the start of a line and cannot be escaped. The advanced
480 command mode enables commands to be entered anywhere and there is an
481 escaping mechanism. After starting s_client with "-adv" type "{help}"
482 to show a list of available commands.
486 * Add Raw Public Key (RFC7250) support. Authentication is supported
487 by matching keys against either local policy (TLSA records synthesised
488 from the expected keys) or DANE (TLSA records obtained by the
489 application from DNS). TLSA records will also match the same key in
490 the server certificate, should RPK use not happen to be negotiated.
494 * Added support for modular exponentiation and CRT offloading for the
499 * Added further assembler code for the RISC-V architecture.
503 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
504 from a given EC_GROUP.
508 * Improved support for non-default library contexts and property queries
509 when parsing PKCS#12 files.
513 * Implemented support for all five instances of EdDSA from RFC8032:
514 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
515 The streaming is not yet supported for the HashEdDSA variants
516 (Ed25519ph and Ed448ph).
520 * Added SM4 optimization for ARM processors using ASIMD and AES HW
525 * Implemented SM4-XTS support.
529 * Added platform-agnostic OSSL_sleep() function.
533 * Implemented deterministic ECDSA signatures (RFC6979) support.
537 * Implemented AES-GCM-SIV (RFC8452) support.
541 * Added support for pluggable (provider-based) TLS signature algorithms.
542 This enables TLS 1.3 authentication operations with algorithms embedded
543 in providers not included by default in OpenSSL. In combination with
544 the already available pluggable KEM and X.509 support, this enables
545 for example suitable providers to deliver post-quantum or quantum-safe
546 cryptography to OpenSSL users.
550 * Added support for pluggable (provider-based) CMS signature algorithms.
551 This enables CMS sign and verify operations with algorithms embedded
552 in providers not included by default in OpenSSL.
556 * Added support for Hybrid Public Key Encryption (HPKE) as defined
557 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
558 Message Layer Security (MLS) and other IETF specifications.
559 HPKE can also be used by other applications that require
560 encrypting "to" an ECDH public key. External APIs are defined in
561 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
565 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
570 * Add support for certificate compression (RFC8879), including
571 library support for Brotli and Zstandard compression.
575 * Add the ability to add custom attributes to PKCS12 files. Add a new API
576 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
577 for a user specified callback and optional argument.
578 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
579 added to the existing STACK_OF attrs.
583 * Major refactor of the libssl record layer.
587 * Add a mac salt length option for the pkcs12 command.
591 * Add more SRTP protection profiles from RFC8723 and RFC8269.
595 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
597 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
599 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
600 supported and enabled.
604 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
605 to the list of ciphersuites providing Perfect Forward Secrecy as
606 required by SECLEVEL >= 3.
608 *Dmitry Belyavskiy, Nicola Tuveri*
610 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
611 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
612 SSL_get0_iana_groups() function-like macro, retrieves the list of
613 supported groups sent by the peer.
614 The function SSL_client_hello_get_extension_order() populates
615 a caller-supplied array with the list of extension types present in the
616 ClientHello, in order of appearance.
620 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
621 to make it possible to use empty passphrase strings.
625 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
629 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
630 to change functions used for allocating the memory of asynchronous call stack.
634 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
638 * A failure exit code is returned when using the openssl x509 command to check
639 certificate attributes and the checks fail.
643 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
644 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
645 of 160 bits and above and less than 224 bits were previously accepted by
646 default but are now no longer allowed. By default TLS compression was
647 already disabled in previous OpenSSL versions. At security level 2 it cannot
652 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
657 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
658 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
659 will need to load the legacy crypto provider.
663 * CCM8 cipher suites in TLS have been downgraded to security level zero
664 because they use a short authentication tag which lowers their strength.
668 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
669 by default. Also spaces surrounding `=` in DN output are removed.
673 * Add X.509 certificate codeSigning purpose and related checks on key usage and
674 extended key usage of the leaf certificate according to the CA/Browser Forum.
678 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
679 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
680 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
681 X.509 version 3 if the certificate information includes X.509 extensions.
685 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
686 such as adding a trace facility for debugging certificate chain building.
690 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
691 in particular supporting requests for central key generation, generalized
692 polling, and various types of genm/genp exchanges defined in CMP Updates.
696 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
697 like correcting the TLS and proxy support and adding tracing for debugging.
701 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
705 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
706 a certificate to be added is already present. `CMS_sign_ex()` and
707 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
708 and no longer throw an error for them.
712 * Fixed and extended `util/check-format.pl` for checking adherence to the
713 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
714 The checks are meanwhile more complete and yield fewer false positives.
718 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
719 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
720 calls. They can be used as the transport BIOs for QUIC.
722 *Hugo Landau, Matt Caswell and Tomáš Mráz*
724 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
725 sending and receiving multiple messages in a single call. An implementation
726 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
730 * Support for loading root certificates from the Windows certificate store
731 has been added. The support is in the form of a store which recognises the
732 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
733 arguments. This store is built by default and can be disabled using the new
734 compile-time option `no-winstore`. This store is not currently used by
735 default and must be loaded explicitly using the above store URI. It is
736 expected to be loaded by default in the future.
740 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
741 kernel versions that support KTLS have a known bug in CCM processing. That
742 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
743 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
748 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
753 * Zerocopy KTLS sendfile() support on Linux.
757 * The OBJ_ calls are now thread safe using a global lock.
761 * New parameter `-digest` for openssl cms command allowing signing
762 pre-computed digests and new CMS API functions supporting that
767 * OPENSSL_malloc() and other allocation functions now raise errors on
768 allocation failures. The callers do not need to explicitly raise errors
769 unless they want to for tracing purposes.
773 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
774 decryption as a protection against Bleichenbacher-like attacks.
775 The RSA decryption API will now return a randomly generated deterministic
776 message instead of an error in case it detects an error when checking
777 padding during PKCS#1 v1.5 decryption. This is a general protection against
778 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
780 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
781 on the RSA decryption context.
785 * Added support for Brainpool curves in TLS-1.3.
787 *Bernd Edlinger and Matt Caswell*
789 * Added OpenBSD specific build targets.
793 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
794 a basic thread pool implementation for select platforms.
801 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
803 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
804 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
805 that alter the key or IV length ([CVE-2023-5363]).
809 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
811 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
813 The POLY1305 MAC (message authentication code) implementation in OpenSSL
814 does not save the contents of non-volatile XMM registers on Windows 64
815 platform when calculating the MAC of data larger than 64 bytes. Before
816 returning to the caller all the XMM registers are set to zero rather than
817 restoring their previous content. The vulnerable code is used only on newer
818 x86_64 processors supporting the AVX512-IFMA instructions.
820 The consequences of this kind of internal application state corruption can
821 be various - from no consequences, if the calling application does not
822 depend on the contents of non-volatile XMM registers at all, to the worst
823 consequences, where the attacker could get complete control of the
824 application process. However given the contents of the registers are just
825 zeroized so the attacker cannot put arbitrary values inside, the most likely
826 consequence, if any, would be an incorrect result of some application
827 dependent calculations or a crash leading to a denial of service.
833 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
835 * Fix excessive time spent checking DH q parameter value.
837 The function DH_check() performs various checks on DH parameters. After
838 fixing CVE-2023-3446 it was discovered that a large q parameter value can
839 also trigger an overly long computation during some of these checks.
840 A correct q value, if present, cannot be larger than the modulus p
841 parameter, thus it is unnecessary to perform these checks if q is larger
844 If DH_check() is called with such q parameter value,
845 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
846 intensive checks are skipped.
852 * Fix DH_check() excessive time with over sized modulus.
854 The function DH_check() performs various checks on DH parameters. One of
855 those checks confirms that the modulus ("p" parameter) is not too large.
856 Trying to use a very large modulus is slow and OpenSSL will not normally use
857 a modulus which is over 10,000 bits in length.
859 However the DH_check() function checks numerous aspects of the key or
860 parameters that have been supplied. Some of those checks use the supplied
861 modulus value even if it has already been found to be too large.
863 A new limit has been added to DH_check of 32,768 bits. Supplying a
864 key/parameters with a modulus over this size will simply cause DH_check() to
871 * Do not ignore empty associated data entries with AES-SIV.
873 The AES-SIV algorithm allows for authentication of multiple associated
874 data entries along with the encryption. To authenticate empty data the
875 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
876 with NULL pointer as the output buffer and 0 as the input buffer length.
877 The AES-SIV implementation in OpenSSL just returns success for such call
878 instead of performing the associated data authentication operation.
879 The empty data thus will not be authenticated. ([CVE-2023-2975])
881 Thanks to Juerg Wullschleger (Google) for discovering the issue.
883 The fix changes the authentication tag value and the ciphertext for
884 applications that use empty associated data entries with AES-SIV.
885 To decrypt data encrypted with previous versions of OpenSSL the application
886 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
891 * When building with the `enable-fips` option and using the resulting
892 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
893 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
894 not operate with truncated digests (FIPS 140-3 IG G.R).
898 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
900 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
901 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
903 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
904 numeric text form. For gigantic sub-identifiers, this would take a very
905 long time, the time complexity being O(n^2) where n is the size of that
906 sub-identifier. ([CVE-2023-2650])
908 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
909 IDENTIFIER to canonical numeric text form if the size of that OBJECT
910 IDENTIFIER is 586 bytes or less, and fail otherwise.
912 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
913 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
914 most 128 sub-identifiers, and that the maximum value that each sub-
915 identifier may have is 2^32-1 (4294967295 decimal).
917 For each byte of every sub-identifier, only the 7 lower bits are part of
918 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
919 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
924 * Multiple algorithm implementation fixes for ARM BE platforms.
928 * Added a -pedantic option to fipsinstall that adjusts the various
929 settings to ensure strict FIPS compliance rather than backwards
934 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
935 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
936 trigger a crash of an application using AES-XTS decryption if the memory
937 just after the buffer being decrypted is not mapped.
938 Thanks to Anton Romanov (Amazon) for discovering the issue.
943 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
944 The previous fix for this timing side channel turned out to cause
945 a severe 2-3x performance regression in the typical use case
946 compared to 3.0.7. The new fix uses existing constant time
947 code paths, and restores the previous performance level while
948 fully eliminating all existing timing side channels.
949 The fix was developed by Bernd Edlinger with testing support
954 * Add FIPS provider configuration option to disallow the use of
955 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
956 The option '-no_drbg_truncated_digests' can optionally be
957 supplied to 'openssl fipsinstall'.
961 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
962 that it does not enable policy checking. Thanks to David Benjamin for
963 discovering this issue.
968 * Fixed an issue where invalid certificate policies in leaf certificates are
969 silently ignored by OpenSSL and other certificate policy checks are skipped
970 for that certificate. A malicious CA could use this to deliberately assert
971 invalid certificate policies in order to circumvent policy checking on the
972 certificate altogether.
977 * Limited the number of nodes created in a policy tree to mitigate
978 against CVE-2023-0464. The default limit is set to 1000 nodes, which
979 should be sufficient for most installations. If required, the limit
980 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
981 time define to a desired maximum number of nodes or zero to allow
987 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
989 * Add FIPS provider configuration option to enforce the
990 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
991 The option '-ems-check' can optionally be supplied to
992 'openssl fipsinstall'.
996 * The FIPS provider includes a few non-approved algorithms for
997 backward compatibility purposes and the "fips=yes" property query
998 must be used for all algorithm fetches to ensure FIPS compliance.
1000 The algorithms that are included but not approved are Triple DES ECB,
1001 Triple DES CBC and EdDSA.
1005 * Added support for KMAC in KBKDF.
1009 * RNDR and RNDRRS support in provider functions to provide
1010 random number generation for Arm CPUs (aarch64).
1014 * s_client and s_server apps now explicitly say when the TLS version
1015 does not include the renegotiation mechanism. This avoids confusion
1016 between that scenario versus when the TLS version includes secure
1017 renegotiation but the peer lacks support for it.
1021 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
1023 *Tomasz Kantecki, Andrey Matyukov*
1025 * The various OBJ_* functions have been made thread safe.
1029 * Parallel dual-prime 1536/2048-bit modular exponentiation for
1030 AVX512_IFMA capable processors.
1032 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1034 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
1035 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
1036 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
1037 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
1038 `OPENSSL_NO_DEPRECATED_3_1`.
1040 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
1041 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
1042 definitions for these functions regardless of whether
1043 `OPENSSL_NO_DEPRECATED_3_1` is defined.
1045 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
1046 functions regardless of whether they are using them. It is recommended that
1047 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
1051 * When generating safe-prime DH parameters set the recommended private key
1052 length equivalent to minimum key lengths as in RFC 7919.
1056 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
1057 maximum size that is smaller or equal to the digest length to comply with
1058 FIPS 186-4 section 5. This is implemented by a new option
1059 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
1060 `rsa_pss_saltlen` parameter, which is now the default. Signature
1061 verification is not affected by this change and continues to work as before.
1068 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
1069 listed here are only a brief description.
1070 The migration guide contains more detailed information related to new features,
1071 breaking changes, and mappings for the large list of deprecated functions.
1073 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
1075 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
1077 * Fixed NULL dereference during PKCS7 data verification.
1079 A NULL pointer can be dereferenced when signatures are being
1080 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
1081 algorithm used for the signature is known to the OpenSSL library but
1082 the implementation of the hash algorithm is not available the digest
1083 initialization will fail. There is a missing check for the return
1084 value from the initialization function which later leads to invalid
1085 usage of the digest API most likely leading to a crash.
1088 PKCS7 data is processed by the SMIME library calls and also by the
1089 time stamp (TS) library calls. The TLS implementation in OpenSSL does
1090 not call these functions however third party applications would be
1091 affected if they call these functions to verify signatures on untrusted
1096 * Fixed X.400 address type confusion in X.509 GeneralName.
1098 There is a type confusion vulnerability relating to X.400 address processing
1099 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
1100 but the public structure definition for GENERAL_NAME incorrectly specified
1101 the type of the x400Address field as ASN1_TYPE. This field is subsequently
1102 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
1103 than an ASN1_STRING.
1105 When CRL checking is enabled (i.e. the application sets the
1106 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
1107 pass arbitrary pointers to a memcmp call, enabling them to read memory
1108 contents or enact a denial of service.
1113 * Fixed NULL dereference validating DSA public key.
1115 An invalid pointer dereference on read can be triggered when an
1116 application tries to check a malformed DSA public key by the
1117 EVP_PKEY_public_check() function. This will most likely lead
1118 to an application crash. This function can be called on public
1119 keys supplied from untrusted sources which could allow an attacker
1120 to cause a denial of service attack.
1122 The TLS implementation in OpenSSL does not call this function
1123 but applications might call the function if there are additional
1124 security requirements imposed by standards such as FIPS 140-3.
1127 *Shane Lontis, Tomáš Mráz*
1129 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
1131 An invalid pointer dereference on read can be triggered when an
1132 application tries to load malformed PKCS7 data with the
1133 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
1135 The result of the dereference is an application crash which could
1136 lead to a denial of service attack. The TLS implementation in OpenSSL
1137 does not call this function however third party applications might
1138 call these functions on untrusted data.
1143 * Fixed Use-after-free following BIO_new_NDEF.
1145 The public API function BIO_new_NDEF is a helper function used for
1146 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
1147 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
1148 be called directly by end user applications.
1150 The function receives a BIO from the caller, prepends a new BIO_f_asn1
1151 filter BIO onto the front of it to form a BIO chain, and then returns
1152 the new head of the BIO chain to the caller. Under certain conditions,
1153 for example if a CMS recipient public key is invalid, the new filter BIO
1154 is freed and the function returns a NULL result indicating a failure.
1155 However, in this case, the BIO chain is not properly cleaned up and the
1156 BIO passed by the caller still retains internal pointers to the previously
1157 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
1158 then a use-after-free will occur. This will most likely result in a crash.
1161 *Viktor Dukhovni, Matt Caswell*
1163 * Fixed Double free after calling PEM_read_bio_ex.
1165 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
1166 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
1167 data. If the function succeeds then the "name_out", "header" and "data"
1168 arguments are populated with pointers to buffers containing the relevant
1169 decoded data. The caller is responsible for freeing those buffers. It is
1170 possible to construct a PEM file that results in 0 bytes of payload data.
1171 In this case PEM_read_bio_ex() will return a failure code but will populate
1172 the header argument with a pointer to a buffer that has already been freed.
1173 If the caller also frees this buffer then a double free will occur. This
1174 will most likely lead to a crash.
1176 The functions PEM_read_bio() and PEM_read() are simple wrappers around
1177 PEM_read_bio_ex() and therefore these functions are also directly affected.
1179 These functions are also called indirectly by a number of other OpenSSL
1180 functions including PEM_X509_INFO_read_bio_ex() and
1181 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
1182 internal uses of these functions are not vulnerable because the caller does
1183 not free the header argument if PEM_read_bio_ex() returns a failure code.
1186 *Kurt Roeckx, Matt Caswell*
1188 * Fixed Timing Oracle in RSA Decryption.
1190 A timing based side channel exists in the OpenSSL RSA Decryption
1191 implementation which could be sufficient to recover a plaintext across
1192 a network in a Bleichenbacher style attack. To achieve a successful
1193 decryption an attacker would have to be able to send a very large number
1194 of trial messages for decryption. The vulnerability affects all RSA padding
1195 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1198 *Dmitry Belyavsky, Hubert Kario*
1200 * Fixed X.509 Name Constraints Read Buffer Overflow.
1202 A read buffer overrun can be triggered in X.509 certificate verification,
1203 specifically in name constraint checking. The read buffer overrun might
1204 result in a crash which could lead to a denial of service attack.
1205 In a TLS client, this can be triggered by connecting to a malicious
1206 server. In a TLS server, this can be triggered if the server requests
1207 client authentication and a malicious client connects.
1212 * Fixed X.509 Policy Constraints Double Locking security issue.
1214 If an X.509 certificate contains a malformed policy constraint and
1215 policy processing is enabled, then a write lock will be taken twice
1216 recursively. On some operating systems (most widely: Windows) this
1217 results in a denial of service when the affected process hangs. Policy
1218 processing being enabled on a publicly facing server is not considered
1219 to be a common setup.
1224 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1225 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1226 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1227 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1228 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1229 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1230 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1231 for legacy EC and SM2 keys is also changed similarly to honor the
1232 equivalent conversion format flag as specified in the underlying
1233 `EC_KEY` object being exported to a provider, when this function is
1234 called through `EVP_PKEY_export()`.
1238 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1240 * Fixed two buffer overflows in punycode decoding functions.
1242 A buffer overrun can be triggered in X.509 certificate verification,
1243 specifically in name constraint checking. Note that this occurs after
1244 certificate chain signature verification and requires either a CA to
1245 have signed the malicious certificate or for the application to continue
1246 certificate verification despite failure to construct a path to a trusted
1249 In a TLS client, this can be triggered by connecting to a malicious
1250 server. In a TLS server, this can be triggered if the server requests
1251 client authentication and a malicious client connects.
1253 An attacker can craft a malicious email address to overflow
1254 an arbitrary number of bytes containing the `.` character (decimal 46)
1255 on the stack. This buffer overflow could result in a crash (causing a
1259 An attacker can craft a malicious email address to overflow four
1260 attacker-controlled bytes on the stack. This buffer overflow could
1261 result in a crash (causing a denial of service) or potentially remote code
1262 execution depending on stack layout for any given platform/compiler.
1267 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1268 parameters in OpenSSL code.
1269 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1270 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1271 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1272 Using these invalid names may cause algorithms to use slower methods
1273 that ignore the CRT parameters.
1277 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1282 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1283 data to be signed before signing the certificate.
1287 * Added RIPEMD160 to the default provider.
1291 * Ensured that the key share group sent or accepted for the key exchange
1292 is allowed for the protocol version.
1296 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1298 * OpenSSL supports creating a custom cipher via the legacy
1299 EVP_CIPHER_meth_new() function and associated function calls. This function
1300 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1301 to use the new provider mechanism in order to implement custom ciphers.
1303 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1304 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1305 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1306 and decryption initialisation functions). Instead of using the custom cipher
1307 directly it incorrectly tries to fetch an equivalent cipher from the
1308 available providers. An equivalent cipher is found based on the NID passed to
1309 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1310 given cipher. However it is possible for an application to incorrectly pass
1311 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1312 is used in this way the OpenSSL encryption/decryption initialisation function
1313 will match the NULL cipher as being equivalent and will fetch this from the
1314 available providers. This will succeed if the default provider has been
1315 loaded (or if a third party provider has been loaded that offers this
1316 cipher). Using the NULL cipher means that the plaintext is emitted as the
1319 Applications are only affected by this issue if they call
1320 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1321 encryption/decryption initialisation function. Applications that only use
1322 SSL/TLS are not impacted by this issue.
1327 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1332 * Fixed the linux-mips64 Configure target which was missing the
1333 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1338 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1343 * Correctly handle a retransmitted ClientHello in DTLS
1347 * Fixed detection of ktls support in cross-compile environment on Linux
1351 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1356 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1357 report correct results in some cases
1361 * Fix UWP builds by defining VirtualLock
1365 * For known safe primes use the minimum key length according to RFC 7919.
1366 Longer private key sizes unnecessarily raise the cycles needed to compute the
1367 shared secret without any increase of the real security. This fixes a
1368 regression from 1.1.1 where these shorter keys were generated for the known
1373 * Added the loongarch64 target
1377 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1378 only passed to the FIPS provider and not to the default or legacy provider.
1382 * Fixed reported performance degradation on aarch64. Restored the
1383 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1384 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1385 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1386 The new algorithm is still used for 32 bit targets.
1390 * Added a missing header for memcmp that caused compilation failure on some
1395 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1397 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1398 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1399 This issue makes the RSA implementation with 2048 bit private keys
1400 incorrect on such machines and memory corruption will happen during
1401 the computation. As a consequence of the memory corruption an attacker
1402 may be able to trigger a remote code execution on the machine performing
1405 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1406 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1407 are affected by this issue.
1412 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1413 implementation would not encrypt the entirety of the data under some
1414 circumstances. This could reveal sixteen bytes of data that was
1415 preexisting in the memory that wasn't written. In the special case of
1416 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1418 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1419 they are both unaffected.
1422 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1424 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1426 * In addition to the c_rehash shell command injection identified in
1427 CVE-2022-1292, further bugs where the c_rehash script does not
1428 properly sanitise shell metacharacters to prevent command injection have been
1431 When the CVE-2022-1292 was fixed it was not discovered that there
1432 are other places in the script where the file names of certificates
1433 being hashed were possibly passed to a command executed through the shell.
1435 This script is distributed by some operating systems in a manner where
1436 it is automatically executed. On such operating systems, an attacker
1437 could execute arbitrary commands with the privileges of the script.
1439 Use of the c_rehash script is considered obsolete and should be replaced
1440 by the OpenSSL rehash command line tool.
1443 *Daniel Fiala, Tomáš Mráz*
1445 * Case insensitive string comparison no longer uses locales. It has instead
1446 been directly implemented.
1450 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1452 * Case insensitive string comparison is reimplemented via new locale-agnostic
1453 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1454 comparison. The previous implementation had problems when the Turkish locale
1459 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1460 metacharacters to prevent command injection. This script is distributed by
1461 some operating systems in a manner where it is automatically executed. On
1462 such operating systems, an attacker could execute arbitrary commands with the
1463 privileges of the script.
1465 Use of the c_rehash script is considered obsolete and should be replaced
1466 by the OpenSSL rehash command line tool.
1471 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1472 certificate on an OCSP response. The bug caused the function in the case
1473 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1474 response (meaning a successful verification) even in the case where the
1475 response signing certificate fails to verify.
1477 It is anticipated that most users of `OCSP_basic_verify` will not use the
1478 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1479 a negative value (indicating a fatal error) in the case of a certificate
1480 verification failure. The normal expected return value in this case would be
1483 This issue also impacts the command line OpenSSL "ocsp" application. When
1484 verifying an ocsp response with the "-no_cert_checks" option the command line
1485 application will report that the verification is successful even though it
1486 has in fact failed. In this case the incorrect successful response will also
1487 be accompanied by error messages showing the failure and contradicting the
1488 apparently successful result.
1493 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1494 AAD data as the MAC key. This made the MAC key trivially predictable.
1496 An attacker could exploit this issue by performing a man-in-the-middle attack
1497 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1498 that the modified data would still pass the MAC integrity check.
1500 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1501 endpoint will always be rejected by the recipient and the connection will
1502 fail at that point. Many application protocols require data to be sent from
1503 the client to the server first. Therefore, in such a case, only an OpenSSL
1504 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1506 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1507 sent in both directions. In this case both clients and servers could be
1508 affected, regardless of the application protocol.
1510 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1511 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1512 the handshake when using this ciphersuite.
1514 The confidentiality of data is not impacted by this issue, i.e. an attacker
1515 cannot decrypt data that has been encrypted using this ciphersuite - they can
1518 In order for this attack to work both endpoints must legitimately negotiate
1519 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1520 OpenSSL 3.0, and is not available within the default provider or the default
1521 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1522 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1523 following must have occurred:
1525 1) OpenSSL must have been compiled with the (non-default) compile time option
1526 enable-weak-ssl-ciphers
1528 2) OpenSSL must have had the legacy provider explicitly loaded (either
1529 through application code or via configuration)
1531 3) The ciphersuite must have been explicitly added to the ciphersuite list
1533 4) The libssl security level must have been set to 0 (default is 1)
1535 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1537 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1538 others that both endpoints have in common
1543 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1544 occupied by the removed hash table entries.
1546 This function is used when decoding certificates or keys. If a long lived
1547 process periodically decodes certificates or keys its memory usage will
1548 expand without bounds and the process might be terminated by the operating
1549 system causing a denial of service. Also traversing the empty hash table
1550 entries will take increasingly more time.
1552 Typically such long lived processes might be TLS clients or TLS servers
1553 configured to accept client certificate authentication.
1556 *Hugo Landau, Aliaksei Levin*
1558 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1559 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1560 statistics are no longer supported. For compatibility, these statistics are
1561 still listed in the output but are now always reported as zero.
1565 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1567 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1568 for non-prime moduli.
1570 Internally this function is used when parsing certificates that contain
1571 elliptic curve public keys in compressed form or explicit elliptic curve
1572 parameters with a base point encoded in compressed form.
1574 It is possible to trigger the infinite loop by crafting a certificate that
1575 has invalid explicit curve parameters.
1577 Since certificate parsing happens prior to verification of the certificate
1578 signature, any process that parses an externally supplied certificate may thus
1579 be subject to a denial of service attack. The infinite loop can also be
1580 reached when parsing crafted private keys as they can contain explicit
1581 elliptic curve parameters.
1583 Thus vulnerable situations include:
1585 - TLS clients consuming server certificates
1586 - TLS servers consuming client certificates
1587 - Hosting providers taking certificates or private keys from customers
1588 - Certificate authorities parsing certification requests from subscribers
1589 - Anything else which parses ASN.1 elliptic curve parameters
1591 Also any other applications that use the BN_mod_sqrt() where the attacker
1592 can control the parameter values are vulnerable to this DoS issue.
1597 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1598 to the list of ciphersuites providing Perfect Forward Secrecy as
1599 required by SECLEVEL >= 3.
1601 *Dmitry Belyavskiy, Nicola Tuveri*
1603 * Made the AES constant time code for no-asm configurations
1604 optional due to the resulting 95% performance degradation.
1605 The AES constant time code can be enabled, for no assembly
1606 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1610 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1615 * The negative return value handling of the certificate verification callback
1616 was reverted. The replacement is to set the verification retry state with
1617 the SSL_set_retry_verify() function.
1621 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1623 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1624 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1625 verify a certificate supplied by a server. That function may return a
1626 negative return value to indicate an internal error (for example out of
1627 memory). Such a negative return value is mishandled by OpenSSL and will cause
1628 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1629 success and a subsequent call to SSL_get_error() to return the value
1630 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1631 returned by OpenSSL if the application has previously called
1632 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1633 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1634 totally unexpected and applications may not behave correctly as a result. The
1635 exact behaviour will depend on the application but it could result in
1636 crashes, infinite loops or other similar incorrect responses.
1638 This issue is made more serious in combination with a separate bug in OpenSSL
1639 3.0 that will cause X509_verify_cert() to indicate an internal error when
1640 processing a certificate chain. This will occur where a certificate does not
1641 include the Subject Alternative Name extension but where a Certificate
1642 Authority has enforced name constraints. This issue can occur even with valid
1648 * Corrected a few file name and file reference bugs in the build,
1649 installation and setup scripts, which lead to installation verification
1650 failures. Slightly enhanced the installation verification script.
1654 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1659 * Fixed PVK encoder to properly query for the passphrase.
1663 * Multiple fixes in the OSSL_HTTP API functions.
1667 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1668 OSSL_PARAM_INTEGER data type and return error on negative numbers
1669 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1670 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1674 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1678 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1682 * Multiple threading fixes.
1686 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1690 * Allow fetching an operation from the provider that owns an unexportable key
1691 as a fallback if that is still allowed by the property query.
1695 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1697 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1702 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1703 S390X capability vector to zero. This simplifies testing of different code
1704 paths on S390X architecture.
1708 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1709 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1710 SP 800-38D". The communication will fail at this point.
1714 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1715 confidential in EC_GROUP data.
1719 * The byte order mark (BOM) character is ignored if encountered at the
1720 beginning of a PEM-formatted file.
1724 * Added CMS support for the Russian GOST algorithms.
1728 * Due to move of the implementation of cryptographic operations
1729 to the providers, validation of various operation parameters can
1730 be postponed until the actual operation is executed where previously
1731 it happened immediately when an operation parameter was set.
1733 For example when setting an unsupported curve with
1734 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1735 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1737 *OpenSSL team members and many third party contributors*
1739 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1740 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1741 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1742 instead to retrieve these algorithms from a provider.
1746 * On build targets where the multilib postfix is set in the build
1747 configuration the libdir directory was changing based on whether
1748 the lib directory with the multilib postfix exists on the system
1749 or not. This unpredictable behavior was removed and eventual
1750 multilib postfix is now always added to the default libdir. Use
1751 `--libdir=lib` to override the libdir if adding the postfix is
1756 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1757 no longer interoperable with OpenSSL 1.1.1.
1761 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1762 function codes, this function can only cause problems for calling
1767 * Add a configurable flag to output date formats as ISO 8601. Does not
1768 change the default date format.
1772 * Version of MSVC earlier than 1300 could get link warnings, which could
1773 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1774 Support for this flag has been removed.
1778 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1779 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1780 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1781 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1782 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1786 * The signatures of the functions to get and set options on SSL and
1787 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1788 Some source code changes may be required.
1792 * The public definitions of conf_method_st and conf_st have been
1793 deprecated. They will be made opaque in a future release.
1795 *Rich Salz and Tomáš Mráz*
1797 * Client-initiated renegotiation is disabled by default. To allow it, use
1798 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1799 flag, or the "ClientRenegotiation" config parameter as appropriate.
1803 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1804 or modify relative pathname inclusion.
1808 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1809 validated. Please consult the README-FIPS and
1810 README-PROVIDERS files, as well as the migration guide.
1812 *OpenSSL team members and many third party contributors*
1814 * For the key types DH and DHX the allowed settable parameters are now different.
1818 * The openssl commands that read keys, certificates, and CRLs now
1819 automatically detect the PEM or DER format of the input files.
1821 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1823 * Added enhanced PKCS#12 APIs which accept a library context.
1827 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1831 * Added support for Kernel TLS (KTLS).
1833 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1835 * Support for RFC 5746 secure renegotiation is now required by default for
1836 SSL or TLS connections to succeed.
1840 * The signature of the `copy` functional parameter of the
1841 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1842 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1843 the signature of the `pub_decode` functional parameter of the
1844 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1845 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1849 * The error return values from some control calls (ctrl) have changed.
1853 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1857 * Many functions in the EVP_ namespace that are getters of values from
1858 implementations or contexts were renamed to include get or get0 in their
1859 names. Old names are provided as macro aliases for compatibility and
1864 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1865 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1866 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1871 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1874 * The output from the command line applications may have minor
1879 * The output from numerous "printing" may have minor changes.
1883 * Windows thread synchronization uses read/write primitives (SRWLock) when
1884 supported by the OS, otherwise CriticalSection continues to be used.
1888 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1889 work on read only BIO source/sinks that do not support these functions.
1890 This allows piping or redirection of a file BIO using stdin to be buffered
1891 into memory. This is used internally in OSSL_DECODER_from_bio().
1895 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1896 this function would return one of the values OSSL_STORE_INFO_NAME,
1897 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1898 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1899 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1900 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1901 using this function should be amended to handle the changed return value.
1905 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1906 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1907 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1908 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1909 contains more than one certificate identifier: This means that all
1910 certificates referenced there MUST be part of the validation chain.
1914 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1915 RC5, DESX and DES have been moved to the legacy provider.
1919 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1920 RIPEMD-160 have been moved to the legacy provider.
1924 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1929 * The deprecated functions EVP_PKEY_get0_RSA(),
1930 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1931 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1932 well as the similarly named "get1" functions behave differently in
1937 * A number of functions handling low-level keys or engines were deprecated
1938 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1939 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1940 EVP_PKEY_get0_siphash().
1944 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1945 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1946 will need to load the legacy crypto provider. This includes these PBE
1947 algorithms which use this KDF:
1948 - NID_pbeWithMD2AndDES_CBC
1949 - NID_pbeWithMD5AndDES_CBC
1950 - NID_pbeWithSHA1AndRC2_CBC
1951 - NID_pbeWithMD2AndRC2_CBC
1952 - NID_pbeWithMD5AndRC2_CBC
1953 - NID_pbeWithSHA1AndDES_CBC
1957 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1958 BIO_debug_callback() functions.
1962 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1963 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1967 * The RAND_METHOD APIs have been deprecated.
1971 * The SRP APIs have been deprecated.
1975 * Add a compile time option to prevent the caching of provider fetched
1976 algorithms. This is enabled by including the no-cached-fetch option
1977 at configuration time.
1981 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1982 count of PKCS12_DEFAULT_ITER.
1984 *Tomáš Mráz and Sahana Prasad*
1986 * The openssl speed command does not use low-level API calls anymore.
1990 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1993 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1995 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1999 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
2000 providers may supply their own group implementations (using either the "key
2001 exchange" or the "key encapsulation" methods) which will automatically be
2002 detected and used by libssl.
2004 *Matt Caswell, Nicola Tuveri*
2006 * The undocumented function X509_certificate_type() has been deprecated;
2010 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
2014 * Removed RSA padding mode for SSLv23 (which was only used for
2015 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
2016 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
2021 * Deprecated the obsolete X9.31 RSA key generation related functions.
2023 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
2024 is not allowed to return a value > 1, this is no more taken as failure.
2026 *Viktor Dukhovni and David von Oheimb*
2028 * Deprecated the obsolete X9.31 RSA key generation related functions
2029 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
2030 BN_X931_generate_prime_ex().
2034 * The default key generation method for the regular 2-prime RSA keys was
2035 changed to the FIPS 186-4 B.3.6 method.
2039 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
2043 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
2047 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
2048 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
2050 *Rich Salz, Richard Levitte, and David von Oheimb*
2052 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
2056 * Deprecated `OCSP_parse_url()`.
2060 * Validation of SM2 keys has been separated from the validation of regular EC
2065 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
2066 switches: a validation failure triggers an early exit, returning a failure
2067 exit status to the parent process.
2071 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
2072 to ignore unknown ciphers.
2076 * The `-cipher-commands` and `-digest-commands` options
2077 of the command line utility `list` have been deprecated.
2078 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
2082 * Added convenience functions for generating asymmetric key pairs:
2083 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
2084 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
2088 * All of the low-level EC_KEY functions have been deprecated.
2090 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
2092 * Deprecated all the libcrypto and libssl error string loading
2097 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
2098 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
2103 * The `-crypt` option to the `passwd` command line tool has been removed.
2107 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
2112 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
2116 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
2117 EVP_PKEY_get1_tls_encodedpoint().
2121 * The security callback, which can be customised by application code, supports
2122 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
2123 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
2127 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
2128 interface. Their functionality remains unchanged.
2132 * Added new option for 'openssl list', '-providers', which will display the
2133 list of loaded providers, their names, version and status. It optionally
2134 displays their gettable parameters.
2138 * Removed EVP_PKEY_set_alias_type().
2142 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
2143 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
2147 * Changed all "STACK" functions to be macros instead of inline functions. Macro
2148 parameters are still checked for type safety at compile time via helper
2153 * Remove the RAND_DRBG API
2155 *Paul Dale and Matthias St. Pierre*
2157 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
2158 as well as actual hostnames.
2162 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2163 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2164 conversely, silently ignore DTLS protocol version bounds when configuring
2165 TLS-based contexts. The commands can be repeated to set bounds of both
2166 types. The same applies with the corresponding "min_protocol" and
2167 "max_protocol" command-line switches, in case some application uses both TLS
2170 SSL_CTX instances that are created for a fixed protocol version (e.g.
2171 `TLSv1_server_method()`) also silently ignore version bounds. Previously
2172 attempts to apply bounds to these protocol versions would result in an
2173 error. Now only the "version-flexible" SSL_CTX instances are subject to
2174 limits in configuration files in command-line options.
2178 * Deprecated the `ENGINE` API. Engines should be replaced with providers
2183 * Reworked the recorded ERR codes to make better space for system errors.
2184 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
2185 given code is a system error (true) or an OpenSSL error (false).
2189 * Reworked the test perl framework to better allow parallel testing.
2191 *Nicola Tuveri and David von Oheimb*
2193 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2194 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2198 * 'Configure' has been changed to figure out the configuration target if
2199 none is given on the command line. Consequently, the 'config' script is
2200 now only a mere wrapper. All documentation is changed to only mention
2203 *Rich Salz and Richard Levitte*
2205 * Added a library context `OSSL_LIB_CTX` that applications as well as
2206 other libraries can use to form a separate context within which
2207 libcrypto operations are performed.
2211 * Added various `_ex` functions to the OpenSSL API that support using
2212 a non-default `OSSL_LIB_CTX`.
2216 * Handshake now fails if Extended Master Secret extension is dropped
2221 * Dropped interactive mode from the `openssl` program.
2225 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2227 *David von Oheimb and Shane Lontis*
2229 * Deprecated `EC_METHOD_get_field_type()`.
2233 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2234 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2235 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2239 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2243 * Add CAdES-BES signature verification support, mostly derived
2244 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2246 *Filipe Raimundo da Silva*
2248 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2252 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2253 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2257 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2261 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2262 EC_KEY_precompute_mult().
2266 * Deprecated EC_POINTs_mul().
2270 * Removed FIPS_mode() and FIPS_mode_set().
2274 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2278 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2279 EC_POINT_get_Jprojective_coordinates_GFp().
2283 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2284 arrays to be more easily constructed via a series of utility functions.
2285 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2286 the various push functions and finally convert to a passable OSSL_PARAM
2287 array using OSSL_PARAM_BLD_to_param().
2291 * The security strength of SHA1 and MD5 based signatures in TLS has been
2296 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2297 contain a provider side internal key.
2301 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2305 * Project text documents not yet having a proper file name extension
2306 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2307 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2311 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2312 have been converted to Markdown with the goal to produce documents
2313 which not only look pretty when viewed online in the browser, but
2314 remain well readable inside a plain text editor.
2316 To achieve this goal, a 'minimalistic' Markdown style has been applied
2317 which avoids formatting elements that interfere too much with the
2318 reading flow in the text file. For example, it
2320 * avoids [ATX headings][] and uses [setext headings][] instead
2321 (which works for `<h1>` and `<h2>` headings only).
2322 * avoids [inline links][] and uses [reference links][] instead.
2323 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2325 [ATX headings]: https://github.github.com/gfm/#atx-headings
2326 [setext headings]: https://github.github.com/gfm/#setext-headings
2327 [inline links]: https://github.github.com/gfm/#inline-link
2328 [reference links]: https://github.github.com/gfm/#reference-link
2329 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2330 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2332 *Matthias St. Pierre*
2334 * The test suite is changed to preserve results of each test recipe.
2335 A new directory test-runs/ with subdirectories named like the
2336 test recipes are created in the build tree for this purpose.
2340 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2341 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2342 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2344 *David von Oheimb, Martin Peylo*
2346 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2347 It supports arbitrary request and response content types, GET redirection,
2348 TLS, connections via HTTP(S) proxies, connections and exchange via
2349 user-defined BIOs (allowing implicit connections), persistent connections,
2350 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2351 The legacy OCSP-focused (and only partly documented) API
2352 is retained for backward compatibility, while most of it is deprecated.
2356 * Added `util/check-format.pl`, a tool for checking adherence to the
2357 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2358 The checks performed are incomplete and yield some false positives.
2359 Still the tool should be useful for detecting most typical glitches.
2363 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2364 If domain name resolution yields multiple IP addresses all of them are tried
2365 after `connect()` failures.
2369 * All of the low-level RSA functions have been deprecated.
2373 * X509 certificates signed using SHA1 are no longer allowed at security
2378 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2379 modified to use PKEY APIs. These commands are now in maintenance mode
2380 and no new features will be added to them.
2384 * The command line utility rsautl has been deprecated.
2388 * The command line utilities genrsa and rsa have been modified to use PKEY
2389 APIs. They now write PKCS#8 keys by default. These commands are now in
2390 maintenance mode and no new features will be added to them.
2394 * All of the low-level DH functions have been deprecated.
2396 *Paul Dale and Matt Caswell*
2398 * All of the low-level DSA functions have been deprecated.
2402 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2403 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2407 * Deprecated low-level ECDH and ECDSA functions.
2411 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2415 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2416 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2417 a new formulation to include all the things it can be used for,
2418 as well as words of caution.
2422 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2426 * All of the low-level HMAC functions have been deprecated.
2428 *Paul Dale and David von Oheimb*
2430 * Over two thousand fixes were made to the documentation, including:
2431 - Common options (such as -rand/-writerand, TLS version control, etc)
2432 were refactored and point to newly-enhanced descriptions in openssl.pod.
2433 - Added style conformance for all options (with help from Richard Levitte),
2434 documented all reported missing options, added a CI build to check
2435 that all options are documented and that no unimplemented options
2437 - Documented some internals, such as all use of environment variables.
2438 - Addressed all internal broken L<> references.
2442 * All of the low-level CMAC functions have been deprecated.
2446 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2447 functions have been deprecated.
2449 *Paul Dale and David von Oheimb*
2451 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2452 set of functions. The documentation mentioned negative values for some
2453 errors, but this was never the case, so the mention of negative values
2456 Code that followed the documentation and thereby check with something
2457 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2461 * All of the low-level cipher functions have been deprecated.
2463 *Matt Caswell and Paul Dale*
2465 * Removed include/openssl/opensslconf.h.in and replaced it with
2466 include/openssl/configuration.h.in, which differs in not including
2467 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2468 was added to include both.
2470 This allows internal hacks where one might need to modify the set
2471 of configured macros, for example this if deprecated symbols are
2472 still supposed to be available internally:
2474 #include <openssl/configuration.h>
2476 #undef OPENSSL_NO_DEPRECATED
2477 #define OPENSSL_SUPPRESS_DEPRECATED
2479 #include <openssl/macros.h>
2481 This should not be used by applications that use the exported
2482 symbols, as that will lead to linking errors.
2486 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2487 used in exponentiation with 512-bit moduli. No EC algorithms are
2488 affected. Analysis suggests that attacks against 2-prime RSA1024,
2489 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2490 difficult to perform and are not believed likely. Attacks against DH512
2491 are considered just feasible. However, for an attack the target would
2492 have to reuse the DH512 private key, which is not recommended anyway.
2493 Also applications directly using the low-level API BN_mod_exp may be
2494 affected if they use BN_FLG_CONSTTIME.
2499 * Most memory-debug features have been deprecated, and the functionality
2500 replaced with no-ops.
2504 * Added documentation for the STACK API.
2508 * Introduced a new method type and API, OSSL_ENCODER, to represent
2509 generic encoders. These do the same sort of job that PEM writers
2510 and d2i functions do, but with support for methods supplied by
2511 providers, and the possibility for providers to support other
2516 * Introduced a new method type and API, OSSL_DECODER, to represent
2517 generic decoders. These do the same sort of job that PEM readers
2518 and i2d functions do, but with support for methods supplied by
2519 providers, and the possibility for providers to support other
2524 * Added a .pragma directive to the syntax of configuration files, to
2525 allow varying behavior in a supported and predictable manner.
2526 Currently added pragma:
2530 This allows dollar signs to be a keyword character unless it's
2531 followed by a opening brace or parenthesis. This is useful for
2532 platforms where dollar signs are commonly used in names, such as
2533 volume names and system directory names on VMS.
2537 * Added functionality to create an EVP_PKEY from user data.
2541 * Change the interpretation of the '--api' configuration option to
2542 mean that this is a desired API compatibility level with no
2543 further meaning. The previous interpretation, that this would
2544 also mean to remove all deprecated symbols up to and including
2545 the given version, no requires that 'no-deprecated' is also used
2546 in the configuration.
2548 When building applications, the desired API compatibility level
2549 can be set with the OPENSSL_API_COMPAT macro like before. For
2550 API compatibility version below 3.0, the old style numerical
2551 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2552 For version 3.0 and on, the value is expected to be the decimal
2553 value calculated from the major and minor version like this:
2555 MAJOR * 10000 + MINOR * 100
2559 -DOPENSSL_API_COMPAT=30000 For 3.0
2560 -DOPENSSL_API_COMPAT=30200 For 3.2
2562 To hide declarations that are deprecated up to and including the
2563 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2564 given when building the application as well.
2568 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2569 access to certificate and CRL stores via URIs and OSSL_STORE
2572 This adds the following functions:
2574 - X509_LOOKUP_store()
2575 - X509_STORE_load_file()
2576 - X509_STORE_load_path()
2577 - X509_STORE_load_store()
2578 - SSL_add_store_cert_subjects_to_stack()
2579 - SSL_CTX_set_default_verify_store()
2580 - SSL_CTX_load_verify_file()
2581 - SSL_CTX_load_verify_dir()
2582 - SSL_CTX_load_verify_store()
2586 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2587 The presence of this system service is determined at run-time.
2591 * Added functionality to create an EVP_PKEY context based on data
2592 for methods from providers. This takes an algorithm name and a
2593 property query string and simply stores them, with the intent
2594 that any operation that uses this context will use those strings
2595 to fetch the needed methods implicitly, thereby making the port
2596 of application written for pre-3.0 OpenSSL easier.
2600 * The undocumented function NCONF_WIN32() has been deprecated; for
2601 conversion details see the HISTORY section of doc/man5/config.pod
2605 * Introduced the new functions EVP_DigestSignInit_ex() and
2606 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2607 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2608 pages for further details.
2612 * Over two thousand fixes were made to the documentation, including:
2613 adding missing command flags, better style conformance, documentation
2616 *Rich Salz, Richard Levitte*
2618 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2619 X25519, X448, Ed25519 and Ed448.
2623 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2628 * Deprecated the public definition of `ERR_STATE` as well as the function
2629 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2634 * Added ERR functionality to give callers access to the stored function
2635 names that have replaced the older function code based functions.
2637 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2638 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2639 ERR_peek_error_all() and ERR_peek_last_error_all().
2641 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2642 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2643 ERR_func_error_string().
2647 * Extended testing to be verbose for failing tests only. The make variables
2648 VERBOSE_FAILURE or VF can be used to enable this:
2650 $ make VF=1 test # Unix
2651 $ mms /macro=(VF=1) test ! OpenVMS
2652 $ nmake VF=1 test # Windows
2656 * Added the `-copy_extensions` option to the `x509` command for use with
2657 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2658 all extensions in the request are copied to the certificate or vice versa.
2660 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2662 * Added the `-copy_extensions` option to the `req` command for use with
2663 `-x509`. When given with the `copy` or `copyall` argument,
2664 all extensions in the certification request are copied to the certificate.
2668 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2669 they generate are by default RFC 5280 compliant in the following sense:
2670 There is a subjectKeyIdentifier extension with a hash value of the public key
2671 and for not self-signed certs there is an authorityKeyIdentifier extension
2672 with a keyIdentifier field or issuer information identifying the signing key.
2673 This is done unless some configuration overrides the new default behavior,
2674 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2678 * Added several checks to `X509_verify_cert()` according to requirements in
2679 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2680 (which may be done by using the CLI option `-x509_strict`):
2681 * The basicConstraints of CA certificates must be marked critical.
2682 * CA certificates must explicitly include the keyUsage extension.
2683 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2684 * The issuer name of any certificate must not be empty.
2685 * The subject name of CA certs, certs with keyUsage crlSign,
2686 and certs without subjectAlternativeName must not be empty.
2687 * If a subjectAlternativeName extension is given it must not be empty.
2688 * The signatureAlgorithm field and the cert signature must be consistent.
2689 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2690 must not be marked critical.
2691 * The authorityKeyIdentifier must be given for X.509v3 certs
2692 unless they are self-signed.
2693 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2697 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2698 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2702 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2703 used even when parsing explicit parameters, when loading a encoded key
2704 or calling `EC_GROUP_new_from_ecpkparameters()`/
2705 `EC_GROUP_new_from_ecparameters()`.
2706 This prevents bypass of security hardening and performance gains,
2707 especially for curves with specialized EC_METHODs.
2708 By default, if a key encoded with explicit parameters is loaded and later
2709 encoded, the output is still encoded with explicit parameters, even if
2710 internally a "named" EC_GROUP is used for computation.
2714 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2715 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2716 NULL. After this change, only the cofactor parameter can be NULL. It also
2717 does some minimal sanity checks on the passed order.
2722 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2723 An attack is simple, if the first CMS_recipientInfo is valid but the
2724 second CMS_recipientInfo is chosen ciphertext. If the second
2725 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2726 encryption key will be replaced by garbage, and the message cannot be
2727 decoded, but if the RSA decryption fails, the correct encryption key is
2728 used and the recipient will not notice the attack.
2729 As a work around for this potential attack the length of the decrypted
2730 key must be equal to the cipher default key length, in case the
2731 certificate is not given and all recipientInfo are tried out.
2732 The old behaviour can be re-enabled in the CMS code by setting the
2733 CMS_DEBUG_DECRYPT flag.
2737 * Early start up entropy quality from the DEVRANDOM seed source has been
2738 improved for older Linux systems. The RAND subsystem will wait for
2739 /dev/random to be producing output before seeding from /dev/urandom.
2740 The seeded state is stored for future library initialisations using
2741 a system global shared memory segment. The shared memory identifier
2742 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2743 the desired value. The default identifier is 114.
2747 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2748 when primes for RSA keys are computed.
2749 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2750 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2751 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2752 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2753 This avoids possible fingerprinting of newly generated RSA modules.
2757 * Correct the extended master secret constant on EBCDIC systems. Without this
2758 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2759 negotiate EMS will fail. Unfortunately this also means that TLS connections
2760 between EBCDIC systems with this fix, and EBCDIC systems without this
2761 fix will fail if they negotiate EMS.
2765 * Changed the library initialisation so that the config file is now loaded
2766 by default. This was already the case for libssl. It now occurs for both
2767 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2768 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2772 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2773 where the former acts as a replacement for `ERR_put_error()`, and the
2774 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2775 `ERR_raise_data()` adds more flexibility by taking a format string and
2776 an arbitrary number of arguments following it, to be processed with
2781 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2782 to check if a named provider is loaded and available. When called, it
2783 will also activate all fallback providers if such are still present.
2787 * Enforce a minimum DH modulus size of 512 bits.
2791 * Changed DH parameters to generate the order q subgroup instead of 2q.
2792 Previously generated DH parameters are still accepted by DH_check
2793 but DH_generate_key works around that by clearing bit 0 of the
2794 private key for those. This avoids leaking bit 0 of the private key.
2798 * Significantly reduce secure memory usage by the randomness pools.
2802 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2807 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2808 algorithms. An implementation of a key exchange algorithm can be obtained
2809 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2810 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2811 the older EVP_PKEY_derive_init() function. See the man pages for the new
2812 functions for further details.
2816 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2820 * Removed the function names from error messages and deprecated the
2825 * Removed NextStep support and the macro OPENSSL_UNISTD
2829 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2830 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2831 Also removed "export var as function" capability; we do not export
2832 variables, only functions.
2836 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2837 an error and 1 indicating success. In previous versions of OpenSSL this
2838 was a void type. If a key was set longer than the maximum possible this
2843 * Support SM2 signing and verification schemes with X509 certificate.
2847 * Use SHA256 as the default digest for TS query in the `ts` app.
2851 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2855 * Default cipher lists/suites are now available via a function, the
2856 #defines are deprecated.
2860 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2861 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2862 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2866 * Join the directories crypto/x509 and crypto/x509v3
2870 * Added command 'openssl kdf' that uses the EVP_KDF API.
2874 * Added command 'openssl mac' that uses the EVP_MAC API.
2878 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2879 as default directories. Also added the command 'openssl info'
2880 for scripting purposes.
2884 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2889 * Add prediction resistance to the DRBG reseeding process.
2893 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2894 mandated by IEEE Std 1619-2018.
2898 * Added newline escaping functionality to a filename when using openssl dgst.
2899 This output format is to replicate the output format found in the `*sum`
2900 checksum programs. This aims to preserve backward compatibility.
2902 *Matt Eaton, Richard Levitte, and Paul Dale*
2904 * Removed the heartbeat message in DTLS feature, as it has very
2905 little usage and doesn't seem to fulfill a valuable purpose.
2906 The configuration option is now deprecated.
2910 * Changed the output of 'openssl {digestname} < file' to display the
2911 digest name in its output.
2915 * Added a new generic trace API which provides support for enabling
2916 instrumentation through trace output.
2918 *Richard Levitte & Matthias St. Pierre*
2920 * Added build tests for C++. These are generated files that only do one
2921 thing, to include one public OpenSSL head file each. This tests that
2922 the public header files can be usefully included in a C++ application.
2924 This test isn't enabled by default. It can be enabled with the option
2925 'enable-buildtest-c++'.
2929 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2933 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2937 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2941 * Added KMAC to EVP_MAC.
2945 * Added property based algorithm implementation selection framework to
2950 * Added SCA hardening for modular field inversion in EC_GROUP through
2951 a new dedicated field_inv() pointer in EC_METHOD.
2952 This also addresses a leakage affecting conversions from projective
2953 to affine coordinates.
2955 *Billy Bob Brumley, Nicola Tuveri*
2957 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2958 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2959 those algorithms that were already supported through the EVP_PKEY API
2960 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2961 and scrypt are now wrappers that call EVP_KDF.
2965 * Build devcrypto engine as a dynamic engine.
2967 *Eneas U de Queiroz*
2969 * Add keyed BLAKE2 to EVP_MAC.
2973 * Fix a bug in the computation of the endpoint-pair shared secret used
2974 by DTLS over SCTP. This breaks interoperability with older versions
2975 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2976 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2977 interoperability with such broken implementations. However, enabling
2978 this switch breaks interoperability with correct implementations.
2980 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2981 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2985 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2989 * Changed the license to the Apache License v2.0.
2993 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2995 - Major releases (indicated by incrementing the MAJOR release number)
2996 may introduce incompatible API/ABI changes.
2997 - Minor releases (indicated by incrementing the MINOR release number)
2998 may introduce new features but retain API/ABI compatibility.
2999 - Patch releases (indicated by incrementing the PATCH number)
3000 are intended for bug fixes and other improvements of existing
3001 features only (like improving performance or adding documentation)
3002 and retain API/ABI compatibility.
3006 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
3010 * Remove the 'dist' target and add a tarball building script. The
3011 'dist' target has fallen out of use, and it shouldn't be
3012 necessary to configure just to create a source distribution.
3016 * Recreate the OS390-Unix config target. It no longer relies on a
3017 special script like it did for OpenSSL pre-1.1.0.
3021 * Instead of having the source directories listed in Configure, add
3022 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
3027 * Add GMAC to EVP_MAC.
3031 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
3035 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
3036 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
3037 to facilitate the continued use of MACs through raw private keys in
3038 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
3042 * Deprecate ECDH_KDF_X9_62().
3046 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
3047 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
3048 are retained for backwards compatibility.
3052 * AES-XTS mode now enforces that its two keys are different to mitigate
3053 the attacked described in "Efficient Instantiations of Tweakable
3054 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
3055 Details of this attack can be obtained from:
3056 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
3060 * Rename the object files, i.e. give them other names than in previous
3061 versions. Their names now include the name of the final product, as
3062 well as its type mnemonic (bin, lib, shlib).
3066 * Added new option for 'openssl list', '-objects', which will display the
3067 list of built in objects, i.e. OIDs with names.
3071 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
3072 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
3077 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
3078 improves application performance by removing data copies and providing
3079 applications with zero-copy system calls such as sendfile and splice.
3083 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
3087 * `PKCS12_parse` now maintains the order of the parsed certificates
3088 when outputting them via `*ca` (rather than reversing it).
3092 * Deprecated pthread fork support methods.
3096 * Added support for FFDHE key exchange in TLS 1.3.
3100 * Added a new concept for OpenSSL plugability: providers. This
3101 functionality is designed to replace the ENGINE API and ENGINE
3102 implementations, and to be much more dynamic, allowing provider
3103 authors to introduce new algorithms among other things, as long as
3104 there's an API that supports the algorithm type.
3106 With this concept comes a new core API for interaction between
3107 libcrypto and provider implementations. Public libcrypto functions
3108 that want to use providers do so through this core API.
3110 The main documentation for this core API is found in
3111 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
3112 refer to other manuals describing the API specific for supported
3113 algorithm types (also called operations).
3120 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
3122 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
3124 * Avoid loading of a dynamic engine twice.
3128 * Prioritise DANE TLSA issuer certs over peer certs
3132 * Fixed random API for MacOS prior to 10.12
3134 These MacOS versions don't support the CommonCrypto APIs
3138 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
3140 * Fixed an SM2 Decryption Buffer Overflow.
3142 In order to decrypt SM2 encrypted data an application is expected to
3143 call the API function EVP_PKEY_decrypt(). Typically an application will
3144 call this function twice. The first time, on entry, the "out" parameter
3145 can be NULL and, on exit, the "outlen" parameter is populated with the
3146 buffer size required to hold the decrypted plaintext. The application
3147 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
3148 again, but this time passing a non-NULL value for the "out" parameter.
3150 A bug in the implementation of the SM2 decryption code means that the
3151 calculation of the buffer size required to hold the plaintext returned
3152 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
3153 size required by the second call. This can lead to a buffer overflow
3154 when EVP_PKEY_decrypt() is called by the application a second time with
3155 a buffer that is too small.
3157 A malicious attacker who is able present SM2 content for decryption to
3158 an application could cause attacker chosen data to overflow the buffer
3159 by up to a maximum of 62 bytes altering the contents of other data held
3160 after the buffer, possibly changing application behaviour or causing
3161 the application to crash. The location of the buffer is application
3162 dependent but is typically heap allocated.
3167 * Fixed various read buffer overruns processing ASN.1 strings
3169 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
3170 structure which contains a buffer holding the string data and a field
3171 holding the buffer length. This contrasts with normal C strings which
3172 are represented as a buffer for the string data which is terminated
3173 with a NUL (0) byte.
3175 Although not a strict requirement, ASN.1 strings that are parsed using
3176 OpenSSL's own "d2i" functions (and other similar parsing functions) as
3177 well as any string whose value has been set with the ASN1_STRING_set()
3178 function will additionally NUL terminate the byte array in the
3179 ASN1_STRING structure.
3181 However, it is possible for applications to directly construct valid
3182 ASN1_STRING structures which do not NUL terminate the byte array by
3183 directly setting the "data" and "length" fields in the ASN1_STRING
3184 array. This can also happen by using the ASN1_STRING_set0() function.
3186 Numerous OpenSSL functions that print ASN.1 data have been found to
3187 assume that the ASN1_STRING byte array will be NUL terminated, even
3188 though this is not guaranteed for strings that have been directly
3189 constructed. Where an application requests an ASN.1 structure to be
3190 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3191 been directly constructed by the application without NUL terminating
3192 the "data" field, then a read buffer overrun can occur.
3194 The same thing can also occur during name constraints processing
3195 of certificates (for example if a certificate has been directly
3196 constructed by the application instead of loading it via the OpenSSL
3197 parsing functions, and the certificate contains non NUL terminated
3198 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3199 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3201 If a malicious actor can cause an application to directly construct an
3202 ASN1_STRING and then process it through one of the affected OpenSSL
3203 functions then this issue could be hit. This might result in a crash
3204 (causing a Denial of Service attack). It could also result in the
3205 disclosure of private memory contents (such as private keys, or
3206 sensitive plaintext).
3211 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3213 * Fixed a problem with verifying a certificate chain when using the
3214 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3215 the certificates present in a certificate chain. It is not set by default.
3217 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3218 the chain that have explicitly encoded elliptic curve parameters was added
3219 as an additional strict check.
3221 An error in the implementation of this check meant that the result of a
3222 previous check to confirm that certificates in the chain are valid CA
3223 certificates was overwritten. This effectively bypasses the check
3224 that non-CA certificates must not be able to issue other certificates.
3226 If a "purpose" has been configured then there is a subsequent opportunity
3227 for checks that the certificate is a valid CA. All of the named "purpose"
3228 values implemented in libcrypto perform this check. Therefore, where
3229 a purpose is set the certificate chain will still be rejected even when the
3230 strict flag has been used. A purpose is set by default in libssl client and
3231 server certificate verification routines, but it can be overridden or
3232 removed by an application.
3234 In order to be affected, an application must explicitly set the
3235 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3236 for the certificate verification or, in the case of TLS client or server
3237 applications, override the default purpose.
3242 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3243 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3244 renegotiation ClientHello omits the signature_algorithms extension (where it
3245 was present in the initial ClientHello), but includes a
3246 signature_algorithms_cert extension then a NULL pointer dereference will
3247 result, leading to a crash and a denial of service attack.
3249 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3250 (which is the default configuration). OpenSSL TLS clients are not impacted by
3254 *Peter Kästle and Samuel Sapalski*
3256 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3258 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3259 create a unique hash value based on the issuer and serial number data
3260 contained within an X509 certificate. However, it was failing to correctly
3261 handle any errors that may occur while parsing the issuer field (which might
3262 occur if the issuer field is maliciously constructed). This may subsequently
3263 result in a NULL pointer deref and a crash leading to a potential denial of
3269 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3270 padding mode to correctly check for rollback attacks. This is considered a
3271 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3276 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3277 functions. Previously they could overflow the output length argument in some
3278 cases where the input length is close to the maximum permissible length for
3279 an integer on the platform. In such cases the return value from the function
3280 call would be 1 (indicating success), but the output length value would be
3281 negative. This could cause applications to behave incorrectly or crash.
3286 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3287 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3288 could be exploited in a side channel attack to recover the password. Since
3289 the attack is local host only this is outside of the current OpenSSL
3290 threat model and therefore no CVE is assigned.
3292 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3297 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3299 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3300 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3301 If an attacker can control both items being compared then this could lead
3302 to a possible denial of service attack. OpenSSL itself uses the
3303 GENERAL_NAME_cmp function for two purposes:
3304 1) Comparing CRL distribution point names between an available CRL and a
3305 CRL distribution point embedded in an X509 certificate
3306 2) When verifying that a timestamp response token signer matches the
3307 timestamp authority name (exposed via the API functions
3308 TS_RESP_verify_response and TS_RESP_verify_token)
3313 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3315 * Certificates with explicit curve parameters are now disallowed in
3316 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3320 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3321 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3322 conversely, silently ignore DTLS protocol version bounds when configuring
3323 TLS-based contexts. The commands can be repeated to set bounds of both
3324 types. The same applies with the corresponding "min_protocol" and
3325 "max_protocol" command-line switches, in case some application uses both TLS
3328 SSL_CTX instances that are created for a fixed protocol version (e.g.
3329 TLSv1_server_method()) also silently ignore version bounds. Previously
3330 attempts to apply bounds to these protocol versions would result in an
3331 error. Now only the "version-flexible" SSL_CTX instances are subject to
3332 limits in configuration files in command-line options.
3336 * Handshake now fails if Extended Master Secret extension is dropped
3341 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3343 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3345 * Fixed segmentation fault in SSL_check_chain()
3346 Server or client applications that call the SSL_check_chain() function
3347 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3348 dereference as a result of incorrect handling of the
3349 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3350 or unrecognised signature algorithm is received from the peer. This could
3351 be exploited by a malicious peer in a Denial of Service attack.
3356 * Added AES consttime code for no-asm configurations
3357 an optional constant time support for AES was added
3358 when building openssl for no-asm.
3359 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3360 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3361 At this time this feature is by default disabled.
3362 It will be enabled by default in 3.0.
3366 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3368 * Revert the change of EOF detection while reading in libssl to avoid
3369 regressions in applications depending on the current way of reporting
3370 the EOF. As the existing method is not fully accurate the change to
3371 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3372 branch and will be present in the 3.0 release.
3376 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3377 when primes for RSA keys are computed.
3378 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3379 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3380 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3381 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3382 This avoids possible fingerprinting of newly generated RSA modules.
3386 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3388 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3389 while reading in libssl then we would report an error back to the
3390 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3391 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3392 therefore give a hint as to what went wrong.
3396 * Check that ed25519 and ed448 are allowed by the security level. Previously
3397 signature algorithms not using an MD were not being checked that they were
3398 allowed by the security level.
3402 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3403 was not quite right. The behaviour was not consistent between resumption
3404 and normal handshakes, and also not quite consistent with historical
3405 behaviour. The behaviour in various scenarios has been clarified and
3406 it has been updated to make it match historical behaviour as closely as
3411 * *[VMS only]* The header files that the VMS compilers include automatically,
3412 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3413 that the C++ compiler doesn't understand. This is a shortcoming in the
3414 compiler, but can be worked around with `__cplusplus` guards.
3416 C++ applications that use OpenSSL libraries must be compiled using the
3417 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3418 functions. Otherwise, only functions with symbols of less than 31
3419 characters can be used, as the linker will not be able to successfully
3420 resolve symbols with longer names.
3424 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3425 The presence of this system service is determined at run-time.
3429 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3434 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3436 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3437 number generator (RNG). This was intended to include protection in the
3438 event of a fork() system call in order to ensure that the parent and child
3439 processes did not share the same RNG state. However, this protection was not
3440 being used in the default case.
3442 A partial mitigation for this issue is that the output from a high
3443 precision timer is mixed into the RNG state so the likelihood of a parent
3444 and child process sharing state is significantly reduced.
3446 If an application already calls OPENSSL_init_crypto() explicitly using
3447 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3450 *Matthias St. Pierre*
3452 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3453 used even when parsing explicit parameters, when loading a encoded key
3454 or calling `EC_GROUP_new_from_ecpkparameters()`/
3455 `EC_GROUP_new_from_ecparameters()`.
3456 This prevents bypass of security hardening and performance gains,
3457 especially for curves with specialized EC_METHODs.
3458 By default, if a key encoded with explicit parameters is loaded and later
3459 encoded, the output is still encoded with explicit parameters, even if
3460 internally a "named" EC_GROUP is used for computation.
3464 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3465 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3466 NULL. After this change, only the cofactor parameter can be NULL. It also
3467 does some minimal sanity checks on the passed order.
3472 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3473 An attack is simple, if the first CMS_recipientInfo is valid but the
3474 second CMS_recipientInfo is chosen ciphertext. If the second
3475 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3476 encryption key will be replaced by garbage, and the message cannot be
3477 decoded, but if the RSA decryption fails, the correct encryption key is
3478 used and the recipient will not notice the attack.
3479 As a work around for this potential attack the length of the decrypted
3480 key must be equal to the cipher default key length, in case the
3481 certificate is not given and all recipientInfo are tried out.
3482 The old behaviour can be re-enabled in the CMS code by setting the
3483 CMS_DEBUG_DECRYPT flag.
3488 * Early start up entropy quality from the DEVRANDOM seed source has been
3489 improved for older Linux systems. The RAND subsystem will wait for
3490 /dev/random to be producing output before seeding from /dev/urandom.
3491 The seeded state is stored for future library initialisations using
3492 a system global shared memory segment. The shared memory identifier
3493 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3494 the desired value. The default identifier is 114.
3498 * Correct the extended master secret constant on EBCDIC systems. Without this
3499 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3500 negotiate EMS will fail. Unfortunately this also means that TLS connections
3501 between EBCDIC systems with this fix, and EBCDIC systems without this
3502 fix will fail if they negotiate EMS.
3506 * Use Windows installation paths in the mingw builds
3508 Mingw isn't a POSIX environment per se, which means that Windows
3509 paths should be used for installation.
3514 * Changed DH_check to accept parameters with order q and 2q subgroups.
3515 With order 2q subgroups the bit 0 of the private key is not secret
3516 but DH_generate_key works around that by clearing bit 0 of the
3517 private key for those. This avoids leaking bit 0 of the private key.
3521 * Significantly reduce secure memory usage by the randomness pools.
3525 * Revert the DEVRANDOM_WAIT feature for Linux systems
3527 The DEVRANDOM_WAIT feature added a select() call to wait for the
3528 /dev/random device to become readable before reading from the
3529 /dev/urandom device.
3531 It turned out that this change had negative side effects on
3532 performance which were not acceptable. After some discussion it
3533 was decided to revert this feature and leave it up to the OS
3534 resp. the platform maintainer to ensure a proper initialization
3535 during early boot time.
3537 *Matthias St. Pierre*
3539 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3541 * Add build tests for C++. These are generated files that only do one
3542 thing, to include one public OpenSSL head file each. This tests that
3543 the public header files can be usefully included in a C++ application.
3545 This test isn't enabled by default. It can be enabled with the option
3546 'enable-buildtest-c++'.
3550 * Enable SHA3 pre-hashing for ECDSA and DSA.
3554 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3555 This changes the size when using the `genpkey` command when no size is given.
3556 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3557 generation commands to use 2048 bits by default.
3561 * Reorganize the manual pages to consistently have RETURN VALUES,
3562 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3563 util/fix-doc-nits accordingly.
3565 *Paul Yang, Joshua Lock*
3567 * Add the missing accessor EVP_PKEY_get0_engine()
3571 * Have commands like `s_client` and `s_server` output the signature scheme
3572 along with other cipher suite parameters when debugging.
3576 * Make OPENSSL_config() error agnostic again.
3580 * Do the error handling in RSA decryption constant time.
3584 * Prevent over long nonces in ChaCha20-Poly1305.
3586 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3587 for every encryption operation. RFC 7539 specifies that the nonce value
3588 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3589 and front pads the nonce with 0 bytes if it is less than 12
3590 bytes. However it also incorrectly allows a nonce to be set of up to 16
3591 bytes. In this case only the last 12 bytes are significant and any
3592 additional leading bytes are ignored.
3594 It is a requirement of using this cipher that nonce values are
3595 unique. Messages encrypted using a reused nonce value are susceptible to
3596 serious confidentiality and integrity attacks. If an application changes
3597 the default nonce length to be longer than 12 bytes and then makes a
3598 change to the leading bytes of the nonce expecting the new value to be a
3599 new unique nonce then such an application could inadvertently encrypt
3600 messages with a reused nonce.
3602 Additionally the ignored bytes in a long nonce are not covered by the
3603 integrity guarantee of this cipher. Any application that relies on the
3604 integrity of these ignored leading bytes of a long nonce may be further
3605 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3606 is safe because no such use sets such a long nonce value. However user
3607 applications that use this cipher directly and set a non-default nonce
3608 length to be longer than 12 bytes may be vulnerable.
3610 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3616 * Add DEVRANDOM_WAIT feature for Linux systems
3618 On older Linux systems where the getrandom() system call is not available,
3619 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3620 Contrary to getrandom(), the /dev/urandom device will not block during
3621 early boot when the kernel CSPRNG has not been seeded yet.
3623 To mitigate this known weakness, use select() to wait for /dev/random to
3624 become readable before reading from /dev/urandom.
3626 * Ensure that SM2 only uses SM3 as digest algorithm
3630 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3632 * Change the info callback signals for the start and end of a post-handshake
3633 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3634 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3635 confused by this and assume that a TLSv1.2 renegotiation has started. This
3636 can break KeyUpdate handling. Instead we no longer signal the start and end
3637 of a post handshake message exchange (although the messages themselves are
3638 still signalled). This could break some applications that were expecting
3639 the old signals. However without this KeyUpdate is not usable for many
3644 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3646 * Timing vulnerability in DSA signature generation
3648 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3649 timing side channel attack. An attacker could use variations in the signing
3650 algorithm to recover the private key.
3652 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3657 * Timing vulnerability in ECDSA signature generation
3659 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3660 timing side channel attack. An attacker could use variations in the signing
3661 algorithm to recover the private key.
3663 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3668 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3669 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3670 of two gigabytes and the error handling improved.
3672 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3673 categorized as a normal bug, not a security issue, because the DRBG reseeds
3674 automatically and is fully functional even without additional randomness
3675 provided by the application.
3677 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3679 * Add a new ClientHello callback. Provides a callback interface that gives
3680 the application the ability to adjust the nascent SSL object at the
3681 earliest stage of ClientHello processing, immediately after extensions have
3682 been collected but before they have been processed. In particular, this
3683 callback can adjust the supported TLS versions in response to the contents
3688 * Add SM2 base algorithm support.
3692 * s390x assembly pack: add (improved) hardware-support for the following
3693 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3694 aes-cfb/cfb8, aes-ecb.
3698 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3699 parameter is no longer accepted, as it leads to a corrupt table. NULL
3700 pem_str is reserved for alias entries only.
3704 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3705 step for prime curves. The new implementation is based on formulae from
3706 differential addition-and-doubling in homogeneous projective coordinates
3707 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3708 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3709 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3710 to work in projective coordinates.
3712 *Billy Bob Brumley, Nicola Tuveri*
3714 * Change generating and checking of primes so that the error rate of not
3715 being prime depends on the intended use based on the size of the input.
3716 For larger primes this will result in more rounds of Miller-Rabin.
3717 The maximal error rate for primes with more than 1080 bits is lowered
3720 *Kurt Roeckx, Annie Yousar*
3722 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3726 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3727 moving between systems, and to avoid confusion when a Windows build is
3728 done with mingw vs with MSVC. For POSIX installs, there's still a
3729 symlink or copy named 'tsget' to avoid that confusion as well.
3733 * Revert blinding in ECDSA sign and instead make problematic addition
3734 length-invariant. Switch even to fixed-length Montgomery multiplication.
3738 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3739 step for binary curves. The new implementation is based on formulae from
3740 differential addition-and-doubling in mixed Lopez-Dahab projective
3741 coordinates, modified to independently blind the operands.
3743 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3745 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3746 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3747 EC_METHODs to implement their own specialized "ladder step", to take
3748 advantage of more favorable coordinate systems or more efficient
3749 differential addition-and-doubling algorithms.
3751 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3753 * Modified the random device based seed sources to keep the relevant
3754 file descriptors open rather than reopening them on each access.
3755 This allows such sources to operate in a chroot() jail without
3756 the associated device nodes being available. This behaviour can be
3757 controlled using RAND_keep_random_devices_open().
3761 * Numerous side-channel attack mitigations have been applied. This may have
3762 performance impacts for some algorithms for the benefit of improved
3763 security. Specific changes are noted in this change log by their respective
3768 * AIX shared library support overhaul. Switch to AIX "natural" way of
3769 handling shared libraries, which means collecting shared objects of
3770 different versions and bitnesses in one common archive. This allows to
3771 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3772 doesn't affect the way 3rd party applications are linked, only how
3773 multi-version installation is managed.
3777 * Make ec_group_do_inverse_ord() more robust and available to other
3778 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3779 mitigations are applied to the fallback BN_mod_inverse().
3780 When using this function rather than BN_mod_inverse() directly, new
3781 EC cryptosystem implementations are then safer-by-default.
3785 * Add coordinate blinding for EC_POINT and implement projective
3786 coordinate blinding for generic prime curves as a countermeasure to
3787 chosen point SCA attacks.
3789 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3791 * Add blinding to ECDSA and DSA signatures to protect against side channel
3792 attacks discovered by Keegan Ryan (NCC Group).
3796 * Enforce checking in the `pkeyutl` command to ensure that the input
3797 length does not exceed the maximum supported digest length when performing
3798 a sign, verify or verifyrecover operation.
3802 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3803 I/O in combination with something like select() or poll() will hang. This
3804 can be turned off again using SSL_CTX_clear_mode().
3805 Many applications do not properly handle non-application data records, and
3806 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3807 around the problems in those applications, but can also break some.
3808 It's recommended to read the manpages about SSL_read(), SSL_write(),
3809 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3810 SSL_CTX_set_read_ahead() again.
3814 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3815 now allow empty (zero character) pass phrases.
3819 * Apply blinding to binary field modular inversion and remove patent
3820 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3824 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3825 binary and prime elliptic curves.
3829 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3830 constant time fixed point multiplication.
3834 * Revise elliptic curve scalar multiplication with timing attack
3835 defenses: ec_wNAF_mul redirects to a constant time implementation
3836 when computing fixed point and variable point multiplication (which
3837 in OpenSSL are mostly used with secret scalars in keygen, sign,
3838 ECDH derive operations).
3839 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3842 * Updated CONTRIBUTING
3846 * Updated DRBG / RAND to request nonce and additional low entropy
3847 randomness from the system.
3849 *Matthias St. Pierre*
3851 * Updated 'openssl rehash' to use OpenSSL consistent default.
3855 * Moved the load of the ssl_conf module to libcrypto, which helps
3856 loading engines that libssl uses before libssl is initialised.
3860 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3864 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3866 *Ingo Schwarze, Rich Salz*
3868 * Added output of accepting IP address and port for 'openssl s_server'
3872 * Added a new API for TLSv1.3 ciphersuites:
3873 SSL_CTX_set_ciphersuites()
3874 SSL_set_ciphersuites()
3878 * Memory allocation failures consistently add an error to the error
3883 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3884 in libcrypto when run as setuid/setgid.
3888 * Load any config file by default when libssl is used.
3892 * Added new public header file <openssl/rand_drbg.h> and documentation
3893 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3895 *Matthias St. Pierre*
3897 * QNX support removed (cannot find contributors to get their approval
3898 for the license change).
3902 * TLSv1.3 replay protection for early data has been implemented. See the
3903 SSL_read_early_data() man page for further details.
3907 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3908 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3909 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3910 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3911 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3912 configuration has been separated out. See the ciphers man page or the
3913 SSL_CTX_set_ciphersuites() man page for more information.
3917 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3918 in responder mode now supports the new "-multi" option, which
3919 spawns the specified number of child processes to handle OCSP
3920 requests. The "-timeout" option now also limits the OCSP
3921 responder's patience to wait to receive the full client request
3922 on a newly accepted connection. Child processes are respawned
3923 as needed, and the CA index file is automatically reloaded
3924 when changed. This makes it possible to run the "ocsp" responder
3925 as a long-running service, making the OpenSSL CA somewhat more
3926 feature-complete. In this mode, most diagnostic messages logged
3927 after entering the event loop are logged via syslog(3) rather than
3932 * Added support for X448 and Ed448. Heavily based on original work by
3937 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3938 objects loaded. This adds the functions OSSL_STORE_expect() and
3939 OSSL_STORE_find() as well as needed tools to construct searches and
3940 get the search data out of them.
3944 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3945 version of OpenSSL should review their configuration settings to ensure
3946 that they are still appropriate for TLSv1.3. For further information see:
3947 <https://wiki.openssl.org/index.php/TLS1.3>
3951 * Grand redesign of the OpenSSL random generator
3953 The default RAND method now utilizes an AES-CTR DRBG according to
3954 NIST standard SP 800-90Ar1. The new random generator is essentially
3955 a port of the default random generator from the OpenSSL FIPS 2.0
3956 object module. It is a hybrid deterministic random bit generator
3957 using an AES-CTR bit stream and which seeds and reseeds itself
3958 automatically using trusted system entropy sources.
3960 Some of its new features are:
3961 - Support for multiple DRBG instances with seed chaining.
3962 - The default RAND method makes use of a DRBG.
3963 - There is a public and private DRBG instance.
3964 - The DRBG instances are fork-safe.
3965 - Keep all global DRBG instances on the secure heap if it is enabled.
3966 - The public and private DRBG instance are per thread for lock free
3969 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3971 * Changed Configure so it only says what it does and doesn't dump
3972 so much data. Instead, ./configdata.pm should be used as a script
3973 to display all sorts of configuration data.
3977 * Added processing of "make variables" to Configure.
3981 * Added SHA512/224 and SHA512/256 algorithm support.
3985 * The last traces of Netware support, first removed in 1.1.0, have
3990 * Get rid of Makefile.shared, and in the process, make the processing
3991 of certain files (rc.obj, or the .def/.map/.opt files produced from
3992 the ordinal files) more visible and hopefully easier to trace and
3993 debug (or make silent).
3997 * Make it possible to have environment variable assignments as
3998 arguments to config / Configure.
4002 * Add multi-prime RSA (RFC 8017) support.
4006 * Add SM3 implemented according to GB/T 32905-2016
4007 *Jack Lloyd <jack.lloyd@ribose.com>,*
4008 *Ronald Tse <ronald.tse@ribose.com>,*
4009 *Erick Borsboom <erick.borsboom@ribose.com>*
4011 * Add 'Maximum Fragment Length' TLS extension negotiation and support
4012 as documented in RFC6066.
4013 Based on a patch from Tomasz Moń
4015 *Filipe Raimundo da Silva*
4017 * Add SM4 implemented according to GB/T 32907-2016.
4018 *Jack Lloyd <jack.lloyd@ribose.com>,*
4019 *Ronald Tse <ronald.tse@ribose.com>,*
4020 *Erick Borsboom <erick.borsboom@ribose.com>*
4022 * Reimplement -newreq-nodes and ERR_error_string_n; the
4023 original author does not agree with the license change.
4027 * Add ARIA AEAD TLS support.
4031 * Some macro definitions to support VS6 have been removed. Visual
4032 Studio 6 has not worked since 1.1.0
4036 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
4037 without clearing the errors.
4041 * Add "atfork" functions. If building on a system that without
4042 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
4043 requirements. The RAND facility now uses/requires this.
4051 * The UI API becomes a permanent and integral part of libcrypto, i.e.
4052 not possible to disable entirely. However, it's still possible to
4053 disable the console reading UI method, UI_OpenSSL() (use UI_null()
4056 To disable, configure with 'no-ui-console'. 'no-ui' is still
4057 possible to use as an alias. Check at compile time with the
4058 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
4059 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
4063 * Add a STORE module, which implements a uniform and URI based reader of
4064 stores that can contain keys, certificates, CRLs and numerous other
4065 objects. The main API is loosely based on a few stdio functions,
4066 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
4067 OSSL_STORE_error and OSSL_STORE_close.
4068 The implementation uses backends called "loaders" to implement arbitrary
4069 URI schemes. There is one built in "loader" for the 'file' scheme.
4073 * Add devcrypto engine. This has been implemented against cryptodev-linux,
4074 then adjusted to work on FreeBSD 8.4 as well.
4075 Enable by configuring with 'enable-devcryptoeng'. This is done by default
4076 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
4080 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
4081 util/mkerr.pl, which is adapted to allow those prefixes, leading to
4082 error code calls like this:
4084 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
4086 With this change, we claim the namespaces OSSL and OPENSSL in a manner
4087 that can be encoded in C. For the foreseeable future, this will only
4090 *Richard Levitte and Tim Hudson*
4092 * Removed BSD cryptodev engine.
4096 * Add a build target 'build_all_generated', to build all generated files
4097 and only that. This can be used to prepare everything that requires
4098 things like perl for a system that lacks perl and then move everything
4099 to that system and do the rest of the build there.
4103 * In the UI interface, make it possible to duplicate the user data. This
4104 can be used by engines that need to retain the data for a longer time
4105 than just the call where this user data is passed.
4109 * Ignore the '-named_curve auto' value for compatibility of applications
4112 *Tomáš Mráz <tmraz@fedoraproject.org>*
4114 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
4115 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
4116 alerts across multiple records (some of which could be empty). In practice
4117 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
4118 prohibits this altogether and other libraries (BoringSSL, NSS) do not
4119 support this at all. Supporting it adds significant complexity to the
4120 record layer, and its removal is unlikely to cause interoperability
4125 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
4126 with Z. These are meant to replace LONG and ZLONG and to be size safe.
4127 The use of LONG and ZLONG is discouraged and scheduled for deprecation
4132 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
4133 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
4135 *Richard Levitte, Andy Polyakov*
4137 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
4142 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4143 platform rather than 'mingw'.
4147 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
4148 success if they are asked to add an object which already exists
4149 in the store. This change cascades to other functions which load
4150 certificates and CRLs.
4154 * x86_64 assembly pack: annotate code with DWARF CFI directives to
4155 facilitate stack unwinding even from assembly subroutines.
4159 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
4160 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
4164 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4165 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4166 which is the minimum version we support.
4170 * Certificate time validation (X509_cmp_time) enforces stricter
4171 compliance with RFC 5280. Fractional seconds and timezone offsets
4172 are no longer allowed.
4176 * Add support for ARIA
4180 * s_client will now send the Server Name Indication (SNI) extension by
4181 default unless the new "-noservername" option is used. The server name is
4182 based on the host provided to the "-connect" option unless overridden by
4183 using "-servername".
4187 * Add support for SipHash
4191 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4192 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4193 prevent issues where no progress is being made and the peer continually
4194 sends unrecognised record types, using up resources processing them.
4198 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4199 using the algorithm defined in
4200 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4204 * Heartbeat support has been removed; the ABI is changed for now.
4206 *Richard Levitte, Rich Salz*
4208 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4212 * The RSA "null" method, which was partially supported to avoid patent
4213 issues, has been replaced to always returns NULL.
4220 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4222 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4223 used even when parsing explicit parameters, when loading a encoded key
4224 or calling `EC_GROUP_new_from_ecpkparameters()`/
4225 `EC_GROUP_new_from_ecparameters()`.
4226 This prevents bypass of security hardening and performance gains,
4227 especially for curves with specialized EC_METHODs.
4228 By default, if a key encoded with explicit parameters is loaded and later
4229 encoded, the output is still encoded with explicit parameters, even if
4230 internally a "named" EC_GROUP is used for computation.
4234 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4235 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4236 NULL. After this change, only the cofactor parameter can be NULL. It also
4237 does some minimal sanity checks on the passed order.
4242 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4243 An attack is simple, if the first CMS_recipientInfo is valid but the
4244 second CMS_recipientInfo is chosen ciphertext. If the second
4245 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4246 encryption key will be replaced by garbage, and the message cannot be
4247 decoded, but if the RSA decryption fails, the correct encryption key is
4248 used and the recipient will not notice the attack.
4249 As a work around for this potential attack the length of the decrypted
4250 key must be equal to the cipher default key length, in case the
4251 certificate is not given and all recipientInfo are tried out.
4252 The old behaviour can be re-enabled in the CMS code by setting the
4253 CMS_DEBUG_DECRYPT flag.
4258 * Use Windows installation paths in the mingw builds
4260 Mingw isn't a POSIX environment per se, which means that Windows
4261 paths should be used for installation.
4266 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4268 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4269 This changes the size when using the `genpkey` command when no size is given.
4270 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4271 generation commands to use 2048 bits by default.
4275 * Prevent over long nonces in ChaCha20-Poly1305.
4277 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4278 for every encryption operation. RFC 7539 specifies that the nonce value
4279 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4280 and front pads the nonce with 0 bytes if it is less than 12
4281 bytes. However it also incorrectly allows a nonce to be set of up to 16
4282 bytes. In this case only the last 12 bytes are significant and any
4283 additional leading bytes are ignored.
4285 It is a requirement of using this cipher that nonce values are
4286 unique. Messages encrypted using a reused nonce value are susceptible to
4287 serious confidentiality and integrity attacks. If an application changes
4288 the default nonce length to be longer than 12 bytes and then makes a
4289 change to the leading bytes of the nonce expecting the new value to be a
4290 new unique nonce then such an application could inadvertently encrypt
4291 messages with a reused nonce.
4293 Additionally the ignored bytes in a long nonce are not covered by the
4294 integrity guarantee of this cipher. Any application that relies on the
4295 integrity of these ignored leading bytes of a long nonce may be further
4296 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4297 is safe because no such use sets such a long nonce value. However user
4298 applications that use this cipher directly and set a non-default nonce
4299 length to be longer than 12 bytes may be vulnerable.
4301 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4307 * Added SCA hardening for modular field inversion in EC_GROUP through
4308 a new dedicated field_inv() pointer in EC_METHOD.
4309 This also addresses a leakage affecting conversions from projective
4310 to affine coordinates.
4312 *Billy Bob Brumley, Nicola Tuveri*
4314 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4315 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4319 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4323 * Remove the 'dist' target and add a tarball building script. The
4324 'dist' target has fallen out of use, and it shouldn't be
4325 necessary to configure just to create a source distribution.
4329 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4331 * Timing vulnerability in DSA signature generation
4333 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4334 timing side channel attack. An attacker could use variations in the signing
4335 algorithm to recover the private key.
4337 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4342 * Timing vulnerability in ECDSA signature generation
4344 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4345 timing side channel attack. An attacker could use variations in the signing
4346 algorithm to recover the private key.
4348 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4353 * Add coordinate blinding for EC_POINT and implement projective
4354 coordinate blinding for generic prime curves as a countermeasure to
4355 chosen point SCA attacks.
4357 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4359 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4361 * Client DoS due to large DH parameter
4363 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4364 malicious server can send a very large prime value to the client. This will
4365 cause the client to spend an unreasonably long period of time generating a
4366 key for this prime resulting in a hang until the client has finished. This
4367 could be exploited in a Denial Of Service attack.
4369 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4374 * Cache timing vulnerability in RSA Key Generation
4376 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4377 a cache timing side channel attack. An attacker with sufficient access to
4378 mount cache timing attacks during the RSA key generation process could
4379 recover the private key.
4381 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4382 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4387 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4388 parameter is no longer accepted, as it leads to a corrupt table. NULL
4389 pem_str is reserved for alias entries only.
4393 * Revert blinding in ECDSA sign and instead make problematic addition
4394 length-invariant. Switch even to fixed-length Montgomery multiplication.
4398 * Change generating and checking of primes so that the error rate of not
4399 being prime depends on the intended use based on the size of the input.
4400 For larger primes this will result in more rounds of Miller-Rabin.
4401 The maximal error rate for primes with more than 1080 bits is lowered
4404 *Kurt Roeckx, Annie Yousar*
4406 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4410 * Add blinding to ECDSA and DSA signatures to protect against side channel
4411 attacks discovered by Keegan Ryan (NCC Group).
4415 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4416 now allow empty (zero character) pass phrases.
4420 * Certificate time validation (X509_cmp_time) enforces stricter
4421 compliance with RFC 5280. Fractional seconds and timezone offsets
4422 are no longer allowed.
4426 * Fixed a text canonicalisation bug in CMS
4428 Where a CMS detached signature is used with text content the text goes
4429 through a canonicalisation process first prior to signing or verifying a
4430 signature. This process strips trailing space at the end of lines, converts
4431 line terminators to CRLF and removes additional trailing line terminators
4432 at the end of a file. A bug in the canonicalisation process meant that
4433 some characters, such as form-feed, were incorrectly treated as whitespace
4434 and removed. This is contrary to the specification (RFC5485). This fix
4435 could mean that detached text data signed with an earlier version of
4436 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4437 signed with a fixed OpenSSL may fail to verify with an earlier version of
4438 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4439 and use the "-binary" flag (for the "cms" command line application) or set
4440 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4444 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4446 * Constructed ASN.1 types with a recursive definition could exceed the stack
4448 Constructed ASN.1 types with a recursive definition (such as can be found
4449 in PKCS7) could eventually exceed the stack given malicious input with
4450 excessive recursion. This could result in a Denial Of Service attack. There
4451 are no such structures used within SSL/TLS that come from untrusted sources
4452 so this is considered safe.
4454 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4460 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4462 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4463 effectively reduced to only comparing the least significant bit of each
4464 byte. This allows an attacker to forge messages that would be considered as
4465 authenticated in an amount of tries lower than that guaranteed by the
4466 security claims of the scheme. The module can only be compiled by the
4467 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4469 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4475 * Add a build target 'build_all_generated', to build all generated files
4476 and only that. This can be used to prepare everything that requires
4477 things like perl for a system that lacks perl and then move everything
4478 to that system and do the rest of the build there.
4482 * Backport SSL_OP_NO_RENGOTIATION
4484 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4485 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4486 changes this is no longer possible in 1.1.0. Therefore, the new
4487 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4488 1.1.0 to provide equivalent functionality.
4490 Note that if an application built against 1.1.0h headers (or above) is run
4491 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4492 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4496 * Removed the OS390-Unix config target. It relied on a script that doesn't
4501 * rsaz_1024_mul_avx2 overflow bug on x86_64
4503 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4504 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4505 Analysis suggests that attacks against RSA and DSA as a result of this
4506 defect would be very difficult to perform and are not believed likely.
4507 Attacks against DH1024 are considered just feasible, because most of the
4508 work necessary to deduce information about a private key may be performed
4509 offline. The amount of resources required for such an attack would be
4510 significant. However, for an attack on TLS to be meaningful, the server
4511 would have to share the DH1024 private key among multiple clients, which is
4512 no longer an option since CVE-2016-0701.
4514 This only affects processors that support the AVX2 but not ADX extensions
4515 like Intel Haswell (4th generation).
4517 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4518 was originally found via the OSS-Fuzz project.
4523 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4525 * bn_sqrx8x_internal carry bug on x86_64
4527 There is a carry propagating bug in the x86_64 Montgomery squaring
4528 procedure. No EC algorithms are affected. Analysis suggests that attacks
4529 against RSA and DSA as a result of this defect would be very difficult to
4530 perform and are not believed likely. Attacks against DH are considered just
4531 feasible (although very difficult) because most of the work necessary to
4532 deduce information about a private key may be performed offline. The amount
4533 of resources required for such an attack would be very significant and
4534 likely only accessible to a limited number of attackers. An attacker would
4535 additionally need online access to an unpatched system using the target
4536 private key in a scenario with persistent DH parameters and a private
4537 key that is shared between multiple clients.
4539 This only affects processors that support the BMI1, BMI2 and ADX extensions
4540 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4542 This issue was reported to OpenSSL by the OSS-Fuzz project.
4547 * Malformed X.509 IPAddressFamily could cause OOB read
4549 If an X.509 certificate has a malformed IPAddressFamily extension,
4550 OpenSSL could do a one-byte buffer overread. The most likely result
4551 would be an erroneous display of the certificate in text format.
4553 This issue was reported to OpenSSL by the OSS-Fuzz project.
4558 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4560 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4561 platform rather than 'mingw'.
4565 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4566 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4567 which is the minimum version we support.
4571 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4573 * Encrypt-Then-Mac renegotiation crash
4575 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4576 negotiated where it was not in the original handshake (or vice-versa) then
4577 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4578 and servers are affected.
4580 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4585 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4587 * Truncated packet could crash via OOB read
4589 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4590 cipher is being used, then a truncated packet can cause that host to
4591 perform an out-of-bounds read, usually resulting in a crash.
4593 This issue was reported to OpenSSL by Robert Święcki of Google.
4598 * Bad (EC)DHE parameters cause a client crash
4600 If a malicious server supplies bad parameters for a DHE or ECDHE key
4601 exchange then this can result in the client attempting to dereference a
4602 NULL pointer leading to a client crash. This could be exploited in a Denial
4605 This issue was reported to OpenSSL by Guido Vranken.
4610 * BN_mod_exp may produce incorrect results on x86_64
4612 There is a carry propagating bug in the x86_64 Montgomery squaring
4613 procedure. No EC algorithms are affected. Analysis suggests that attacks
4614 against RSA and DSA as a result of this defect would be very difficult to
4615 perform and are not believed likely. Attacks against DH are considered just
4616 feasible (although very difficult) because most of the work necessary to
4617 deduce information about a private key may be performed offline. The amount
4618 of resources required for such an attack would be very significant and
4619 likely only accessible to a limited number of attackers. An attacker would
4620 additionally need online access to an unpatched system using the target
4621 private key in a scenario with persistent DH parameters and a private
4622 key that is shared between multiple clients. For example this can occur by
4623 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4624 similar to CVE-2015-3193 but must be treated as a separate problem.
4626 This issue was reported to OpenSSL by the OSS-Fuzz project.
4631 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4633 * ChaCha20/Poly1305 heap-buffer-overflow
4635 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4636 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4637 crash. This issue is not considered to be exploitable beyond a DoS.
4639 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4644 * CMS Null dereference
4646 Applications parsing invalid CMS structures can crash with a NULL pointer
4647 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4648 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4649 structure callback if an attempt is made to free certain invalid encodings.
4650 Only CHOICE structures using a callback which do not handle NULL value are
4653 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4658 * Montgomery multiplication may produce incorrect results
4660 There is a carry propagating bug in the Broadwell-specific Montgomery
4661 multiplication procedure that handles input lengths divisible by, but
4662 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4663 and DH private keys are impossible. This is because the subroutine in
4664 question is not used in operations with the private key itself and an input
4665 of the attacker's direct choice. Otherwise the bug can manifest itself as
4666 transient authentication and key negotiation failures or reproducible
4667 erroneous outcome of public-key operations with specially crafted input.
4668 Among EC algorithms only Brainpool P-512 curves are affected and one
4669 presumably can attack ECDH key negotiation. Impact was not analyzed in
4670 detail, because pre-requisites for attack are considered unlikely. Namely
4671 multiple clients have to choose the curve in question and the server has to
4672 share the private key among them, neither of which is default behaviour.
4673 Even then only clients that chose the curve will be affected.
4675 This issue was publicly reported as transient failures and was not
4676 initially recognized as a security issue. Thanks to Richard Morgan for
4677 providing reproducible case.
4682 * Removed automatic addition of RPATH in shared libraries and executables,
4683 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4687 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4689 * Fix Use After Free for large message sizes
4691 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4692 message larger than approx 16k is received then the underlying buffer to
4693 store the incoming message is reallocated and moved. Unfortunately a
4694 dangling pointer to the old location is left which results in an attempt to
4695 write to the previously freed location. This is likely to result in a
4696 crash, however it could potentially lead to execution of arbitrary code.
4698 This issue only affects OpenSSL 1.1.0a.
4700 This issue was reported to OpenSSL by Robert Święcki.
4705 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4707 * OCSP Status Request extension unbounded memory growth
4709 A malicious client can send an excessively large OCSP Status Request
4710 extension. If that client continually requests renegotiation, sending a
4711 large OCSP Status Request extension each time, then there will be unbounded
4712 memory growth on the server. This will eventually lead to a Denial Of
4713 Service attack through memory exhaustion. Servers with a default
4714 configuration are vulnerable even if they do not support OCSP. Builds using
4715 the "no-ocsp" build time option are not affected.
4717 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4722 * SSL_peek() hang on empty record
4724 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4725 sends an empty record. This could be exploited by a malicious peer in a
4726 Denial Of Service attack.
4728 This issue was reported to OpenSSL by Alex Gaynor.
4733 * Excessive allocation of memory in tls_get_message_header() and
4734 dtls1_preprocess_fragment()
4736 A (D)TLS message includes 3 bytes for its length in the header for the
4737 message. This would allow for messages up to 16Mb in length. Messages of
4738 this length are excessive and OpenSSL includes a check to ensure that a
4739 peer is sending reasonably sized messages in order to avoid too much memory
4740 being consumed to service a connection. A flaw in the logic of version
4741 1.1.0 means that memory for the message is allocated too early, prior to
4742 the excessive message length check. Due to way memory is allocated in
4743 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4744 to service a connection. This could lead to a Denial of Service through
4745 memory exhaustion. However, the excessive message length check still takes
4746 place, and this would cause the connection to immediately fail. Assuming
4747 that the application calls SSL_free() on the failed connection in a timely
4748 manner then the 21Mb of allocated memory will then be immediately freed
4749 again. Therefore, the excessive memory allocation will be transitory in
4750 nature. This then means that there is only a security impact if:
4752 1) The application does not call SSL_free() in a timely manner in the event
4753 that the connection fails
4755 2) The application is working in a constrained environment where there is
4756 very little free memory
4758 3) The attacker initiates multiple connection attempts such that there are
4759 multiple connections in a state where memory has been allocated for the
4760 connection; SSL_free() has not yet been called; and there is insufficient
4761 memory to service the multiple requests.
4763 Except in the instance of (1) above any Denial Of Service is likely to be
4764 transitory because as soon as the connection fails the memory is
4765 subsequently freed again in the SSL_free() call. However there is an
4766 increased risk during this period of application crashes due to the lack of
4767 memory - which would then mean a more serious Denial of Service.
4769 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4770 (CVE-2016-6307 and CVE-2016-6308)
4774 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4775 had to be removed. Primary reason is that vendor assembler can't
4776 assemble our modules with -KPIC flag. As result it, assembly
4777 support, was not even available as option. But its lack means
4778 lack of side-channel resistant code, which is incompatible with
4779 security by todays standards. Fortunately gcc is readily available
4780 prepackaged option, which we firmly point at...
4784 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4786 * Windows command-line tool supports UTF-8 opt-in option for arguments
4787 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4788 (to any value) allows Windows user to access PKCS#12 file generated
4789 with Windows CryptoAPI and protected with non-ASCII password, as well
4790 as files generated under UTF-8 locale on Linux also protected with
4795 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4796 have been disabled by default and removed from DEFAULT, just like RC4.
4797 See the RC4 item below to re-enable both.
4801 * The method for finding the storage location for the Windows RAND seed file
4802 has changed. First we check %RANDFILE%. If that is not set then we check
4803 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4804 all else fails we fall back to C:\.
4808 * The EVP_EncryptUpdate() function has had its return type changed from void
4809 to int. A return of 0 indicates and error while a return of 1 indicates
4814 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4815 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4816 off the constant time implementation for RSA, DSA and DH have been made
4817 no-ops and deprecated.
4821 * Windows RAND implementation was simplified to only get entropy by
4822 calling CryptGenRandom(). Various other RAND-related tickets
4825 *Joseph Wylie Yandle, Rich Salz*
4827 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4828 and `OPENSSL_LH_`, respectively. The old names are available
4829 with API compatibility. They new names are now completely documented.
4833 * Unify TYPE_up_ref(obj) methods signature.
4834 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4835 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4836 int (instead of void) like all others TYPE_up_ref() methods.
4837 So now these methods also check the return value of CRYPTO_atomic_add(),
4838 and the validity of object reference counter.
4840 *fdasilvayy@gmail.com*
4842 * With Windows Visual Studio builds, the .pdb files are installed
4843 alongside the installed libraries and executables. For a static
4844 library installation, ossl_static.pdb is the associate compiler
4845 generated .pdb file to be used when linking programs.
4849 * Remove openssl.spec. Packaging files belong with the packagers.
4853 * Automatic Darwin/OSX configuration has had a refresh, it will now
4854 recognise x86_64 architectures automatically. You can still decide
4855 to build for a different bitness with the environment variable
4856 KERNEL_BITS (can be 32 or 64), for example:
4858 KERNEL_BITS=32 ./config
4862 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4863 256 bit AES and HMAC with SHA256.
4867 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4871 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4875 * To enable users to have their own config files and build file templates,
4876 Configure looks in the directory indicated by the environment variable
4877 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4878 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4879 name and is used as is.
4883 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4884 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4885 X509_CERT_FILE_CTX was removed.
4889 * "shared" builds are now the default. To create only static libraries use
4890 the "no-shared" Configure option.
4894 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4895 All of these option have not worked for some while and are fundamental
4900 * Make various cleanup routines no-ops and mark them as deprecated. Most
4901 global cleanup functions are no longer required because they are handled
4902 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4903 Explicitly de-initing can cause problems (e.g. where a library that uses
4904 OpenSSL de-inits, but an application is still using it). The affected
4905 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4906 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4907 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4908 COMP_zlib_cleanup().
4912 * --strict-warnings no longer enables runtime debugging options
4913 such as REF_DEBUG. Instead, debug options are automatically
4914 enabled with '--debug' builds.
4916 *Andy Polyakov, Emilia Käsper*
4918 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4919 have been moved out of the public header files. New functions for managing
4920 these have been added.
4924 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4925 objects have been moved out of the public header files. New
4926 functions for managing these have been added.
4930 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4931 have been moved out of the public header files. New functions for managing
4932 these have been added.
4936 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4937 moved out of the public header files. New functions for managing these
4942 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4946 * Removed the mk1mf build scripts.
4950 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4951 it is always safe to #include a header now.
4955 * Removed the aged BC-32 config and all its supporting scripts
4959 * Removed support for Ultrix, Netware, and OS/2.
4963 * Add support for HKDF.
4965 *Alessandro Ghedini*
4967 * Add support for blake2b and blake2s
4971 * Added support for "pipelining". Ciphers that have the
4972 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4973 encryptions/decryptions simultaneously. There are currently no built-in
4974 ciphers with this property but the expectation is that engines will be able
4975 to offer it to significantly improve throughput. Support has been extended
4976 into libssl so that multiple records for a single connection can be
4977 processed in one go (for >=TLS 1.1).
4981 * Added the AFALG engine. This is an async capable engine which is able to
4982 offload work to the Linux kernel. In this initial version it only supports
4983 AES128-CBC. The kernel must be version 4.1.0 or greater.
4987 * OpenSSL now uses a new threading API. It is no longer necessary to
4988 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4989 are two supported threading models: pthreads and windows threads. It is
4990 also possible to configure OpenSSL at compile time for "no-threads". The
4991 old threading API should no longer be used. The functions have been
4992 replaced with "no-op" compatibility macros.
4994 *Alessandro Ghedini, Matt Caswell*
4996 * Modify behavior of ALPN to invoke callback after SNI/servername
4997 callback, such that updates to the SSL_CTX affect ALPN.
5001 * Add SSL_CIPHER queries for authentication and key-exchange.
5005 * Changes to the DEFAULT cipherlist:
5006 - Prefer (EC)DHE handshakes over plain RSA.
5007 - Prefer AEAD ciphers over legacy ciphers.
5008 - Prefer ECDSA over RSA when both certificates are available.
5009 - Prefer TLSv1.2 ciphers/PRF.
5010 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
5015 * Change the ECC default curve list to be this, in order: x25519,
5016 secp256r1, secp521r1, secp384r1.
5020 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
5021 disabled by default. They can be re-enabled using the
5022 enable-weak-ssl-ciphers option to Configure.
5026 * If the server has ALPN configured, but supports no protocols that the
5027 client advertises, send a fatal "no_application_protocol" alert.
5028 This behaviour is SHALL in RFC 7301, though it isn't universally
5029 implemented by other servers.
5033 * Add X25519 support.
5034 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
5035 for public and private key encoding using the format documented in
5036 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
5037 key generation and key derivation.
5039 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
5044 * Deprecate SRP_VBASE_get_by_user.
5045 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5046 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
5047 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
5048 seed, even if the seed is configured.
5050 Users should use SRP_VBASE_get1_by_user instead. Note that in
5051 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5052 also that even though configuring the SRP seed attempts to hide
5053 invalid usernames by continuing the handshake with fake
5054 credentials, this behaviour is not constant time and no strong
5055 guarantees are made that the handshake is indistinguishable from
5056 that of a valid user.
5060 * Configuration change; it's now possible to build dynamic engines
5061 without having to build shared libraries and vice versa. This
5062 only applies to the engines in `engines/`, those in `crypto/engine/`
5063 will always be built into libcrypto (i.e. "static").
5065 Building dynamic engines is enabled by default; to disable, use
5066 the configuration option "disable-dynamic-engine".
5068 The only requirements for building dynamic engines are the
5069 presence of the DSO module and building with position independent
5070 code, so they will also automatically be disabled if configuring
5071 with "disable-dso" or "disable-pic".
5073 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
5074 are also taken away from openssl/opensslconf.h, as they are
5079 * Configuration change; if there is a known flag to compile
5080 position independent code, it will always be applied on the
5081 libcrypto and libssl object files, and never on the application
5082 object files. This means other libraries that use routines from
5083 libcrypto / libssl can be made into shared libraries regardless
5084 of how OpenSSL was configured.
5086 If this isn't desirable, the configuration options "disable-pic"
5087 or "no-pic" can be used to disable the use of PIC. This will
5088 also disable building shared libraries and dynamic engines.
5092 * Removed JPAKE code. It was experimental and has no wide use.
5096 * The INSTALL_PREFIX Makefile variable has been renamed to
5097 DESTDIR. That makes for less confusion on what this variable
5098 is for. Also, the configuration option --install_prefix is
5103 * Heartbeat for TLS has been removed and is disabled by default
5104 for DTLS; configure with enable-heartbeats. Code that uses the
5105 old #define's might need to be updated.
5107 *Emilia Käsper, Rich Salz*
5109 * Rename REF_CHECK to REF_DEBUG.
5113 * New "unified" build system
5115 The "unified" build system is aimed to be a common system for all
5116 platforms we support. With it comes new support for VMS.
5118 This system builds supports building in a different directory tree
5119 than the source tree. It produces one Makefile (for unix family
5120 or lookalikes), or one descrip.mms (for VMS).
5122 The source of information to make the Makefile / descrip.mms is
5123 small files called 'build.info', holding the necessary
5124 information for each directory with source to compile, and a
5125 template in Configurations, like unix-Makefile.tmpl or
5128 With this change, the library names were also renamed on Windows
5129 and on VMS. They now have names that are closer to the standard
5130 on Unix, and include the major version number, and in certain
5131 cases, the architecture they are built for. See "Notes on shared
5132 libraries" in INSTALL.
5134 We rely heavily on the perl module Text::Template.
5138 * Added support for auto-initialisation and de-initialisation of the library.
5139 OpenSSL no longer requires explicit init or deinit routines to be called,
5140 except in certain circumstances. See the OPENSSL_init_crypto() and
5141 OPENSSL_init_ssl() man pages for further information.
5145 * The arguments to the DTLSv1_listen function have changed. Specifically the
5146 "peer" argument is now expected to be a BIO_ADDR object.
5148 * Rewrite of BIO networking library. The BIO library lacked consistent
5149 support of IPv6, and adding it required some more extensive
5150 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
5151 which hold all types of addresses and chains of address information.
5152 It also introduces a new API, with functions like BIO_socket,
5153 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
5154 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
5155 have been adapted accordingly.
5159 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
5164 * CRIME protection: disable compression by default, even if OpenSSL is
5165 compiled with zlib enabled. Applications can still enable compression
5166 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
5167 using the SSL_CONF library to configure compression.
5171 * The signature of the session callback configured with
5172 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
5173 was explicitly marked as `const unsigned char*` instead of
5178 * Always DPURIFY. Remove the use of uninitialized memory in the
5179 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
5183 * Removed many obsolete configuration items, including
5184 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
5185 MD2_CHAR, MD2_INT, MD2_LONG
5187 IDEA_SHORT, IDEA_LONG
5188 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5190 *Rich Salz, with advice from Andy Polyakov*
5192 * Many BN internals have been moved to an internal header file.
5194 *Rich Salz with help from Andy Polyakov*
5196 * Configuration and writing out the results from it has changed.
5197 Files such as Makefile include/openssl/opensslconf.h and are now
5198 produced through general templates, such as Makefile.in and
5199 crypto/opensslconf.h.in and some help from the perl module
5202 Also, the center of configuration information is no longer
5203 Makefile. Instead, Configure produces a perl module in
5204 configdata.pm which holds most of the config data (in the hash
5205 table %config), the target data that comes from the target
5206 configuration in one of the `Configurations/*.conf` files (in
5211 * To clarify their intended purposes, the Configure options
5212 --prefix and --openssldir change their semantics, and become more
5213 straightforward and less interdependent.
5215 --prefix shall be used exclusively to give the location INSTALLTOP
5216 where programs, scripts, libraries, include files and manuals are
5217 going to be installed. The default is now /usr/local.
5219 --openssldir shall be used exclusively to give the default
5220 location OPENSSLDIR where certificates, private keys, CRLs are
5221 managed. This is also where the default openssl.cnf gets
5223 If the directory given with this option is a relative path, the
5224 values of both the --prefix value and the --openssldir value will
5225 be combined to become OPENSSLDIR.
5226 The default for --openssldir is INSTALLTOP/ssl.
5228 Anyone who uses --openssldir to specify where OpenSSL is to be
5229 installed MUST change to use --prefix instead.
5233 * The GOST engine was out of date and therefore it has been removed. An up
5234 to date GOST engine is now being maintained in an external repository.
5235 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5236 support for GOST ciphersuites (these are only activated if a GOST engine
5241 * EGD is no longer supported by default; use enable-egd when
5244 *Ben Kaduk and Rich Salz*
5246 * The distribution now has Makefile.in files, which are used to
5247 create Makefile's when Configure is run. *Configure must be run
5248 before trying to build now.*
5252 * The return value for SSL_CIPHER_description() for error conditions
5257 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5259 Obtaining and performing DNSSEC validation of TLSA records is
5260 the application's responsibility. The application provides
5261 the TLSA records of its choice to OpenSSL, and these are then
5262 used to authenticate the peer.
5264 The TLSA records need not even come from DNS. They can, for
5265 example, be used to implement local end-entity certificate or
5266 trust-anchor "pinning", where the "pin" data takes the form
5267 of TLSA records, which can augment or replace verification
5268 based on the usual WebPKI public certification authorities.
5272 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5273 continues to support deprecated interfaces in default builds.
5274 However, applications are strongly advised to compile their
5275 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5276 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5277 or the 1.1.0 releases.
5279 In environments in which all applications have been ported to
5280 not use any deprecated interfaces OpenSSL's Configure script
5281 should be used with the --api=1.1.0 option to entirely remove
5282 support for the deprecated features from the library and
5283 unconditionally disable them in the installed headers.
5284 Essentially the same effect can be achieved with the "no-deprecated"
5285 argument to Configure, except that this will always restrict
5286 the build to just the latest API, rather than a fixed API
5289 As applications are ported to future revisions of the API,
5290 they should update their compile-time OPENSSL_API_COMPAT define
5291 accordingly, but in most cases should be able to continue to
5292 compile with later releases.
5294 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5295 0x10000000L and 0x00908000L, respectively. However those
5296 versions did not support the OPENSSL_API_COMPAT feature, and
5297 so applications are not typically tested for explicit support
5298 of just the undeprecated features of either release.
5302 * Add support for setting the minimum and maximum supported protocol.
5303 It can bet set via the SSL_set_min_proto_version() and
5304 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5305 MaxProtocol. It's recommended to use the new APIs to disable
5306 protocols instead of disabling individual protocols using
5307 SSL_set_options() or SSL_CONF's Protocol. This change also
5308 removes support for disabling TLS 1.2 in the OpenSSL TLS
5309 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5313 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5317 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5318 and integrates ECDSA and ECDH functionality into EC. Implementations can
5319 now redirect key generation and no longer need to convert to or from
5322 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5323 include the ec.h header file instead.
5327 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5328 ciphers who are no longer supported and drops support the ephemeral RSA key
5329 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5333 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5334 opaque. For HMAC_CTX, the following constructors and destructors
5337 HMAC_CTX *HMAC_CTX_new(void);
5338 void HMAC_CTX_free(HMAC_CTX *ctx);
5340 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5341 destroy such methods has been added. See EVP_MD_meth_new(3) and
5342 EVP_CIPHER_meth_new(3) for documentation.
5345 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5346 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5347 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5348 an already created structure.
5349 2) For consistency with the majority of our object creators and
5350 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5351 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5352 for deprecated builds.
5356 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5357 cryptographic operations to be performed asynchronously as long as an
5358 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5359 further details. Libssl has also had this capability integrated with the
5360 introduction of the new mode SSL_MODE_ASYNC and associated error
5361 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5362 pages. This work was developed in partnership with Intel Corp.
5366 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5367 always enabled now. If you want to disable the support you should
5368 exclude it using the list of supported ciphers. This also means that the
5369 "-no_ecdhe" option has been removed from s_server.
5373 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5374 SSL_{CTX_}set1_curves() which can set a list.
5378 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5379 curve you want to support using SSL_{CTX_}set1_curves().
5383 * State machine rewrite. The state machine code has been significantly
5384 refactored in order to remove much duplication of code and solve issues
5385 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5386 further details). This change does have some associated API changes.
5387 Notably the SSL_state() function has been removed and replaced by
5388 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5389 SSL_set_state() has been removed altogether. The previous handshake states
5390 defined in ssl.h and ssl3.h have also been removed.
5394 * All instances of the string "ssleay" in the public API were replaced
5395 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5396 Some error codes related to internal RSA_eay API's were renamed.
5400 * The demo files in crypto/threads were moved to demo/threads.
5404 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5407 *Matt Caswell, Rich Salz*
5409 * New ASN.1 embed macro.
5411 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5412 structure is not allocated: it is part of the parent. That is instead of
5420 This reduces memory fragmentation and make it impossible to accidentally
5421 set a mandatory field to NULL.
5423 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5424 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5425 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5430 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5434 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5435 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5436 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5437 DES and RC4 ciphersuites.
5441 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5442 This changes the decoding behaviour for some invalid messages,
5443 though the change is mostly in the more lenient direction, and
5444 legacy behaviour is preserved as much as possible.
5448 * Fix no-stdio build.
5449 *David Woodhouse <David.Woodhouse@intel.com> and also*
5450 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5452 * New testing framework
5453 The testing framework has been largely rewritten and is now using
5454 perl and the perl modules Test::Harness and an extended variant of
5455 Test::More called OpenSSL::Test to do its work. All test scripts in
5456 test/ have been rewritten into test recipes, and all direct calls to
5457 executables in test/Makefile have become individual recipes using the
5458 simplified testing OpenSSL::Test::Simple.
5460 For documentation on our testing modules, do:
5462 perldoc test/testlib/OpenSSL/Test/Simple.pm
5463 perldoc test/testlib/OpenSSL/Test.pm
5467 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5468 are used; the latter aborts on memory leaks (usually checked on exit).
5469 Some undocumented "set malloc, etc., hooks" functions were removed
5470 and others were changed. All are now documented.
5474 * In DSA_generate_parameters_ex, if the provided seed is too short,
5477 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5479 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5480 from RFC4279, RFC4785, RFC5487, RFC5489.
5482 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5483 original RSA_PSK patch.
5487 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5488 era flag was never set throughout the codebase (only read). Also removed
5489 SSL3_FLAGS_POP_BUFFER which was only used if
5490 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5494 * Changed the default name options in the "ca", "crl", "req" and "x509"
5495 to be "oneline" instead of "compat".
5499 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5500 not aware of clients that still exhibit this bug, and the workaround
5501 hasn't been working properly for a while.
5505 * The return type of BIO_number_read() and BIO_number_written() as well as
5506 the corresponding num_read and num_write members in the BIO structure has
5507 changed from unsigned long to uint64_t. On platforms where an unsigned
5508 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5513 * Given the pervasive nature of TLS extensions it is inadvisable to run
5514 OpenSSL without support for them. It also means that maintaining
5515 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5516 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5520 * Removed support for the two export grade static DH ciphersuites
5521 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5522 were newly added (along with a number of other static DH ciphersuites) to
5523 1.0.2. However the two export ones have *never* worked since they were
5524 introduced. It seems strange in any case to be adding new export
5525 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5529 * Version negotiation has been rewritten. In particular SSLv23_method(),
5530 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5531 and turned into macros which simply call the new preferred function names
5532 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5533 should use the new names instead. Also as part of this change the ssl23.h
5534 header file has been removed.
5538 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5539 code and the associated standard is no longer considered fit-for-purpose.
5543 * RT2547 was closed. When generating a private key, try to make the
5544 output file readable only by the owner. This behavior change might
5545 be noticeable when interacting with other software.
5547 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5552 * Added HTTP GET support to the ocsp command.
5556 * Changed default digest for the dgst and enc commands from MD5 to
5561 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5565 * Added support for TLS extended master secret from
5566 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5567 initial patch which was a great help during development.
5571 * All libssl internal structures have been removed from the public header
5572 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5573 now redundant). Users should not attempt to access internal structures
5574 directly. Instead they should use the provided API functions.
5578 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5579 Access to deprecated functions can be re-enabled by running config with
5580 "enable-deprecated". In addition applications wishing to use deprecated
5581 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5582 will, by default, disable some transitive includes that previously existed
5583 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5587 * Added support for OCB mode. OpenSSL has been granted a patent license
5588 compatible with the OpenSSL license for use of OCB. Details are available
5589 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5590 for OCB can be removed by calling config with no-ocb.
5594 * SSLv2 support has been removed. It still supports receiving an SSLv2
5595 compatible client hello.
5599 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5600 done while fixing the error code for the key-too-small case.
5602 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5604 * CA.sh has been removed; use CA.pl instead.
5608 * Removed old DES API.
5612 * Remove various unsupported platforms:
5618 Sinix/ReliantUNIX RM400
5623 16-bit platforms such as WIN16
5627 * Clean up OPENSSL_NO_xxx #define's
5628 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5629 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5630 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5631 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5632 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5633 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5634 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5635 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5636 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5637 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5641 * Cleaned up dead code
5642 Remove all but one '#ifdef undef' which is to be looked at.
5646 * Clean up calling of xxx_free routines.
5647 Just like free(), fix most of the xxx_free routines to accept
5648 NULL. Remove the non-null checks from callers. Save much code.
5652 * Add secure heap for storage of private keys (when possible).
5653 Add BIO_s_secmem(), CBIGNUM, etc.
5654 Contributed by Akamai Technologies under our Corporate CLA.
5658 * Experimental support for a new, fast, unbiased prime candidate generator,
5659 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5661 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5663 * New output format NSS in the sess_id command line tool. This allows
5664 exporting the session id and the master key in NSS keylog format.
5666 *Martin Kaiser <martin@kaiser.cx>*
5668 * Harmonize version and its documentation. -f flag is used to display
5671 *mancha <mancha1@zoho.com>*
5673 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5674 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5676 *mancha <mancha1@zoho.com>*
5678 * Fix some double frees. These are not thought to be exploitable.
5680 *mancha <mancha1@zoho.com>*
5682 * A missing bounds check in the handling of the TLS heartbeat extension
5683 can be used to reveal up to 64k of memory to a connected client or
5686 Thanks for Neel Mehta of Google Security for discovering this bug and to
5687 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5688 preparing the fix ([CVE-2014-0160])
5690 *Adam Langley, Bodo Moeller*
5692 * Fix for the attack described in the paper "Recovering OpenSSL
5693 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5694 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5695 <http://eprint.iacr.org/2014/140>
5697 Thanks to Yuval Yarom and Naomi Benger for discovering this
5698 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5700 *Yuval Yarom and Naomi Benger*
5702 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5703 this fixes a limitation in previous versions of OpenSSL.
5707 * Experimental encrypt-then-mac support.
5709 Experimental support for encrypt then mac from
5710 draft-gutmann-tls-encrypt-then-mac-02.txt
5712 To enable it set the appropriate extension number (0x42 for the test
5713 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5715 For non-compliant peers (i.e. just about everything) this should have no
5718 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5722 * Add EVP support for key wrapping algorithms, to avoid problems with
5723 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5724 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5725 algorithms and include tests cases.
5729 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5734 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5735 MGF1 digest and OAEP label.
5739 * Make openssl verify return errors.
5741 *Chris Palmer <palmer@google.com> and Ben Laurie*
5743 * New function ASN1_TIME_diff to calculate the difference between two
5744 ASN1_TIME structures or one structure and the current time.
5748 * Update fips_test_suite to support multiple command line options. New
5749 test to induce all self test errors in sequence and check expected
5754 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5755 sign or verify all in one operation.
5759 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5760 test programs and fips_test_suite. Includes functionality to parse
5761 the minimal script output of fipsalgest.pl directly.
5765 * Add authorisation parameter to FIPS_module_mode_set().
5769 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5773 * Use separate DRBG fields for internal and external flags. New function
5774 FIPS_drbg_health_check() to perform on demand health checking. Add
5775 generation tests to fips_test_suite with reduced health check interval to
5776 demonstrate periodic health checking. Add "nodh" option to
5777 fips_test_suite to skip very slow DH test.
5781 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5786 * More extensive health check for DRBG checking many more failure modes.
5787 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5788 combination: call this in fips_test_suite.
5792 * Add support for canonical generation of DSA parameter 'g'. See
5795 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5796 POST to handle HMAC cases.
5800 * Add functions FIPS_module_version() and FIPS_module_version_text()
5801 to return numerical and string versions of the FIPS module number.
5805 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5806 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5807 outside the validated module in the FIPS capable OpenSSL.
5811 * Minor change to DRBG entropy callback semantics. In some cases
5812 there is no multiple of the block length between min_len and
5813 max_len. Allow the callback to return more than max_len bytes
5814 of entropy but discard any extra: it is the callback's responsibility
5815 to ensure that the extra data discarded does not impact the
5816 requested amount of entropy.
5820 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5821 information in FIPS186-3, SP800-57 and SP800-131A.
5825 * CCM support via EVP. Interface is very similar to GCM case except we
5826 must supply all data in one chunk (i.e. no update, final) and the
5827 message length must be supplied if AAD is used. Add algorithm test
5832 * Initial version of POST overhaul. Add POST callback to allow the status
5833 of POST to be monitored and/or failures induced. Modify fips_test_suite
5834 to use callback. Always run all selftests even if one fails.
5838 * XTS support including algorithm test driver in the fips_gcmtest program.
5839 Note: this does increase the maximum key length from 32 to 64 bytes but
5840 there should be no binary compatibility issues as existing applications
5841 will never use XTS mode.
5845 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5846 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5847 performs algorithm blocking for unapproved PRNG types. Also do not
5848 set PRNG type in FIPS_mode_set(): leave this to the application.
5849 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5850 the standard OpenSSL PRNG: set additional data to a date time vector.
5854 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5855 This shouldn't present any incompatibility problems because applications
5856 shouldn't be using these directly and any that are will need to rethink
5857 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5861 * Extensive self tests and health checking required by SP800-90 DRBG.
5862 Remove strength parameter from FIPS_drbg_instantiate and always
5863 instantiate at maximum supported strength.
5867 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5871 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5875 * New function DH_compute_key_padded() to compute a DH key and pad with
5876 leading zeroes if needed: this complies with SP800-56A et al.
5880 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5881 anything, incomplete, subject to change and largely untested at present.
5885 * Modify fipscanisteronly build option to only build the necessary object
5886 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5890 * Add experimental option FIPSSYMS to give all symbols in
5891 fipscanister.o and FIPS or fips prefix. This will avoid
5892 conflicts with future versions of OpenSSL. Add perl script
5893 util/fipsas.pl to preprocess assembly language source files
5894 and rename any affected symbols.
5898 * Add selftest checks and algorithm block of non-fips algorithms in
5899 FIPS mode. Remove DES2 from selftests.
5903 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5904 return internal method without any ENGINE dependencies. Add new
5905 tiny fips sign and verify functions.
5909 * New build option no-ec2m to disable characteristic 2 code.
5913 * New build option "fipscanisteronly". This only builds fipscanister.o
5914 and (currently) associated fips utilities. Uses the file Makefile.fips
5915 instead of Makefile.org as the prototype.
5919 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5920 Update fips_gcmtest to use IV generator.
5924 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5925 setting output buffer to NULL. The `*Final` function must be
5926 called although it will not retrieve any additional data. The tag
5927 can be set or retrieved with a ctrl. The IV length is by default 12
5928 bytes (96 bits) but can be set to an alternative value. If the IV
5929 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5934 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5935 underlying do_cipher function handles all cipher semantics itself
5936 including padding and finalisation. This is useful if (for example)
5937 an ENGINE cipher handles block padding itself. The behaviour of
5938 do_cipher is subtly changed if this flag is set: the return value
5939 is the number of characters written to the output buffer (zero is
5940 no longer an error code) or a negative error code. Also if the
5941 input buffer is NULL and length 0 finalisation should be performed.
5945 * If a candidate issuer certificate is already part of the constructed
5946 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5950 * Improve forward-security support: add functions
5952 void SSL_CTX_set_not_resumable_session_callback(
5953 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5954 void SSL_set_not_resumable_session_callback(
5955 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5957 for use by SSL/TLS servers; the callback function will be called whenever a
5958 new session is created, and gets to decide whether the session may be
5959 cached to make it resumable (return 0) or not (return 1). (As by the
5960 SSL/TLS protocol specifications, the session_id sent by the server will be
5961 empty to indicate that the session is not resumable; also, the server will
5962 not generate RFC 4507 (RFC 5077) session tickets.)
5964 A simple reasonable callback implementation is to return is_forward_secure.
5965 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5966 by the SSL/TLS server library, indicating whether it can provide forward
5969 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5971 * New -verify_name option in command line utilities to set verification
5976 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5977 Add CMAC pkey methods.
5981 * Experimental renegotiation in s_server -www mode. If the client
5982 browses /reneg connection is renegotiated. If /renegcert it is
5983 renegotiated requesting a certificate.
5987 * Add an "external" session cache for debugging purposes to s_server. This
5988 should help trace issues which normally are only apparent in deployed
5989 multi-process servers.
5993 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5994 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5995 BIO_set_cipher() and some obscure PEM functions were changed so they
5996 can now return an error. The RAND changes required a change to the
5997 RAND_METHOD structure.
6001 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
6002 a gcc attribute to warn if the result of a function is ignored. This
6003 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
6004 whose return value is often ignored.
6008 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
6009 These allow SCTs (signed certificate timestamps) to be requested and
6010 validated when establishing a connection.
6012 *Rob Percival <robpercival@google.com>*
6017 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
6019 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
6020 used even when parsing explicit parameters, when loading a encoded key
6021 or calling `EC_GROUP_new_from_ecpkparameters()`/
6022 `EC_GROUP_new_from_ecparameters()`.
6023 This prevents bypass of security hardening and performance gains,
6024 especially for curves with specialized EC_METHODs.
6025 By default, if a key encoded with explicit parameters is loaded and later
6026 encoded, the output is still encoded with explicit parameters, even if
6027 internally a "named" EC_GROUP is used for computation.
6031 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
6032 this change, EC_GROUP_set_generator would accept order and/or cofactor as
6033 NULL. After this change, only the cofactor parameter can be NULL. It also
6034 does some minimal sanity checks on the passed order.
6039 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
6040 An attack is simple, if the first CMS_recipientInfo is valid but the
6041 second CMS_recipientInfo is chosen ciphertext. If the second
6042 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
6043 encryption key will be replaced by garbage, and the message cannot be
6044 decoded, but if the RSA decryption fails, the correct encryption key is
6045 used and the recipient will not notice the attack.
6046 As a work around for this potential attack the length of the decrypted
6047 key must be equal to the cipher default key length, in case the
6048 certificate is not given and all recipientInfo are tried out.
6049 The old behaviour can be re-enabled in the CMS code by setting the
6050 CMS_DEBUG_DECRYPT flag.
6055 * Document issue with installation paths in diverse Windows builds
6057 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
6058 binaries and run-time config file.
6063 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
6065 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
6066 This changes the size when using the `genpkey` command when no size is given.
6067 It fixes an omission in earlier changes that changed all RSA, DSA and DH
6068 generation commands to use 2048 bits by default.
6072 * Add FIPS support for Android Arm 64-bit
6074 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
6075 Module in Version 2.0.10. For some reason, the corresponding target
6076 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
6077 built with FIPS support on Android Arm 64-bit. This omission has been
6080 *Matthias St. Pierre*
6082 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
6084 * 0-byte record padding oracle
6086 If an application encounters a fatal protocol error and then calls
6087 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
6088 then OpenSSL can respond differently to the calling application if a 0 byte
6089 record is received with invalid padding compared to if a 0 byte record is
6090 received with an invalid MAC. If the application then behaves differently
6091 based on that in a way that is detectable to the remote peer, then this
6092 amounts to a padding oracle that could be used to decrypt data.
6094 In order for this to be exploitable "non-stitched" ciphersuites must be in
6095 use. Stitched ciphersuites are optimised implementations of certain
6096 commonly used ciphersuites. Also the application must call SSL_shutdown()
6097 twice even if a protocol error has occurred (applications should not do
6098 this but some do anyway).
6100 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
6101 Aviram, with additional investigation by Steven Collison and Andrew
6102 Hourselt. It was reported to OpenSSL on 10th December 2018.
6107 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
6111 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
6113 * Microarchitecture timing vulnerability in ECC scalar multiplication
6115 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
6116 shown to be vulnerable to a microarchitecture timing side channel attack.
6117 An attacker with sufficient access to mount local timing attacks during
6118 ECDSA signature generation could recover the private key.
6120 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
6121 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
6127 * Timing vulnerability in DSA signature generation
6129 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
6130 timing side channel attack. An attacker could use variations in the signing
6131 algorithm to recover the private key.
6133 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
6138 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
6139 Module, accidentally introduced while backporting security fixes from the
6140 development branch and hindering the use of ECC in FIPS mode.
6144 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
6146 * Client DoS due to large DH parameter
6148 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
6149 malicious server can send a very large prime value to the client. This will
6150 cause the client to spend an unreasonably long period of time generating a
6151 key for this prime resulting in a hang until the client has finished. This
6152 could be exploited in a Denial Of Service attack.
6154 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
6159 * Cache timing vulnerability in RSA Key Generation
6161 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
6162 a cache timing side channel attack. An attacker with sufficient access to
6163 mount cache timing attacks during the RSA key generation process could
6164 recover the private key.
6166 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
6167 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
6172 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
6173 parameter is no longer accepted, as it leads to a corrupt table. NULL
6174 pem_str is reserved for alias entries only.
6178 * Revert blinding in ECDSA sign and instead make problematic addition
6179 length-invariant. Switch even to fixed-length Montgomery multiplication.
6183 * Change generating and checking of primes so that the error rate of not
6184 being prime depends on the intended use based on the size of the input.
6185 For larger primes this will result in more rounds of Miller-Rabin.
6186 The maximal error rate for primes with more than 1080 bits is lowered
6189 *Kurt Roeckx, Annie Yousar*
6191 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6195 * Add blinding to ECDSA and DSA signatures to protect against side channel
6196 attacks discovered by Keegan Ryan (NCC Group).
6200 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6201 now allow empty (zero character) pass phrases.
6205 * Certificate time validation (X509_cmp_time) enforces stricter
6206 compliance with RFC 5280. Fractional seconds and timezone offsets
6207 are no longer allowed.
6211 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6213 * Constructed ASN.1 types with a recursive definition could exceed the stack
6215 Constructed ASN.1 types with a recursive definition (such as can be found
6216 in PKCS7) could eventually exceed the stack given malicious input with
6217 excessive recursion. This could result in a Denial Of Service attack. There
6218 are no such structures used within SSL/TLS that come from untrusted sources
6219 so this is considered safe.
6221 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6227 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6229 * Read/write after SSL object in error state
6231 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6232 mechanism. The intent was that if a fatal error occurred during a handshake
6233 then OpenSSL would move into the error state and would immediately fail if
6234 you attempted to continue the handshake. This works as designed for the
6235 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6236 SSL_connect()), however due to a bug it does not work correctly if
6237 SSL_read() or SSL_write() is called directly. In that scenario, if the
6238 handshake fails then a fatal error will be returned in the initial function
6239 call. If SSL_read()/SSL_write() is subsequently called by the application
6240 for the same SSL object then it will succeed and the data is passed without
6241 being decrypted/encrypted directly from the SSL/TLS record layer.
6243 In order to exploit this issue an application bug would have to be present
6244 that resulted in a call to SSL_read()/SSL_write() being issued after having
6245 already received a fatal error.
6247 This issue was reported to OpenSSL by David Benjamin (Google).
6252 * rsaz_1024_mul_avx2 overflow bug on x86_64
6254 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6255 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6256 Analysis suggests that attacks against RSA and DSA as a result of this
6257 defect would be very difficult to perform and are not believed likely.
6258 Attacks against DH1024 are considered just feasible, because most of the
6259 work necessary to deduce information about a private key may be performed
6260 offline. The amount of resources required for such an attack would be
6261 significant. However, for an attack on TLS to be meaningful, the server
6262 would have to share the DH1024 private key among multiple clients, which is
6263 no longer an option since CVE-2016-0701.
6265 This only affects processors that support the AVX2 but not ADX extensions
6266 like Intel Haswell (4th generation).
6268 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6269 was originally found via the OSS-Fuzz project.
6274 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6276 * bn_sqrx8x_internal carry bug on x86_64
6278 There is a carry propagating bug in the x86_64 Montgomery squaring
6279 procedure. No EC algorithms are affected. Analysis suggests that attacks
6280 against RSA and DSA as a result of this defect would be very difficult to
6281 perform and are not believed likely. Attacks against DH are considered just
6282 feasible (although very difficult) because most of the work necessary to
6283 deduce information about a private key may be performed offline. The amount
6284 of resources required for such an attack would be very significant and
6285 likely only accessible to a limited number of attackers. An attacker would
6286 additionally need online access to an unpatched system using the target
6287 private key in a scenario with persistent DH parameters and a private
6288 key that is shared between multiple clients.
6290 This only affects processors that support the BMI1, BMI2 and ADX extensions
6291 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6293 This issue was reported to OpenSSL by the OSS-Fuzz project.
6298 * Malformed X.509 IPAddressFamily could cause OOB read
6300 If an X.509 certificate has a malformed IPAddressFamily extension,
6301 OpenSSL could do a one-byte buffer overread. The most likely result
6302 would be an erroneous display of the certificate in text format.
6304 This issue was reported to OpenSSL by the OSS-Fuzz project.
6308 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6310 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6311 platform rather than 'mingw'.
6315 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6317 * Truncated packet could crash via OOB read
6319 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6320 cipher is being used, then a truncated packet can cause that host to
6321 perform an out-of-bounds read, usually resulting in a crash.
6323 This issue was reported to OpenSSL by Robert Święcki of Google.
6328 * BN_mod_exp may produce incorrect results on x86_64
6330 There is a carry propagating bug in the x86_64 Montgomery squaring
6331 procedure. No EC algorithms are affected. Analysis suggests that attacks
6332 against RSA and DSA as a result of this defect would be very difficult to
6333 perform and are not believed likely. Attacks against DH are considered just
6334 feasible (although very difficult) because most of the work necessary to
6335 deduce information about a private key may be performed offline. The amount
6336 of resources required for such an attack would be very significant and
6337 likely only accessible to a limited number of attackers. An attacker would
6338 additionally need online access to an unpatched system using the target
6339 private key in a scenario with persistent DH parameters and a private
6340 key that is shared between multiple clients. For example this can occur by
6341 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6342 similar to CVE-2015-3193 but must be treated as a separate problem.
6344 This issue was reported to OpenSSL by the OSS-Fuzz project.
6349 * Montgomery multiplication may produce incorrect results
6351 There is a carry propagating bug in the Broadwell-specific Montgomery
6352 multiplication procedure that handles input lengths divisible by, but
6353 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6354 and DH private keys are impossible. This is because the subroutine in
6355 question is not used in operations with the private key itself and an input
6356 of the attacker's direct choice. Otherwise the bug can manifest itself as
6357 transient authentication and key negotiation failures or reproducible
6358 erroneous outcome of public-key operations with specially crafted input.
6359 Among EC algorithms only Brainpool P-512 curves are affected and one
6360 presumably can attack ECDH key negotiation. Impact was not analyzed in
6361 detail, because pre-requisites for attack are considered unlikely. Namely
6362 multiple clients have to choose the curve in question and the server has to
6363 share the private key among them, neither of which is default behaviour.
6364 Even then only clients that chose the curve will be affected.
6366 This issue was publicly reported as transient failures and was not
6367 initially recognized as a security issue. Thanks to Richard Morgan for
6368 providing reproducible case.
6373 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6374 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6375 prevent issues where no progress is being made and the peer continually
6376 sends unrecognised record types, using up resources processing them.
6380 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6382 * Missing CRL sanity check
6384 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6385 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6386 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6388 This issue only affects the OpenSSL 1.0.2i
6393 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6395 * OCSP Status Request extension unbounded memory growth
6397 A malicious client can send an excessively large OCSP Status Request
6398 extension. If that client continually requests renegotiation, sending a
6399 large OCSP Status Request extension each time, then there will be unbounded
6400 memory growth on the server. This will eventually lead to a Denial Of
6401 Service attack through memory exhaustion. Servers with a default
6402 configuration are vulnerable even if they do not support OCSP. Builds using
6403 the "no-ocsp" build time option are not affected.
6405 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6410 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6413 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6419 * OOB write in MDC2_Update()
6421 An overflow can occur in MDC2_Update() either if called directly or
6422 through the EVP_DigestUpdate() function using MDC2. If an attacker
6423 is able to supply very large amounts of input data after a previous
6424 call to EVP_EncryptUpdate() with a partial block then a length check
6425 can overflow resulting in a heap corruption.
6427 The amount of data needed is comparable to SIZE_MAX which is impractical
6430 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6435 * Malformed SHA512 ticket DoS
6437 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6438 DoS attack where a malformed ticket will result in an OOB read which will
6441 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6442 a custom server callback and ticket lookup mechanism.
6444 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6449 * OOB write in BN_bn2dec()
6451 The function BN_bn2dec() does not check the return value of BN_div_word().
6452 This can cause an OOB write if an application uses this function with an
6453 overly large BIGNUM. This could be a problem if an overly large certificate
6454 or CRL is printed out from an untrusted source. TLS is not affected because
6455 record limits will reject an oversized certificate before it is parsed.
6457 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6462 * OOB read in TS_OBJ_print_bio()
6464 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6465 the total length the OID text representation would use and not the amount
6466 of data written. This will result in OOB reads when large OIDs are
6469 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6474 * Pointer arithmetic undefined behaviour
6476 Avoid some undefined pointer arithmetic
6478 A common idiom in the codebase is to check limits in the following manner:
6481 Where "p" points to some malloc'd data of SIZE bytes and
6484 "len" here could be from some externally supplied data (e.g. from a TLS
6487 The rules of C pointer arithmetic are such that "p + len" is only well
6488 defined where len <= SIZE. Therefore the above idiom is actually
6489 undefined behaviour.
6491 For example this could cause problems if some malloc implementation
6492 provides an address for "p" such that "p + len" actually overflows for
6493 values of len that are too big and therefore p + len < limit.
6495 This issue was reported to OpenSSL by Guido Vranken
6500 * Constant time flag not preserved in DSA signing
6502 Operations in the DSA signing algorithm should run in constant time in
6503 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6504 implementation means that a non-constant time codepath is followed for
6505 certain operations. This has been demonstrated through a cache-timing
6506 attack to be sufficient for an attacker to recover the private DSA key.
6508 This issue was reported by César Pereida (Aalto University), Billy Brumley
6509 (Tampere University of Technology), and Yuval Yarom (The University of
6510 Adelaide and NICTA).
6515 * DTLS buffered message DoS
6517 In a DTLS connection where handshake messages are delivered out-of-order
6518 those messages that OpenSSL is not yet ready to process will be buffered
6519 for later use. Under certain circumstances, a flaw in the logic means that
6520 those messages do not get removed from the buffer even though the handshake
6521 has been completed. An attacker could force up to approx. 15 messages to
6522 remain in the buffer when they are no longer required. These messages will
6523 be cleared when the DTLS connection is closed. The default maximum size for
6524 a message is 100k. Therefore, the attacker could force an additional 1500k
6525 to be consumed per connection. By opening many simultaneous connections an
6526 attacker could cause a DoS attack through memory exhaustion.
6528 This issue was reported to OpenSSL by Quan Luo.
6533 * DTLS replay protection DoS
6535 A flaw in the DTLS replay attack protection mechanism means that records
6536 that arrive for future epochs update the replay protection "window" before
6537 the MAC for the record has been validated. This could be exploited by an
6538 attacker by sending a record for the next epoch (which does not have to
6539 decrypt or have a valid MAC), with a very large sequence number. This means
6540 that all subsequent legitimate packets are dropped causing a denial of
6541 service for a specific DTLS connection.
6543 This issue was reported to OpenSSL by the OCAP audit team.
6548 * Certificate message OOB reads
6550 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6551 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6552 theoretical DoS risk but this has not been observed in practice on common
6555 The messages affected are client certificate, client certificate request
6556 and server certificate. As a result the attack can only be performed
6557 against a client or a server which enables client authentication.
6559 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6564 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6566 * Prevent padding oracle in AES-NI CBC MAC check
6568 A MITM attacker can use a padding oracle attack to decrypt traffic
6569 when the connection uses an AES CBC cipher and the server support
6572 This issue was introduced as part of the fix for Lucky 13 padding
6573 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6574 constant time by making sure that always the same bytes are read and
6575 compared against either the MAC or padding bytes. But it no longer
6576 checked that there was enough data to have both the MAC and padding
6579 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6583 * Fix EVP_EncodeUpdate overflow
6585 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6586 Base64 encoding of binary data. If an attacker is able to supply very large
6587 amounts of input data then a length check can overflow resulting in a heap
6590 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6591 the `PEM_write_bio*` family of functions. These are mainly used within the
6592 OpenSSL command line applications, so any application which processes data
6593 from an untrusted source and outputs it as a PEM file should be considered
6594 vulnerable to this issue. User applications that call these APIs directly
6595 with large amounts of untrusted data may also be vulnerable.
6597 This issue was reported by Guido Vranken.
6602 * Fix EVP_EncryptUpdate overflow
6604 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6605 is able to supply very large amounts of input data after a previous call to
6606 EVP_EncryptUpdate() with a partial block then a length check can overflow
6607 resulting in a heap corruption. Following an analysis of all OpenSSL
6608 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6609 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6610 the first called function after an EVP_EncryptInit(), and therefore that
6611 specific call must be safe. The second form is where the length passed to
6612 EVP_EncryptUpdate() can be seen from the code to be some small value and
6613 therefore there is no possibility of an overflow. Since all instances are
6614 one of these two forms, it is believed that there can be no overflows in
6615 internal code due to this problem. It should be noted that
6616 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6617 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6618 of these calls have also been analysed too and it is believed there are no
6619 instances in internal usage where an overflow could occur.
6621 This issue was reported by Guido Vranken.
6626 * Prevent ASN.1 BIO excessive memory allocation
6628 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6629 a short invalid encoding can cause allocation of large amounts of memory
6630 potentially consuming excessive resources or exhausting memory.
6632 Any application parsing untrusted data through d2i BIO functions is
6633 affected. The memory based functions such as d2i_X509() are *not* affected.
6634 Since the memory based functions are used by the TLS library, TLS
6635 applications are not affected.
6637 This issue was reported by Brian Carpenter.
6644 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6645 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6646 in arbitrary stack data being returned in the buffer.
6648 This issue was reported by Guido Vranken.
6653 * Modify behavior of ALPN to invoke callback after SNI/servername
6654 callback, such that updates to the SSL_CTX affect ALPN.
6658 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6663 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6664 methods are enabled and ssl2 is disabled the methods return NULL.
6668 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6670 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6671 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6672 provide any "EXPORT" or "LOW" strength ciphers.
6676 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6677 is by default disabled at build-time. Builds that are not configured with
6678 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6679 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6680 will need to explicitly call either of:
6682 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6684 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6686 as appropriate. Even if either of those is used, or the application
6687 explicitly uses the version-specific SSLv2_method() or its client and
6688 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6689 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6690 ciphers, and SSLv2 56-bit DES are no longer available.
6695 * Fix a double-free in DSA code
6697 A double free bug was discovered when OpenSSL parses malformed DSA private
6698 keys and could lead to a DoS attack or memory corruption for applications
6699 that receive DSA private keys from untrusted sources. This scenario is
6702 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6708 * Disable SRP fake user seed to address a server memory leak.
6710 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6712 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6713 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6714 was changed to ignore the "fake user" SRP seed, even if the seed
6717 Users should use SRP_VBASE_get1_by_user instead. Note that in
6718 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6719 also that even though configuring the SRP seed attempts to hide
6720 invalid usernames by continuing the handshake with fake
6721 credentials, this behaviour is not constant time and no strong
6722 guarantees are made that the handshake is indistinguishable from
6723 that of a valid user.
6728 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6730 In the BN_hex2bn function the number of hex digits is calculated using an
6731 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6732 large values of `i` this can result in `bn_expand` not allocating any
6733 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6734 field as NULL leading to a subsequent NULL ptr deref. For very large values
6735 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6736 In this case memory is allocated to the internal BIGNUM data field, but it
6737 is insufficiently sized leading to heap corruption. A similar issue exists
6738 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6739 is ever called by user applications with very large untrusted hex/dec data.
6740 This is anticipated to be a rare occurrence.
6742 All OpenSSL internal usage of these functions use data that is not expected
6743 to be untrusted, e.g. config file data or application command line
6744 arguments. If user developed applications generate config file data based
6745 on untrusted data then it is possible that this could also lead to security
6746 consequences. This is also anticipated to be rare.
6748 This issue was reported to OpenSSL by Guido Vranken.
6753 * Fix memory issues in `BIO_*printf` functions
6755 The internal `fmtstr` function used in processing a "%s" format string in
6756 the `BIO_*printf` functions could overflow while calculating the length of a
6757 string and cause an OOB read when printing very long strings.
6759 Additionally the internal `doapr_outch` function can attempt to write to an
6760 OOB memory location (at an offset from the NULL pointer) in the event of a
6761 memory allocation failure. In 1.0.2 and below this could be caused where
6762 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6763 could be in processing a very long "%s" format string. Memory leaks can
6766 The first issue may mask the second issue dependent on compiler behaviour.
6767 These problems could enable attacks where large amounts of untrusted data
6768 is passed to the `BIO_*printf` functions. If applications use these functions
6769 in this way then they could be vulnerable. OpenSSL itself uses these
6770 functions when printing out human-readable dumps of ASN.1 data. Therefore
6771 applications that print this data could be vulnerable if the data is from
6772 untrusted sources. OpenSSL command line applications could also be
6773 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6774 as command line arguments.
6776 Libssl is not considered directly vulnerable. Additionally certificates etc
6777 received via remote connections via libssl are also unlikely to be able to
6778 trigger these issues because of message size limits enforced within libssl.
6780 This issue was reported to OpenSSL Guido Vranken.
6785 * Side channel attack on modular exponentiation
6787 A side-channel attack was found which makes use of cache-bank conflicts on
6788 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6789 of RSA keys. The ability to exploit this issue is limited as it relies on
6790 an attacker who has control of code in a thread running on the same
6791 hyper-threaded core as the victim thread which is performing decryptions.
6793 This issue was reported to OpenSSL by Yuval Yarom, The University of
6794 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6795 Nadia Heninger, University of Pennsylvania with more information at
6796 <http://cachebleed.info>.
6801 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6802 if no keysize is specified with default_bits. This fixes an
6803 omission in an earlier change that changed all RSA/DSA key generation
6804 commands to use 2048 bits by default.
6808 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6810 * DH small subgroups
6812 Historically OpenSSL only ever generated DH parameters based on "safe"
6813 primes. More recently (in version 1.0.2) support was provided for
6814 generating X9.42 style parameter files such as those required for RFC 5114
6815 support. The primes used in such files may not be "safe". Where an
6816 application is using DH configured with parameters based on primes that are
6817 not "safe" then an attacker could use this fact to find a peer's private
6818 DH exponent. This attack requires that the attacker complete multiple
6819 handshakes in which the peer uses the same private DH exponent. For example
6820 this could be used to discover a TLS server's private DH exponent if it's
6821 reusing the private DH exponent or it's using a static DH ciphersuite.
6823 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6824 TLS. It is not on by default. If the option is not set then the server
6825 reuses the same private DH exponent for the life of the server process and
6826 would be vulnerable to this attack. It is believed that many popular
6827 applications do set this option and would therefore not be at risk.
6829 The fix for this issue adds an additional check where a "q" parameter is
6830 available (as is the case in X9.42 based parameters). This detects the
6831 only known attack, and is the only possible defense for static DH
6832 ciphersuites. This could have some performance impact.
6834 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6835 default and cannot be disabled. This could have some performance impact.
6837 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6842 * SSLv2 doesn't block disabled ciphers
6844 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6845 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6846 been disabled, provided that the SSLv2 protocol was not also disabled via
6849 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6850 and Sebastian Schinzel.
6855 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6857 * BN_mod_exp may produce incorrect results on x86_64
6859 There is a carry propagating bug in the x86_64 Montgomery squaring
6860 procedure. No EC algorithms are affected. Analysis suggests that attacks
6861 against RSA and DSA as a result of this defect would be very difficult to
6862 perform and are not believed likely. Attacks against DH are considered just
6863 feasible (although very difficult) because most of the work necessary to
6864 deduce information about a private key may be performed offline. The amount
6865 of resources required for such an attack would be very significant and
6866 likely only accessible to a limited number of attackers. An attacker would
6867 additionally need online access to an unpatched system using the target
6868 private key in a scenario with persistent DH parameters and a private
6869 key that is shared between multiple clients. For example this can occur by
6870 default in OpenSSL DHE based SSL/TLS ciphersuites.
6872 This issue was reported to OpenSSL by Hanno Böck.
6877 * Certificate verify crash with missing PSS parameter
6879 The signature verification routines will crash with a NULL pointer
6880 dereference if presented with an ASN.1 signature using the RSA PSS
6881 algorithm and absent mask generation function parameter. Since these
6882 routines are used to verify certificate signature algorithms this can be
6883 used to crash any certificate verification operation and exploited in a
6884 DoS attack. Any application which performs certificate verification is
6885 vulnerable including OpenSSL clients and servers which enable client
6888 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6893 * X509_ATTRIBUTE memory leak
6895 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6896 memory. This structure is used by the PKCS#7 and CMS routines so any
6897 application which reads PKCS#7 or CMS data from untrusted sources is
6898 affected. SSL/TLS is not affected.
6900 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6906 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6907 This changes the decoding behaviour for some invalid messages,
6908 though the change is mostly in the more lenient direction, and
6909 legacy behaviour is preserved as much as possible.
6913 * In DSA_generate_parameters_ex, if the provided seed is too short,
6916 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6918 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6920 * Alternate chains certificate forgery
6922 During certificate verification, OpenSSL will attempt to find an
6923 alternative certificate chain if the first attempt to build such a chain
6924 fails. An error in the implementation of this logic can mean that an
6925 attacker could cause certain checks on untrusted certificates to be
6926 bypassed, such as the CA flag, enabling them to use a valid leaf
6927 certificate to act as a CA and "issue" an invalid certificate.
6929 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6934 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6936 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6937 incompatibility in the handling of HMAC. The previous ABI has now been
6942 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6944 * Malformed ECParameters causes infinite loop
6946 When processing an ECParameters structure OpenSSL enters an infinite loop
6947 if the curve specified is over a specially malformed binary polynomial
6950 This can be used to perform denial of service against any
6951 system which processes public keys, certificate requests or
6952 certificates. This includes TLS clients and TLS servers with
6953 client authentication enabled.
6955 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6960 * Exploitable out-of-bounds read in X509_cmp_time
6962 X509_cmp_time does not properly check the length of the ASN1_TIME
6963 string and can read a few bytes out of bounds. In addition,
6964 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6967 An attacker can use this to craft malformed certificates and CRLs of
6968 various sizes and potentially cause a segmentation fault, resulting in
6969 a DoS on applications that verify certificates or CRLs. TLS clients
6970 that verify CRLs are affected. TLS clients and servers with client
6971 authentication enabled may be affected if they use custom verification
6974 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6975 independently by Hanno Böck.
6980 * PKCS7 crash with missing EnvelopedContent
6982 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6983 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6984 with missing content and trigger a NULL pointer dereference on parsing.
6986 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6987 structures from untrusted sources are affected. OpenSSL clients and
6988 servers are not affected.
6990 This issue was reported to OpenSSL by Michal Zalewski (Google).
6995 * CMS verify infinite loop with unknown hash function
6997 When verifying a signedData message the CMS code can enter an infinite loop
6998 if presented with an unknown hash function OID. This can be used to perform
6999 denial of service against any system which verifies signedData messages using
7001 This issue was reported to OpenSSL by Johannes Bauer.
7006 * Race condition handling NewSessionTicket
7008 If a NewSessionTicket is received by a multi-threaded client when attempting to
7009 reuse a previous ticket then a race condition can occur potentially leading to
7010 a double free of the ticket data.
7015 * Only support 256-bit or stronger elliptic curves with the
7016 'ecdh_auto' setting (server) or by default (client). Of supported
7017 curves, prefer P-256 (both).
7021 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
7023 * ClientHello sigalgs DoS fix
7025 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
7026 invalid signature algorithms extension a NULL pointer dereference will
7027 occur. This can be exploited in a DoS attack against the server.
7029 This issue was was reported to OpenSSL by David Ramos of Stanford
7033 *Stephen Henson and Matt Caswell*
7035 * Multiblock corrupted pointer fix
7037 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
7038 feature only applies on 64 bit x86 architecture platforms that support AES
7039 NI instructions. A defect in the implementation of "multiblock" can cause
7040 OpenSSL's internal write buffer to become incorrectly set to NULL when
7041 using non-blocking IO. Typically, when the user application is using a
7042 socket BIO for writing, this will only result in a failed connection.
7043 However if some other BIO is used then it is likely that a segmentation
7044 fault will be triggered, thus enabling a potential DoS attack.
7046 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
7051 * Segmentation fault in DTLSv1_listen fix
7053 The DTLSv1_listen function is intended to be stateless and processes the
7054 initial ClientHello from many peers. It is common for user code to loop
7055 over the call to DTLSv1_listen until a valid ClientHello is received with
7056 an associated cookie. A defect in the implementation of DTLSv1_listen means
7057 that state is preserved in the SSL object from one invocation to the next
7058 that can lead to a segmentation fault. Errors processing the initial
7059 ClientHello can trigger this scenario. An example of such an error could be
7060 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
7063 This issue was reported to OpenSSL by Per Allansson.
7068 * Segmentation fault in ASN1_TYPE_cmp fix
7070 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7071 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7072 certificate signature algorithm consistency this can be used to crash any
7073 certificate verification operation and exploited in a DoS attack. Any
7074 application which performs certificate verification is vulnerable including
7075 OpenSSL clients and servers which enable client authentication.
7080 * Segmentation fault for invalid PSS parameters fix
7082 The signature verification routines will crash with a NULL pointer
7083 dereference if presented with an ASN.1 signature using the RSA PSS
7084 algorithm and invalid parameters. Since these routines are used to verify
7085 certificate signature algorithms this can be used to crash any
7086 certificate verification operation and exploited in a DoS attack. Any
7087 application which performs certificate verification is vulnerable including
7088 OpenSSL clients and servers which enable client authentication.
7090 This issue was was reported to OpenSSL by Brian Carpenter.
7095 * ASN.1 structure reuse memory corruption fix
7097 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7098 memory corruption via an invalid write. Such reuse is and has been
7099 strongly discouraged and is believed to be rare.
7101 Applications that parse structures containing CHOICE or ANY DEFINED BY
7102 components may be affected. Certificate parsing (d2i_X509 and related
7103 functions) are however not affected. OpenSSL clients and servers are
7109 * PKCS7 NULL pointer dereferences fix
7111 The PKCS#7 parsing code does not handle missing outer ContentInfo
7112 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7113 missing content and trigger a NULL pointer dereference on parsing.
7115 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7116 otherwise parse PKCS#7 structures from untrusted sources are
7117 affected. OpenSSL clients and servers are not affected.
7119 This issue was reported to OpenSSL by Michal Zalewski (Google).
7124 * DoS via reachable assert in SSLv2 servers fix
7126 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7127 servers that both support SSLv2 and enable export cipher suites by sending
7128 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7130 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7131 (OpenSSL development team).
7136 * Empty CKE with client auth and DHE fix
7138 If client auth is used then a server can seg fault in the event of a DHE
7139 ciphersuite being selected and a zero length ClientKeyExchange message
7140 being sent by the client. This could be exploited in a DoS attack.
7145 * Handshake with unseeded PRNG fix
7147 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
7148 with an unseeded PRNG. The conditions are:
7149 - The client is on a platform where the PRNG has not been seeded
7150 automatically, and the user has not seeded manually
7151 - A protocol specific client method version has been used (i.e. not
7152 SSL_client_methodv23)
7153 - A ciphersuite is used that does not require additional random data from
7154 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
7156 If the handshake succeeds then the client random that has been used will
7157 have been generated from a PRNG with insufficient entropy and therefore the
7158 output may be predictable.
7160 For example using the following command with an unseeded openssl will
7161 succeed on an unpatched platform:
7163 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
7168 * Use After Free following d2i_ECPrivatekey error fix
7170 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7171 could cause a use after free condition. This, in turn, could cause a double
7172 free in several private key parsing functions (such as d2i_PrivateKey
7173 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7174 for applications that receive EC private keys from untrusted
7175 sources. This scenario is considered rare.
7177 This issue was discovered by the BoringSSL project and fixed in their
7183 * X509_to_X509_REQ NULL pointer deref fix
7185 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7186 the certificate key is invalid. This function is rarely used in practice.
7188 This issue was discovered by Brian Carpenter.
7193 * Removed the export ciphers from the DEFAULT ciphers
7197 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7199 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7200 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7201 So far those who have to target multiple platforms would compromise
7202 and argue that binary targeting say ARMv5 would still execute on
7203 ARMv8. "Universal" build resolves this compromise by providing
7204 near-optimal performance even on newer platforms.
7208 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7209 (other platforms pending).
7211 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7213 * Add support for the SignedCertificateTimestampList certificate and
7214 OCSP response extensions from RFC6962.
7218 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7219 for corner cases. (Certain input points at infinity could lead to
7220 bogus results, with non-infinity inputs mapped to infinity too.)
7224 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7225 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7226 common cases are optimized and there still is room for further
7227 improvements. Vector Permutation AES for Altivec is also added.
7231 * Add support for little-endian ppc64 Linux target.
7233 *Marcelo Cerri (IBM)*
7235 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7236 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7237 are optimized and there still is room for further improvements.
7238 Both 32- and 64-bit modes are supported.
7240 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7242 * Improved ARMv7 NEON support.
7246 * Support for SPARC Architecture 2011 crypto extensions, first
7247 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7248 SHA256/512, MD5, GHASH and modular exponentiation.
7250 *Andy Polyakov, David Miller*
7252 * Accelerated modular exponentiation for Intel processors, a.k.a.
7255 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7257 * Support for new and upcoming Intel processors, including AVX2,
7258 BMI and SHA ISA extensions. This includes additional "stitched"
7259 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7262 This work was sponsored by Intel Corp.
7266 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7267 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7268 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7272 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7273 this fixes a limitation in previous versions of OpenSSL.
7277 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7278 MGF1 digest and OAEP label.
7282 * Add EVP support for key wrapping algorithms, to avoid problems with
7283 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7284 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7285 algorithms and include tests cases.
7289 * Add functions to allocate and set the fields of an ECDSA_METHOD
7292 *Douglas E. Engert, Steve Henson*
7294 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7295 difference in days and seconds between two tm or ASN1_TIME structures.
7299 * Add -rev test option to s_server to just reverse order of characters
7300 received by client and send back to server. Also prints an abbreviated
7301 summary of the connection parameters.
7305 * New option -brief for s_client and s_server to print out a brief summary
7306 of connection parameters.
7310 * Add callbacks for arbitrary TLS extensions.
7312 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7314 * New option -crl_download in several openssl utilities to download CRLs
7315 from CRLDP extension in certificates.
7319 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7323 * New function X509_CRL_diff to generate a delta CRL from the difference
7324 of two full CRLs. Add support to "crl" utility.
7328 * New functions to set lookup_crls function and to retrieve
7329 X509_STORE from X509_STORE_CTX.
7333 * Print out deprecated issuer and subject unique ID fields in
7338 * Extend OCSP I/O functions so they can be used for simple general purpose
7339 HTTP as well as OCSP. New wrapper function which can be used to download
7340 CRLs using the OCSP API.
7344 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7348 * `SSL_CONF*` functions. These provide a common framework for application
7349 configuration using configuration files or command lines.
7353 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7354 message callback and prints the results. Needs compile time option
7355 "enable-ssl-trace". New options to s_client and s_server to enable
7360 * New ctrl and macro to retrieve supported points extensions.
7361 Print out extension in s_server and s_client.
7365 * New functions to retrieve certificate signature and signature
7370 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7375 * New Suite B modes for TLS code. These use and enforce the requirements
7376 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7377 only use Suite B curves. The Suite B modes can be set by using the
7378 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7382 * New chain verification flags for Suite B levels of security. Check
7383 algorithms are acceptable when flags are set in X509_verify_cert.
7387 * Make tls1_check_chain return a set of flags indicating checks passed
7388 by a certificate chain. Add additional tests to handle client
7389 certificates: checks for matching certificate type and issuer name
7394 * If an attempt is made to use a signature algorithm not in the peer
7395 preference list abort the handshake. If client has no suitable
7396 signature algorithms in response to a certificate request do not
7397 use the certificate.
7401 * If server EC tmp key is not in client preference list abort handshake.
7405 * Add support for certificate stores in CERT structure. This makes it
7406 possible to have different stores per SSL structure or one store in
7407 the parent SSL_CTX. Include distinct stores for certificate chain
7408 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7409 to build and store a certificate chain in CERT structure: returning
7410 an error if the chain cannot be built: this will allow applications
7411 to test if a chain is correctly configured.
7413 Note: if the CERT based stores are not set then the parent SSL_CTX
7414 store is used to retain compatibility with existing behaviour.
7418 * New function ssl_set_client_disabled to set a ciphersuite disabled
7419 mask based on the current session, check mask when sending client
7420 hello and checking the requested ciphersuite.
7424 * New ctrls to retrieve and set certificate types in a certificate
7425 request message. Print out received values in s_client. If certificate
7426 types is not set with custom values set sensible values based on
7427 supported signature algorithms.
7431 * Support for distinct client and server supported signature algorithms.
7435 * Add certificate callback. If set this is called whenever a certificate
7436 is required by client or server. An application can decide which
7437 certificate chain to present based on arbitrary criteria: for example
7438 supported signature algorithms. Add very simple example to s_server.
7439 This fixes many of the problems and restrictions of the existing client
7440 certificate callback: for example you can now clear an existing
7441 certificate and specify the whole chain.
7445 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7446 the certificate can be used for (if anything). Set valid_flags field
7447 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7448 to have similar checks in it.
7450 Add new "cert_flags" field to CERT structure and include a "strict mode".
7451 This enforces some TLS certificate requirements (such as only permitting
7452 certificate signature algorithms contained in the supported algorithms
7453 extension) which some implementations ignore: this option should be used
7454 with caution as it could cause interoperability issues.
7458 * Update and tidy signature algorithm extension processing. Work out
7459 shared signature algorithms based on preferences and peer algorithms
7460 and print them out in s_client and s_server. Abort handshake if no
7461 shared signature algorithms.
7465 * Add new functions to allow customised supported signature algorithms
7466 for SSL and SSL_CTX structures. Add options to s_client and s_server
7471 * New function SSL_certs_clear() to delete all references to certificates
7472 from an SSL structure. Before this once a certificate had been added
7473 it couldn't be removed.
7477 * Integrate hostname, email address and IP address checking with certificate
7478 verification. New verify options supporting checking in openssl utility.
7482 * Fixes and wildcard matching support to hostname and email checking
7483 functions. Add manual page.
7485 *Florian Weimer (Red Hat Product Security Team)*
7487 * New functions to check a hostname email or IP address against a
7488 certificate. Add options x509 utility to print results of checks against
7493 * Fix OCSP checking.
7495 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7497 * Initial experimental support for explicitly trusted non-root CAs.
7498 OpenSSL still tries to build a complete chain to a root but if an
7499 intermediate CA has a trust setting included that is used. The first
7500 setting is used: whether to trust (e.g., -addtrust option to the x509
7505 * Add -trusted_first option which attempts to find certificates in the
7506 trusted store even if an untrusted chain is also supplied.
7510 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7511 platform support for Linux and Android.
7515 * Support for linux-x32, ILP32 environment in x86_64 framework.
7519 * Experimental multi-implementation support for FIPS capable OpenSSL.
7520 When in FIPS mode the approved implementations are used as normal,
7521 when not in FIPS mode the internal unapproved versions are used instead.
7522 This means that the FIPS capable OpenSSL isn't forced to use the
7523 (often lower performance) FIPS implementations outside FIPS mode.
7527 * Transparently support X9.42 DH parameters when calling
7528 PEM_read_bio_DHparameters. This means existing applications can handle
7529 the new parameter format automatically.
7533 * Initial experimental support for X9.42 DH parameter format: mainly
7534 to support use of 'q' parameter for RFC5114 parameters.
7538 * Add DH parameters from RFC5114 including test data to dhtest.
7542 * Support for automatic EC temporary key parameter selection. If enabled
7543 the most preferred EC parameters are automatically used instead of
7544 hardcoded fixed parameters. Now a server just has to call:
7545 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7546 support ECDH and use the most appropriate parameters.
7550 * Enhance and tidy EC curve and point format TLS extension code. Use
7551 static structures instead of allocation if default values are used.
7552 New ctrls to set curves we wish to support and to retrieve shared curves.
7553 Print out shared curves in s_server. New options to s_server and s_client
7554 to set list of supported curves.
7558 * New ctrls to retrieve supported signature algorithms and
7559 supported curve values as an array of NIDs. Extend openssl utility
7560 to print out received values.
7564 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7565 between NIDs and the more common NIST names such as "P-256". Enhance
7566 ecparam utility and ECC method to recognise the NIST names for curves.
7570 * Enhance SSL/TLS certificate chain handling to support different
7571 chains for each certificate instead of one chain in the parent SSL_CTX.
7575 * Support for fixed DH ciphersuite client authentication: where both
7576 server and client use DH certificates with common parameters.
7580 * Support for fixed DH ciphersuites: those requiring DH server
7585 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7587 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7588 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7589 X509_CINF_get_signature were reverted post internal team review.
7594 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7596 * OCSP Status Request extension unbounded memory growth
7598 A malicious client can send an excessively large OCSP Status Request
7599 extension. If that client continually requests renegotiation, sending a
7600 large OCSP Status Request extension each time, then there will be unbounded
7601 memory growth on the server. This will eventually lead to a Denial Of
7602 Service attack through memory exhaustion. Servers with a default
7603 configuration are vulnerable even if they do not support OCSP. Builds using
7604 the "no-ocsp" build time option are not affected.
7606 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7611 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7614 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7620 * OOB write in MDC2_Update()
7622 An overflow can occur in MDC2_Update() either if called directly or
7623 through the EVP_DigestUpdate() function using MDC2. If an attacker
7624 is able to supply very large amounts of input data after a previous
7625 call to EVP_EncryptUpdate() with a partial block then a length check
7626 can overflow resulting in a heap corruption.
7628 The amount of data needed is comparable to SIZE_MAX which is impractical
7631 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7636 * Malformed SHA512 ticket DoS
7638 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7639 DoS attack where a malformed ticket will result in an OOB read which will
7642 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7643 a custom server callback and ticket lookup mechanism.
7645 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7650 * OOB write in BN_bn2dec()
7652 The function BN_bn2dec() does not check the return value of BN_div_word().
7653 This can cause an OOB write if an application uses this function with an
7654 overly large BIGNUM. This could be a problem if an overly large certificate
7655 or CRL is printed out from an untrusted source. TLS is not affected because
7656 record limits will reject an oversized certificate before it is parsed.
7658 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7663 * OOB read in TS_OBJ_print_bio()
7665 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7666 the total length the OID text representation would use and not the amount
7667 of data written. This will result in OOB reads when large OIDs are
7670 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7675 * Pointer arithmetic undefined behaviour
7677 Avoid some undefined pointer arithmetic
7679 A common idiom in the codebase is to check limits in the following manner:
7682 Where "p" points to some malloc'd data of SIZE bytes and
7685 "len" here could be from some externally supplied data (e.g. from a TLS
7688 The rules of C pointer arithmetic are such that "p + len" is only well
7689 defined where len <= SIZE. Therefore, the above idiom is actually
7690 undefined behaviour.
7692 For example this could cause problems if some malloc implementation
7693 provides an address for "p" such that "p + len" actually overflows for
7694 values of len that are too big and therefore p + len < limit.
7696 This issue was reported to OpenSSL by Guido Vranken
7701 * Constant time flag not preserved in DSA signing
7703 Operations in the DSA signing algorithm should run in constant time in
7704 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7705 implementation means that a non-constant time codepath is followed for
7706 certain operations. This has been demonstrated through a cache-timing
7707 attack to be sufficient for an attacker to recover the private DSA key.
7709 This issue was reported by César Pereida (Aalto University), Billy Brumley
7710 (Tampere University of Technology), and Yuval Yarom (The University of
7711 Adelaide and NICTA).
7716 * DTLS buffered message DoS
7718 In a DTLS connection where handshake messages are delivered out-of-order
7719 those messages that OpenSSL is not yet ready to process will be buffered
7720 for later use. Under certain circumstances, a flaw in the logic means that
7721 those messages do not get removed from the buffer even though the handshake
7722 has been completed. An attacker could force up to approx. 15 messages to
7723 remain in the buffer when they are no longer required. These messages will
7724 be cleared when the DTLS connection is closed. The default maximum size for
7725 a message is 100k. Therefore, the attacker could force an additional 1500k
7726 to be consumed per connection. By opening many simultaneous connections an
7727 attacker could cause a DoS attack through memory exhaustion.
7729 This issue was reported to OpenSSL by Quan Luo.
7734 * DTLS replay protection DoS
7736 A flaw in the DTLS replay attack protection mechanism means that records
7737 that arrive for future epochs update the replay protection "window" before
7738 the MAC for the record has been validated. This could be exploited by an
7739 attacker by sending a record for the next epoch (which does not have to
7740 decrypt or have a valid MAC), with a very large sequence number. This means
7741 that all subsequent legitimate packets are dropped causing a denial of
7742 service for a specific DTLS connection.
7744 This issue was reported to OpenSSL by the OCAP audit team.
7749 * Certificate message OOB reads
7751 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7752 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7753 theoretical DoS risk but this has not been observed in practice on common
7756 The messages affected are client certificate, client certificate request
7757 and server certificate. As a result the attack can only be performed
7758 against a client or a server which enables client authentication.
7760 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7765 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7767 * Prevent padding oracle in AES-NI CBC MAC check
7769 A MITM attacker can use a padding oracle attack to decrypt traffic
7770 when the connection uses an AES CBC cipher and the server support
7773 This issue was introduced as part of the fix for Lucky 13 padding
7774 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7775 constant time by making sure that always the same bytes are read and
7776 compared against either the MAC or padding bytes. But it no longer
7777 checked that there was enough data to have both the MAC and padding
7780 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7785 * Fix EVP_EncodeUpdate overflow
7787 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7788 Base64 encoding of binary data. If an attacker is able to supply very large
7789 amounts of input data then a length check can overflow resulting in a heap
7792 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7793 the `PEM_write_bio*` family of functions. These are mainly used within the
7794 OpenSSL command line applications, so any application which processes data
7795 from an untrusted source and outputs it as a PEM file should be considered
7796 vulnerable to this issue. User applications that call these APIs directly
7797 with large amounts of untrusted data may also be vulnerable.
7799 This issue was reported by Guido Vranken.
7804 * Fix EVP_EncryptUpdate overflow
7806 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7807 is able to supply very large amounts of input data after a previous call to
7808 EVP_EncryptUpdate() with a partial block then a length check can overflow
7809 resulting in a heap corruption. Following an analysis of all OpenSSL
7810 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7811 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7812 the first called function after an EVP_EncryptInit(), and therefore that
7813 specific call must be safe. The second form is where the length passed to
7814 EVP_EncryptUpdate() can be seen from the code to be some small value and
7815 therefore there is no possibility of an overflow. Since all instances are
7816 one of these two forms, it is believed that there can be no overflows in
7817 internal code due to this problem. It should be noted that
7818 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7819 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7820 of these calls have also been analysed too and it is believed there are no
7821 instances in internal usage where an overflow could occur.
7823 This issue was reported by Guido Vranken.
7828 * Prevent ASN.1 BIO excessive memory allocation
7830 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7831 a short invalid encoding can casuse allocation of large amounts of memory
7832 potentially consuming excessive resources or exhausting memory.
7834 Any application parsing untrusted data through d2i BIO functions is
7835 affected. The memory based functions such as d2i_X509() are *not* affected.
7836 Since the memory based functions are used by the TLS library, TLS
7837 applications are not affected.
7839 This issue was reported by Brian Carpenter.
7846 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7847 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7848 in arbitrary stack data being returned in the buffer.
7850 This issue was reported by Guido Vranken.
7855 * Modify behavior of ALPN to invoke callback after SNI/servername
7856 callback, such that updates to the SSL_CTX affect ALPN.
7860 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7865 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7866 methods are enabled and ssl2 is disabled the methods return NULL.
7870 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7872 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7873 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7874 provide any "EXPORT" or "LOW" strength ciphers.
7878 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7879 is by default disabled at build-time. Builds that are not configured with
7880 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7881 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7882 will need to explicitly call either of:
7884 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7886 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7888 as appropriate. Even if either of those is used, or the application
7889 explicitly uses the version-specific SSLv2_method() or its client and
7890 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7891 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7892 ciphers, and SSLv2 56-bit DES are no longer available.
7897 * Fix a double-free in DSA code
7899 A double free bug was discovered when OpenSSL parses malformed DSA private
7900 keys and could lead to a DoS attack or memory corruption for applications
7901 that receive DSA private keys from untrusted sources. This scenario is
7904 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7910 * Disable SRP fake user seed to address a server memory leak.
7912 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7914 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7915 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7916 was changed to ignore the "fake user" SRP seed, even if the seed
7919 Users should use SRP_VBASE_get1_by_user instead. Note that in
7920 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7921 also that even though configuring the SRP seed attempts to hide
7922 invalid usernames by continuing the handshake with fake
7923 credentials, this behaviour is not constant time and no strong
7924 guarantees are made that the handshake is indistinguishable from
7925 that of a valid user.
7930 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7932 In the BN_hex2bn function the number of hex digits is calculated using an
7933 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7934 large values of `i` this can result in `bn_expand` not allocating any
7935 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7936 field as NULL leading to a subsequent NULL ptr deref. For very large values
7937 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7938 In this case memory is allocated to the internal BIGNUM data field, but it
7939 is insufficiently sized leading to heap corruption. A similar issue exists
7940 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7941 is ever called by user applications with very large untrusted hex/dec data.
7942 This is anticipated to be a rare occurrence.
7944 All OpenSSL internal usage of these functions use data that is not expected
7945 to be untrusted, e.g. config file data or application command line
7946 arguments. If user developed applications generate config file data based
7947 on untrusted data then it is possible that this could also lead to security
7948 consequences. This is also anticipated to be rare.
7950 This issue was reported to OpenSSL by Guido Vranken.
7955 * Fix memory issues in `BIO_*printf` functions
7957 The internal `fmtstr` function used in processing a "%s" format string in
7958 the `BIO_*printf` functions could overflow while calculating the length of a
7959 string and cause an OOB read when printing very long strings.
7961 Additionally the internal `doapr_outch` function can attempt to write to an
7962 OOB memory location (at an offset from the NULL pointer) in the event of a
7963 memory allocation failure. In 1.0.2 and below this could be caused where
7964 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7965 could be in processing a very long "%s" format string. Memory leaks can
7968 The first issue may mask the second issue dependent on compiler behaviour.
7969 These problems could enable attacks where large amounts of untrusted data
7970 is passed to the `BIO_*printf` functions. If applications use these functions
7971 in this way then they could be vulnerable. OpenSSL itself uses these
7972 functions when printing out human-readable dumps of ASN.1 data. Therefore
7973 applications that print this data could be vulnerable if the data is from
7974 untrusted sources. OpenSSL command line applications could also be
7975 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7976 as command line arguments.
7978 Libssl is not considered directly vulnerable. Additionally certificates etc
7979 received via remote connections via libssl are also unlikely to be able to
7980 trigger these issues because of message size limits enforced within libssl.
7982 This issue was reported to OpenSSL Guido Vranken.
7987 * Side channel attack on modular exponentiation
7989 A side-channel attack was found which makes use of cache-bank conflicts on
7990 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7991 of RSA keys. The ability to exploit this issue is limited as it relies on
7992 an attacker who has control of code in a thread running on the same
7993 hyper-threaded core as the victim thread which is performing decryptions.
7995 This issue was reported to OpenSSL by Yuval Yarom, The University of
7996 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7997 Nadia Heninger, University of Pennsylvania with more information at
7998 <http://cachebleed.info>.
8003 * Change the req command to generate a 2048-bit RSA/DSA key by default,
8004 if no keysize is specified with default_bits. This fixes an
8005 omission in an earlier change that changed all RSA/DSA key generation
8006 commands to use 2048 bits by default.
8010 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
8012 * Protection for DH small subgroup attacks
8014 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
8015 switched on by default and cannot be disabled. This could have some
8020 * SSLv2 doesn't block disabled ciphers
8022 A malicious client can negotiate SSLv2 ciphers that have been disabled on
8023 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
8024 been disabled, provided that the SSLv2 protocol was not also disabled via
8027 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
8028 and Sebastian Schinzel.
8033 * Reject DH handshakes with parameters shorter than 1024 bits.
8037 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
8039 * Certificate verify crash with missing PSS parameter
8041 The signature verification routines will crash with a NULL pointer
8042 dereference if presented with an ASN.1 signature using the RSA PSS
8043 algorithm and absent mask generation function parameter. Since these
8044 routines are used to verify certificate signature algorithms this can be
8045 used to crash any certificate verification operation and exploited in a
8046 DoS attack. Any application which performs certificate verification is
8047 vulnerable including OpenSSL clients and servers which enable client
8050 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
8055 * X509_ATTRIBUTE memory leak
8057 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8058 memory. This structure is used by the PKCS#7 and CMS routines so any
8059 application which reads PKCS#7 or CMS data from untrusted sources is
8060 affected. SSL/TLS is not affected.
8062 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8068 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
8069 This changes the decoding behaviour for some invalid messages,
8070 though the change is mostly in the more lenient direction, and
8071 legacy behaviour is preserved as much as possible.
8075 * In DSA_generate_parameters_ex, if the provided seed is too short,
8076 use a random seed, as already documented.
8078 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
8080 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
8082 * Alternate chains certificate forgery
8084 During certificate verification, OpenSSL will attempt to find an
8085 alternative certificate chain if the first attempt to build such a chain
8086 fails. An error in the implementation of this logic can mean that an
8087 attacker could cause certain checks on untrusted certificates to be
8088 bypassed, such as the CA flag, enabling them to use a valid leaf
8089 certificate to act as a CA and "issue" an invalid certificate.
8091 This issue was reported to OpenSSL by Adam Langley/David Benjamin
8097 * Race condition handling PSK identify hint
8099 If PSK identity hints are received by a multi-threaded client then
8100 the values are wrongly updated in the parent SSL_CTX structure. This can
8101 result in a race condition potentially leading to a double free of the
8107 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
8109 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
8110 incompatibility in the handling of HMAC. The previous ABI has now been
8113 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
8115 * Malformed ECParameters causes infinite loop
8117 When processing an ECParameters structure OpenSSL enters an infinite loop
8118 if the curve specified is over a specially malformed binary polynomial
8121 This can be used to perform denial of service against any
8122 system which processes public keys, certificate requests or
8123 certificates. This includes TLS clients and TLS servers with
8124 client authentication enabled.
8126 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8131 * Exploitable out-of-bounds read in X509_cmp_time
8133 X509_cmp_time does not properly check the length of the ASN1_TIME
8134 string and can read a few bytes out of bounds. In addition,
8135 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8138 An attacker can use this to craft malformed certificates and CRLs of
8139 various sizes and potentially cause a segmentation fault, resulting in
8140 a DoS on applications that verify certificates or CRLs. TLS clients
8141 that verify CRLs are affected. TLS clients and servers with client
8142 authentication enabled may be affected if they use custom verification
8145 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8146 independently by Hanno Böck.
8151 * PKCS7 crash with missing EnvelopedContent
8153 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8154 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8155 with missing content and trigger a NULL pointer dereference on parsing.
8157 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8158 structures from untrusted sources are affected. OpenSSL clients and
8159 servers are not affected.
8161 This issue was reported to OpenSSL by Michal Zalewski (Google).
8166 * CMS verify infinite loop with unknown hash function
8168 When verifying a signedData message the CMS code can enter an infinite loop
8169 if presented with an unknown hash function OID. This can be used to perform
8170 denial of service against any system which verifies signedData messages using
8172 This issue was reported to OpenSSL by Johannes Bauer.
8177 * Race condition handling NewSessionTicket
8179 If a NewSessionTicket is received by a multi-threaded client when attempting to
8180 reuse a previous ticket then a race condition can occur potentially leading to
8181 a double free of the ticket data.
8186 * Reject DH handshakes with parameters shorter than 768 bits.
8188 *Kurt Roeckx and Emilia Kasper*
8190 * dhparam: generate 2048-bit parameters by default.
8192 *Kurt Roeckx and Emilia Kasper*
8194 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8196 * Segmentation fault in ASN1_TYPE_cmp fix
8198 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8199 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8200 certificate signature algorithm consistency this can be used to crash any
8201 certificate verification operation and exploited in a DoS attack. Any
8202 application which performs certificate verification is vulnerable including
8203 OpenSSL clients and servers which enable client authentication.
8208 * ASN.1 structure reuse memory corruption fix
8210 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8211 memory corruption via an invalid write. Such reuse is and has been
8212 strongly discouraged and is believed to be rare.
8214 Applications that parse structures containing CHOICE or ANY DEFINED BY
8215 components may be affected. Certificate parsing (d2i_X509 and related
8216 functions) are however not affected. OpenSSL clients and servers are
8222 * PKCS7 NULL pointer dereferences fix
8224 The PKCS#7 parsing code does not handle missing outer ContentInfo
8225 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8226 missing content and trigger a NULL pointer dereference on parsing.
8228 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8229 otherwise parse PKCS#7 structures from untrusted sources are
8230 affected. OpenSSL clients and servers are not affected.
8232 This issue was reported to OpenSSL by Michal Zalewski (Google).
8237 * DoS via reachable assert in SSLv2 servers fix
8239 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8240 servers that both support SSLv2 and enable export cipher suites by sending
8241 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8243 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8244 (OpenSSL development team).
8249 * Use After Free following d2i_ECPrivatekey error fix
8251 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8252 could cause a use after free condition. This, in turn, could cause a double
8253 free in several private key parsing functions (such as d2i_PrivateKey
8254 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8255 for applications that receive EC private keys from untrusted
8256 sources. This scenario is considered rare.
8258 This issue was discovered by the BoringSSL project and fixed in their
8264 * X509_to_X509_REQ NULL pointer deref fix
8266 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8267 the certificate key is invalid. This function is rarely used in practice.
8269 This issue was discovered by Brian Carpenter.
8274 * Removed the export ciphers from the DEFAULT ciphers
8278 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8280 * Build fixes for the Windows and OpenVMS platforms
8282 *Matt Caswell and Richard Levitte*
8284 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8286 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8287 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8288 dereference. This could lead to a Denial Of Service attack. Thanks to
8289 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8294 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8295 dtls1_buffer_record function under certain conditions. In particular this
8296 could occur if an attacker sent repeated DTLS records with the same
8297 sequence number but for the next epoch. The memory leak could be exploited
8298 by an attacker in a Denial of Service attack through memory exhaustion.
8299 Thanks to Chris Mueller for reporting this issue.
8304 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8305 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8306 method would be set to NULL which could later result in a NULL pointer
8307 dereference. Thanks to Frank Schmirler for reporting this issue.
8312 * Abort handshake if server key exchange message is omitted for ephemeral
8315 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8316 reporting this issue.
8321 * Remove non-export ephemeral RSA code on client and server. This code
8322 violated the TLS standard by allowing the use of temporary RSA keys in
8323 non-export ciphersuites and could be used by a server to effectively
8324 downgrade the RSA key length used to a value smaller than the server
8325 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8326 INRIA or reporting this issue.
8331 * Fixed issue where DH client certificates are accepted without verification.
8332 An OpenSSL server will accept a DH certificate for client authentication
8333 without the certificate verify message. This effectively allows a client to
8334 authenticate without the use of a private key. This only affects servers
8335 which trust a client certificate authority which issues certificates
8336 containing DH keys: these are extremely rare and hardly ever encountered.
8337 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8343 * Ensure that the session ID context of an SSL is updated when its
8344 SSL_CTX is updated via SSL_set_SSL_CTX.
8346 The session ID context is typically set from the parent SSL_CTX,
8347 and can vary with the CTX.
8351 * Fix various certificate fingerprint issues.
8353 By using non-DER or invalid encodings outside the signed portion of a
8354 certificate the fingerprint can be changed without breaking the signature.
8355 Although no details of the signed portion of the certificate can be changed
8356 this can cause problems with some applications: e.g. those using the
8357 certificate fingerprint for blacklists.
8359 1. Reject signatures with non zero unused bits.
8361 If the BIT STRING containing the signature has non zero unused bits reject
8362 the signature. All current signature algorithms require zero unused bits.
8364 2. Check certificate algorithm consistency.
8366 Check the AlgorithmIdentifier inside TBS matches the one in the
8367 certificate signature. NB: this will result in signature failure
8368 errors for some broken certificates.
8370 Thanks to Konrad Kraszewski from Google for reporting this issue.
8372 3. Check DSA/ECDSA signatures use DER.
8374 Re-encode DSA/ECDSA signatures and compare with the original received
8375 signature. Return an error if there is a mismatch.
8377 This will reject various cases including garbage after signature
8378 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8379 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8380 (negative or with leading zeroes).
8382 Further analysis was conducted and fixes were developed by Stephen Henson
8383 of the OpenSSL core team.
8389 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8390 results on some platforms, including x86_64. This bug occurs at random
8391 with a very low probability, and is not known to be exploitable in any
8392 way, though its exact impact is difficult to determine. Thanks to Pieter
8393 Wuille (Blockstream) who reported this issue and also suggested an initial
8394 fix. Further analysis was conducted by the OpenSSL development team and
8395 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8396 the OpenSSL core team.
8401 * Do not resume sessions on the server if the negotiated protocol
8402 version does not match the session's version. Resuming with a different
8403 version, while not strictly forbidden by the RFC, is of questionable
8404 sanity and breaks all known clients.
8406 *David Benjamin, Emilia Käsper*
8408 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8409 early CCS messages during renegotiation. (Note that because
8410 renegotiation is encrypted, this early CCS was not exploitable.)
8414 * Tighten client-side session ticket handling during renegotiation:
8415 ensure that the client only accepts a session ticket if the server sends
8416 the extension anew in the ServerHello. Previously, a TLS client would
8417 reuse the old extension state and thus accept a session ticket if one was
8418 announced in the initial ServerHello.
8420 Similarly, ensure that the client requires a session ticket if one
8421 was advertised in the ServerHello. Previously, a TLS client would
8422 ignore a missing NewSessionTicket message.
8426 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8430 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8431 sends a carefully crafted handshake message, to cause OpenSSL to fail
8432 to free up to 64k of memory causing a memory leak. This could be
8433 exploited in a Denial Of Service attack. This issue affects OpenSSL
8434 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8435 whether SRTP is used or configured. Implementations of OpenSSL that
8436 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8438 The fix was developed by the OpenSSL team.
8443 * Session Ticket Memory Leak.
8445 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8446 integrity of that ticket is first verified. In the event of a session
8447 ticket integrity check failing, OpenSSL will fail to free memory
8448 causing a memory leak. By sending a large number of invalid session
8449 tickets an attacker could exploit this issue in a Denial Of Service
8455 * Build option no-ssl3 is incomplete.
8457 When OpenSSL is configured with "no-ssl3" as a build option, servers
8458 could accept and complete an SSL 3.0 handshake, and clients could be
8459 configured to send them.
8462 *Akamai and the OpenSSL team*
8464 * Add support for TLS_FALLBACK_SCSV.
8465 Client applications doing fallback retries should call
8466 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8469 *Adam Langley, Bodo Moeller*
8471 * Add additional DigestInfo checks.
8473 Re-encode DigestInto in DER and check against the original when
8474 verifying RSA signature: this will reject any improperly encoded
8475 DigestInfo structures.
8477 Note: this is a precautionary measure and no attacks are currently known.
8481 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8483 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8484 SRP code can be overrun an internal buffer. Add sanity check that
8485 g, A, B < N to SRP code.
8487 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8488 Group for discovering this issue.
8493 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8494 TLS 1.0 instead of higher protocol versions when the ClientHello message
8495 is badly fragmented. This allows a man-in-the-middle attacker to force a
8496 downgrade to TLS 1.0 even if both the server and the client support a
8497 higher protocol version, by modifying the client's TLS records.
8499 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8500 researching this issue.
8505 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8506 to a denial of service attack. A malicious server can crash the client
8507 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8508 ciphersuite and sending carefully crafted handshake messages.
8510 Thanks to Felix Gröbert (Google) for discovering and researching this
8516 * By sending carefully crafted DTLS packets an attacker could cause openssl
8517 to leak memory. This can be exploited through a Denial of Service attack.
8518 Thanks to Adam Langley for discovering and researching this issue.
8523 * An attacker can force openssl to consume large amounts of memory whilst
8524 processing DTLS handshake messages. This can be exploited through a
8525 Denial of Service attack.
8526 Thanks to Adam Langley for discovering and researching this issue.
8531 * An attacker can force an error condition which causes openssl to crash
8532 whilst processing DTLS packets due to memory being freed twice. This
8533 can be exploited through a Denial of Service attack.
8534 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8540 * If a multithreaded client connects to a malicious server using a resumed
8541 session and the server sends an ec point format extension it could write
8542 up to 255 bytes to freed memory.
8544 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8550 * A malicious server can crash an OpenSSL client with a null pointer
8551 dereference (read) by specifying an SRP ciphersuite even though it was not
8552 properly negotiated with the client. This can be exploited through a
8553 Denial of Service attack.
8555 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8556 discovering and researching this issue.
8561 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8562 X509_name_oneline, X509_name_print_ex et al. to leak some information
8563 from the stack. Applications may be affected if they echo pretty printing
8564 output to the attacker.
8566 Thanks to Ivan Fratric (Google) for discovering this issue.
8569 *Emilia Käsper, and Steve Henson*
8571 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8572 for corner cases. (Certain input points at infinity could lead to
8573 bogus results, with non-infinity inputs mapped to infinity too.)
8577 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8579 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8580 handshake can force the use of weak keying material in OpenSSL
8581 SSL/TLS clients and servers.
8583 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8584 researching this issue. ([CVE-2014-0224])
8586 *KIKUCHI Masashi, Steve Henson*
8588 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8589 OpenSSL DTLS client the code can be made to recurse eventually crashing
8592 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8595 *Imre Rad, Steve Henson*
8597 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8598 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8599 client or server. This is potentially exploitable to run arbitrary
8600 code on a vulnerable client or server.
8602 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8604 *Jüri Aedla, Steve Henson*
8606 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8607 are subject to a denial of service attack.
8609 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8610 this issue. ([CVE-2014-3470])
8612 *Felix Gröbert, Ivan Fratric, Steve Henson*
8614 * Harmonize version and its documentation. -f flag is used to display
8617 *mancha <mancha1@zoho.com>*
8619 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8620 in i2d_ECPrivateKey.
8622 *mancha <mancha1@zoho.com>*
8624 * Fix some double frees. These are not thought to be exploitable.
8626 *mancha <mancha1@zoho.com>*
8628 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8630 * A missing bounds check in the handling of the TLS heartbeat extension
8631 can be used to reveal up to 64k of memory to a connected client or
8634 Thanks for Neel Mehta of Google Security for discovering this bug and to
8635 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8636 preparing the fix ([CVE-2014-0160])
8638 *Adam Langley, Bodo Moeller*
8640 * Fix for the attack described in the paper "Recovering OpenSSL
8641 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8642 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8643 <http://eprint.iacr.org/2014/140>
8645 Thanks to Yuval Yarom and Naomi Benger for discovering this
8646 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8648 *Yuval Yarom and Naomi Benger*
8650 * TLS pad extension: draft-agl-tls-padding-03
8652 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8653 TLS client Hello record length value would otherwise be > 255 and
8654 less that 512 pad with a dummy extension containing zeroes so it
8655 is at least 512 bytes long.
8657 *Adam Langley, Steve Henson*
8659 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8661 * Fix for TLS record tampering bug. A carefully crafted invalid
8662 handshake could crash OpenSSL with a NULL pointer exception.
8663 Thanks to Anton Johansson for reporting this issues.
8666 * Keep original DTLS digest and encryption contexts in retransmission
8667 structures so we can use the previous session parameters if they need
8668 to be resent. ([CVE-2013-6450])
8672 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8673 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8674 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8675 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8676 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8677 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8679 *Rob Stradling, Adam Langley*
8681 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8683 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8684 supporting platforms or when small records were transferred.
8686 *Andy Polyakov, Steve Henson*
8688 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8690 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8692 This addresses the flaw in CBC record processing discovered by
8693 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8694 at: <http://www.isg.rhul.ac.uk/tls/>
8696 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8697 Security Group at Royal Holloway, University of London
8698 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8699 Emilia Käsper for the initial patch.
8702 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8704 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8705 ciphersuites which can be exploited in a denial of service attack.
8706 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8707 and detecting this bug and to Wolfgang Ettlinger
8708 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8713 * Return an error when checking OCSP signatures when key is NULL.
8714 This fixes a DoS attack. ([CVE-2013-0166])
8718 * Make openssl verify return errors.
8720 *Chris Palmer <palmer@google.com> and Ben Laurie*
8722 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8723 the right response is stapled. Also change SSL_get_certificate()
8724 so it returns the certificate actually sent.
8725 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8727 *Rob Stradling <rob.stradling@comodo.com>*
8729 * Fix possible deadlock when decoding public keys.
8733 * Don't use TLS 1.0 record version number in initial client hello
8738 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8740 * Sanity check record length before skipping explicit IV in TLS
8741 1.2, 1.1 and DTLS to fix DoS attack.
8743 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8744 fuzzing as a service testing platform.
8749 * Initialise tkeylen properly when encrypting CMS messages.
8750 Thanks to Solar Designer of Openwall for reporting this issue.
8754 * In FIPS mode don't try to use composite ciphers as they are not
8759 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8761 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8762 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8763 mean any application compiled against OpenSSL 1.0.0 headers setting
8764 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8765 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8766 0x10000000L Any application which was previously compiled against
8767 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8768 will need to be recompiled as a result. Letting be results in
8769 inability to disable specifically TLS 1.1 and in client context,
8770 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8774 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8775 disable just protocol X, but all protocols above X *if* there are
8776 protocols *below* X still enabled. In more practical terms it means
8777 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8778 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8779 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8784 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8786 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8787 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8788 in CRYPTO_realloc_clean.
8790 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8791 issue and to Adam Langley <agl@chromium.org> for fixing it.
8794 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8796 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8800 * Workarounds for some broken servers that "hang" if a client hello
8801 record length exceeds 255 bytes.
8803 1. Do not use record version number > TLS 1.0 in initial client
8804 hello: some (but not all) hanging servers will now work.
8805 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8806 the number of ciphers sent in the client hello. This should be
8807 set to an even number, such as 50, for example by passing:
8808 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8809 Most broken servers should now work.
8810 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8811 TLS 1.2 client support entirely.
8815 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8819 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8821 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8822 STRING form instead of a DigestInfo.
8826 * The format used for MDC2 RSA signatures is inconsistent between EVP
8827 and the RSA_sign/RSA_verify functions. This was made more apparent when
8828 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8829 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8830 the correct format in RSA_verify so both forms transparently work.
8834 * Some servers which support TLS 1.0 can choke if we initially indicate
8835 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8836 encrypted premaster secret. As a workaround use the maximum permitted
8837 client version in client hello, this should keep such servers happy
8838 and still work with previous versions of OpenSSL.
8842 * Add support for TLS/DTLS heartbeats.
8844 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8846 * Add support for SCTP.
8848 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8850 * Improved PRNG seeding for VOS.
8852 *Paul Green <Paul.Green@stratus.com>*
8854 * Extensive assembler packs updates, most notably:
8856 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8857 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8858 - x86_64: bit-sliced AES implementation;
8859 - ARM: NEON support, contemporary platforms optimizations;
8860 - s390x: z196 support;
8861 - `*`: GHASH and GF(2^m) multiplication implementations;
8865 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8866 (removal of unnecessary code)
8868 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8870 * Add TLS key material exporter from RFC 5705.
8874 * Add DTLS-SRTP negotiation from RFC 5764.
8878 * Add Next Protocol Negotiation,
8879 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8880 disabled with a no-npn flag to config or Configure. Code donated
8883 *Adam Langley <agl@google.com> and Ben Laurie*
8885 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8886 NIST-P256, NIST-P521, with constant-time single point multiplication on
8887 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8888 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8889 Code made available under Apache License version 2.0.
8891 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8892 line to include this in your build of OpenSSL, and run "make depend" (or
8893 "make update"). This enables the following EC_METHODs:
8895 EC_GFp_nistp224_method()
8896 EC_GFp_nistp256_method()
8897 EC_GFp_nistp521_method()
8899 EC_GROUP_new_by_curve_name() will automatically use these (while
8900 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8903 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8905 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8906 all platforms. Move ssize_t definition from e_os.h to the public
8907 header file e_os2.h as it now appears in public header file cms.h
8911 * New -sigopt option to the ca, req and x509 utilities. Additional
8912 signature parameters can be passed using this option and in
8917 * Add RSA PSS signing function. This will generate and set the
8918 appropriate AlgorithmIdentifiers for PSS based on those in the
8919 corresponding EVP_MD_CTX structure. No application support yet.
8923 * Support for companion algorithm specific ASN1 signing routines.
8924 New function ASN1_item_sign_ctx() signs a pre-initialised
8925 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8926 the appropriate parameters.
8930 * Add new algorithm specific ASN1 verification initialisation function
8931 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8932 handling will be the same no matter what EVP_PKEY_METHOD is used.
8933 Add a PSS handler to support verification of PSS signatures: checked
8934 against a number of sample certificates.
8938 * Add signature printing for PSS. Add PSS OIDs.
8940 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8942 * Add algorithm specific signature printing. An individual ASN1 method
8943 can now print out signatures instead of the standard hex dump.
8945 More complex signatures (e.g. PSS) can print out more meaningful
8946 information. Include DSA version that prints out the signature
8951 * Password based recipient info support for CMS library: implementing
8956 * Split password based encryption into PBES2 and PBKDF2 functions. This
8957 neatly separates the code into cipher and PBE sections and is required
8958 for some algorithms that split PBES2 into separate pieces (such as
8959 password based CMS).
8963 * Session-handling fixes:
8964 - Fix handling of connections that are resuming with a session ID,
8965 but also support Session Tickets.
8966 - Fix a bug that suppressed issuing of a new ticket if the client
8967 presented a ticket with an expired session.
8968 - Try to set the ticket lifetime hint to something reasonable.
8969 - Make tickets shorter by excluding irrelevant information.
8970 - On the client side, don't ignore renewed tickets.
8972 *Adam Langley, Bodo Moeller (Google)*
8974 * Fix PSK session representation.
8978 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8980 This work was sponsored by Intel.
8984 * Add GCM support to TLS library. Some custom code is needed to split
8985 the IV between the fixed (from PRF) and explicit (from TLS record)
8986 portions. This adds all GCM ciphersuites supported by RFC5288 and
8987 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8988 add a special AESGCM string for GCM only.
8992 * Expand range of ctrls for AES GCM. Permit setting invocation
8993 field on decrypt and retrieval of invocation field only on encrypt.
8997 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8998 As required by RFC5289 these ciphersuites cannot be used if for
8999 versions of TLS earlier than 1.2.
9003 * For FIPS capable OpenSSL interpret a NULL default public key method
9004 as unset and return the appropriate default but do *not* set the default.
9005 This means we can return the appropriate method in applications that
9006 switch between FIPS and non-FIPS modes.
9010 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
9011 ENGINE is used then we cannot handle that in the FIPS module so we
9012 keep original code iff non-FIPS operations are allowed.
9016 * Add -attime option to openssl utilities.
9018 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
9020 * Redirect DSA and DH operations to FIPS module in FIPS mode.
9024 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
9025 FIPS EC methods unconditionally for now.
9029 * New build option no-ec2m to disable characteristic 2 code.
9033 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
9034 all cases can be covered as some introduce binary incompatibilities.
9038 * Redirect RSA operations to FIPS module including keygen,
9039 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
9043 * Add similar low-level API blocking to ciphers.
9047 * low-level digest APIs are not approved in FIPS mode: any attempt
9048 to use these will cause a fatal error. Applications that *really* want
9049 to use them can use the `private_*` version instead.
9053 * Redirect cipher operations to FIPS module for FIPS builds.
9057 * Redirect digest operations to FIPS module for FIPS builds.
9061 * Update build system to add "fips" flag which will link in fipscanister.o
9062 for static and shared library builds embedding a signature if needed.
9066 * Output TLS supported curves in preference order instead of numerical
9067 order. This is currently hardcoded for the highest order curves first.
9068 This should be configurable so applications can judge speed vs strength.
9072 * Add TLS v1.2 server support for client authentication.
9076 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
9081 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
9082 FIPS modules versions.
9086 * Add TLS v1.2 client side support for client authentication. Keep cache
9087 of handshake records longer as we don't know the hash algorithm to use
9088 until after the certificate request message is received.
9092 * Initial TLS v1.2 client support. Add a default signature algorithms
9093 extension including all the algorithms we support. Parse new signature
9094 format in client key exchange. Relax some ECC signing restrictions for
9095 TLS v1.2 as indicated in RFC5246.
9099 * Add server support for TLS v1.2 signature algorithms extension. Switch
9100 to new signature format when needed using client digest preference.
9101 All server ciphersuites should now work correctly in TLS v1.2. No client
9102 support yet and no support for client certificates.
9106 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
9107 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
9108 ciphersuites. At present only RSA key exchange ciphersuites work with
9109 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
9110 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
9111 and version checking.
9115 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
9116 with this defined it will not be affected by any changes to ssl internal
9117 structures. Add several utility functions to allow openssl application
9118 to work with OPENSSL_NO_SSL_INTERN defined.
9122 * A long standing patch to add support for SRP from EdelWeb (Peter
9123 Sylvester and Christophe Renou) was integrated.
9124 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
9125 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
9128 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
9132 * Permit abbreviated handshakes when renegotiating using the function
9133 SSL_renegotiate_abbreviated().
9135 *Robin Seggelmann <seggelmann@fh-muenster.de>*
9137 * Add call to ENGINE_register_all_complete() to
9138 ENGINE_load_builtin_engines(), so some implementations get used
9139 automatically instead of needing explicit application support.
9143 * Add support for TLS key exporter as described in RFC5705.
9145 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
9147 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
9148 a few changes are required:
9150 Add SSL_OP_NO_TLSv1_1 flag.
9151 Add TLSv1_1 methods.
9152 Update version checking logic to handle version 1.1.
9153 Add explicit IV handling (ported from DTLS code).
9154 Add command line options to s_client/s_server.
9161 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
9163 * X509_ATTRIBUTE memory leak
9165 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
9166 memory. This structure is used by the PKCS#7 and CMS routines so any
9167 application which reads PKCS#7 or CMS data from untrusted sources is
9168 affected. SSL/TLS is not affected.
9170 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
9176 * Race condition handling PSK identify hint
9178 If PSK identity hints are received by a multi-threaded client then
9179 the values are wrongly updated in the parent SSL_CTX structure. This can
9180 result in a race condition potentially leading to a double free of the
9186 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
9188 * Malformed ECParameters causes infinite loop
9190 When processing an ECParameters structure OpenSSL enters an infinite loop
9191 if the curve specified is over a specially malformed binary polynomial
9194 This can be used to perform denial of service against any
9195 system which processes public keys, certificate requests or
9196 certificates. This includes TLS clients and TLS servers with
9197 client authentication enabled.
9199 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9204 * Exploitable out-of-bounds read in X509_cmp_time
9206 X509_cmp_time does not properly check the length of the ASN1_TIME
9207 string and can read a few bytes out of bounds. In addition,
9208 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9211 An attacker can use this to craft malformed certificates and CRLs of
9212 various sizes and potentially cause a segmentation fault, resulting in
9213 a DoS on applications that verify certificates or CRLs. TLS clients
9214 that verify CRLs are affected. TLS clients and servers with client
9215 authentication enabled may be affected if they use custom verification
9218 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9219 independently by Hanno Böck.
9224 * PKCS7 crash with missing EnvelopedContent
9226 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9227 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9228 with missing content and trigger a NULL pointer dereference on parsing.
9230 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9231 structures from untrusted sources are affected. OpenSSL clients and
9232 servers are not affected.
9234 This issue was reported to OpenSSL by Michal Zalewski (Google).
9239 * CMS verify infinite loop with unknown hash function
9241 When verifying a signedData message the CMS code can enter an infinite loop
9242 if presented with an unknown hash function OID. This can be used to perform
9243 denial of service against any system which verifies signedData messages using
9245 This issue was reported to OpenSSL by Johannes Bauer.
9250 * Race condition handling NewSessionTicket
9252 If a NewSessionTicket is received by a multi-threaded client when attempting to
9253 reuse a previous ticket then a race condition can occur potentially leading to
9254 a double free of the ticket data.
9259 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9261 * Segmentation fault in ASN1_TYPE_cmp fix
9263 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9264 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9265 certificate signature algorithm consistency this can be used to crash any
9266 certificate verification operation and exploited in a DoS attack. Any
9267 application which performs certificate verification is vulnerable including
9268 OpenSSL clients and servers which enable client authentication.
9273 * ASN.1 structure reuse memory corruption fix
9275 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9276 memory corruption via an invalid write. Such reuse is and has been
9277 strongly discouraged and is believed to be rare.
9279 Applications that parse structures containing CHOICE or ANY DEFINED BY
9280 components may be affected. Certificate parsing (d2i_X509 and related
9281 functions) are however not affected. OpenSSL clients and servers are
9287 * PKCS7 NULL pointer dereferences fix
9289 The PKCS#7 parsing code does not handle missing outer ContentInfo
9290 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9291 missing content and trigger a NULL pointer dereference on parsing.
9293 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9294 otherwise parse PKCS#7 structures from untrusted sources are
9295 affected. OpenSSL clients and servers are not affected.
9297 This issue was reported to OpenSSL by Michal Zalewski (Google).
9302 * DoS via reachable assert in SSLv2 servers fix
9304 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9305 servers that both support SSLv2 and enable export cipher suites by sending
9306 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9308 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9309 (OpenSSL development team).
9314 * Use After Free following d2i_ECPrivatekey error fix
9316 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9317 could cause a use after free condition. This, in turn, could cause a double
9318 free in several private key parsing functions (such as d2i_PrivateKey
9319 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9320 for applications that receive EC private keys from untrusted
9321 sources. This scenario is considered rare.
9323 This issue was discovered by the BoringSSL project and fixed in their
9329 * X509_to_X509_REQ NULL pointer deref fix
9331 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9332 the certificate key is invalid. This function is rarely used in practice.
9334 This issue was discovered by Brian Carpenter.
9339 * Removed the export ciphers from the DEFAULT ciphers
9343 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9345 * Build fixes for the Windows and OpenVMS platforms
9347 *Matt Caswell and Richard Levitte*
9349 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9351 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9352 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9353 dereference. This could lead to a Denial Of Service attack. Thanks to
9354 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9359 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9360 dtls1_buffer_record function under certain conditions. In particular this
9361 could occur if an attacker sent repeated DTLS records with the same
9362 sequence number but for the next epoch. The memory leak could be exploited
9363 by an attacker in a Denial of Service attack through memory exhaustion.
9364 Thanks to Chris Mueller for reporting this issue.
9369 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9370 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9371 method would be set to NULL which could later result in a NULL pointer
9372 dereference. Thanks to Frank Schmirler for reporting this issue.
9377 * Abort handshake if server key exchange message is omitted for ephemeral
9380 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9381 reporting this issue.
9386 * Remove non-export ephemeral RSA code on client and server. This code
9387 violated the TLS standard by allowing the use of temporary RSA keys in
9388 non-export ciphersuites and could be used by a server to effectively
9389 downgrade the RSA key length used to a value smaller than the server
9390 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9391 INRIA or reporting this issue.
9396 * Fixed issue where DH client certificates are accepted without verification.
9397 An OpenSSL server will accept a DH certificate for client authentication
9398 without the certificate verify message. This effectively allows a client to
9399 authenticate without the use of a private key. This only affects servers
9400 which trust a client certificate authority which issues certificates
9401 containing DH keys: these are extremely rare and hardly ever encountered.
9402 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9408 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9409 results on some platforms, including x86_64. This bug occurs at random
9410 with a very low probability, and is not known to be exploitable in any
9411 way, though its exact impact is difficult to determine. Thanks to Pieter
9412 Wuille (Blockstream) who reported this issue and also suggested an initial
9413 fix. Further analysis was conducted by the OpenSSL development team and
9414 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9415 the OpenSSL core team.
9420 * Fix various certificate fingerprint issues.
9422 By using non-DER or invalid encodings outside the signed portion of a
9423 certificate the fingerprint can be changed without breaking the signature.
9424 Although no details of the signed portion of the certificate can be changed
9425 this can cause problems with some applications: e.g. those using the
9426 certificate fingerprint for blacklists.
9428 1. Reject signatures with non zero unused bits.
9430 If the BIT STRING containing the signature has non zero unused bits reject
9431 the signature. All current signature algorithms require zero unused bits.
9433 2. Check certificate algorithm consistency.
9435 Check the AlgorithmIdentifier inside TBS matches the one in the
9436 certificate signature. NB: this will result in signature failure
9437 errors for some broken certificates.
9439 Thanks to Konrad Kraszewski from Google for reporting this issue.
9441 3. Check DSA/ECDSA signatures use DER.
9443 Re-encode DSA/ECDSA signatures and compare with the original received
9444 signature. Return an error if there is a mismatch.
9446 This will reject various cases including garbage after signature
9447 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9448 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9449 (negative or with leading zeroes).
9451 Further analysis was conducted and fixes were developed by Stephen Henson
9452 of the OpenSSL core team.
9458 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9460 * Session Ticket Memory Leak.
9462 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9463 integrity of that ticket is first verified. In the event of a session
9464 ticket integrity check failing, OpenSSL will fail to free memory
9465 causing a memory leak. By sending a large number of invalid session
9466 tickets an attacker could exploit this issue in a Denial Of Service
9472 * Build option no-ssl3 is incomplete.
9474 When OpenSSL is configured with "no-ssl3" as a build option, servers
9475 could accept and complete an SSL 3.0 handshake, and clients could be
9476 configured to send them.
9479 *Akamai and the OpenSSL team*
9481 * Add support for TLS_FALLBACK_SCSV.
9482 Client applications doing fallback retries should call
9483 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9486 *Adam Langley, Bodo Moeller*
9488 * Add additional DigestInfo checks.
9490 Re-encode DigestInto in DER and check against the original when
9491 verifying RSA signature: this will reject any improperly encoded
9492 DigestInfo structures.
9494 Note: this is a precautionary measure and no attacks are currently known.
9498 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9500 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9501 to a denial of service attack. A malicious server can crash the client
9502 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9503 ciphersuite and sending carefully crafted handshake messages.
9505 Thanks to Felix Gröbert (Google) for discovering and researching this
9511 * By sending carefully crafted DTLS packets an attacker could cause openssl
9512 to leak memory. This can be exploited through a Denial of Service attack.
9513 Thanks to Adam Langley for discovering and researching this issue.
9518 * An attacker can force openssl to consume large amounts of memory whilst
9519 processing DTLS handshake messages. This can be exploited through a
9520 Denial of Service attack.
9521 Thanks to Adam Langley for discovering and researching this issue.
9526 * An attacker can force an error condition which causes openssl to crash
9527 whilst processing DTLS packets due to memory being freed twice. This
9528 can be exploited through a Denial of Service attack.
9529 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9535 * If a multithreaded client connects to a malicious server using a resumed
9536 session and the server sends an ec point format extension it could write
9537 up to 255 bytes to freed memory.
9539 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9545 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9546 X509_name_oneline, X509_name_print_ex et al. to leak some information
9547 from the stack. Applications may be affected if they echo pretty printing
9548 output to the attacker.
9550 Thanks to Ivan Fratric (Google) for discovering this issue.
9553 *Emilia Käsper, and Steve Henson*
9555 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9556 for corner cases. (Certain input points at infinity could lead to
9557 bogus results, with non-infinity inputs mapped to infinity too.)
9561 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9563 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9564 handshake can force the use of weak keying material in OpenSSL
9565 SSL/TLS clients and servers.
9567 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9568 researching this issue. ([CVE-2014-0224])
9570 *KIKUCHI Masashi, Steve Henson*
9572 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9573 OpenSSL DTLS client the code can be made to recurse eventually crashing
9576 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9579 *Imre Rad, Steve Henson*
9581 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9582 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9583 client or server. This is potentially exploitable to run arbitrary
9584 code on a vulnerable client or server.
9586 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9588 *Jüri Aedla, Steve Henson*
9590 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9591 are subject to a denial of service attack.
9593 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9594 this issue. ([CVE-2014-3470])
9596 *Felix Gröbert, Ivan Fratric, Steve Henson*
9598 * Harmonize version and its documentation. -f flag is used to display
9601 *mancha <mancha1@zoho.com>*
9603 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9604 in i2d_ECPrivateKey.
9606 *mancha <mancha1@zoho.com>*
9608 * Fix some double frees. These are not thought to be exploitable.
9610 *mancha <mancha1@zoho.com>*
9612 * Fix for the attack described in the paper "Recovering OpenSSL
9613 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9614 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9615 <http://eprint.iacr.org/2014/140>
9617 Thanks to Yuval Yarom and Naomi Benger for discovering this
9618 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9620 *Yuval Yarom and Naomi Benger*
9622 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9624 * Keep original DTLS digest and encryption contexts in retransmission
9625 structures so we can use the previous session parameters if they need
9626 to be resent. ([CVE-2013-6450])
9630 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9631 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9632 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9633 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9634 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9635 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9637 *Rob Stradling, Adam Langley*
9639 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9641 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9643 This addresses the flaw in CBC record processing discovered by
9644 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9645 at: <http://www.isg.rhul.ac.uk/tls/>
9647 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9648 Security Group at Royal Holloway, University of London
9649 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9650 Emilia Käsper for the initial patch.
9653 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9655 * Return an error when checking OCSP signatures when key is NULL.
9656 This fixes a DoS attack. ([CVE-2013-0166])
9660 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9661 the right response is stapled. Also change SSL_get_certificate()
9662 so it returns the certificate actually sent.
9663 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9664 (This is a backport)
9666 *Rob Stradling <rob.stradling@comodo.com>*
9668 * Fix possible deadlock when decoding public keys.
9672 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9674 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9677 * Sanity check record length before skipping explicit IV in DTLS
9680 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9681 fuzzing as a service testing platform.
9686 * Initialise tkeylen properly when encrypting CMS messages.
9687 Thanks to Solar Designer of Openwall for reporting this issue.
9691 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9693 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9694 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9695 in CRYPTO_realloc_clean.
9697 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9698 issue and to Adam Langley <agl@chromium.org> for fixing it.
9701 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9703 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9705 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9706 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9707 content decryption and always return the same error. Note: this attack
9708 needs on average 2^20 messages so it only affects automated senders. The
9709 old behaviour can be re-enabled in the CMS code by setting the
9710 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9711 an MMA defence is not necessary.
9712 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9713 this issue. ([CVE-2012-0884])
9717 * Fix CVE-2011-4619: make sure we really are receiving a
9718 client hello before rejecting multiple SGC restarts. Thanks to
9719 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9723 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9725 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9726 Thanks to Antonio Martin, Enterprise Secure Access Research and
9727 Development, Cisco Systems, Inc. for discovering this bug and
9728 preparing a fix. ([CVE-2012-0050])
9732 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9734 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9735 of the Vaudenay padding oracle attack on CBC mode encryption
9736 which enables an efficient plaintext recovery attack against
9737 the OpenSSL implementation of DTLS. Their attack exploits timing
9738 differences arising during decryption processing. A research
9739 paper describing this attack can be found at:
9740 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9741 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9742 Security Group at Royal Holloway, University of London
9743 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9744 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9745 for preparing the fix. ([CVE-2011-4108])
9747 *Robin Seggelmann, Michael Tuexen*
9749 * Clear bytes used for block padding of SSL 3.0 records.
9752 *Adam Langley (Google)*
9754 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9755 Kadianakis <desnacked@gmail.com> for discovering this issue and
9756 Adam Langley for preparing the fix. ([CVE-2011-4619])
9758 *Adam Langley (Google)*
9760 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9762 *Andrey Kulikov <amdeich@gmail.com>*
9764 * Prevent malformed RFC3779 data triggering an assertion failure.
9765 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9766 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9768 *Rob Austein <sra@hactrn.net>*
9770 * Improved PRNG seeding for VOS.
9772 *Paul Green <Paul.Green@stratus.com>*
9774 * Fix ssl_ciph.c set-up race.
9776 *Adam Langley (Google)*
9778 * Fix spurious failures in ecdsatest.c.
9780 *Emilia Käsper (Google)*
9782 * Fix the BIO_f_buffer() implementation (which was mixing different
9783 interpretations of the `..._len` fields).
9785 *Adam Langley (Google)*
9787 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9788 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9789 threads won't reuse the same blinding coefficients.
9791 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9792 lock to call BN_BLINDING_invert_ex, and avoids one use of
9793 BN_BLINDING_update for each BN_BLINDING structure (previously,
9794 the last update always remained unused).
9796 *Emilia Käsper (Google)*
9798 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9800 *Bob Buckholz (Google)*
9802 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9804 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9805 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9807 *Kaspar Brand <ossl@velox.ch>*
9809 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9810 for multi-threaded use of ECDH. ([CVE-2011-3210])
9812 *Adam Langley (Google)*
9814 * Fix x509_name_ex_d2i memory leak on bad inputs.
9818 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9819 signature public key algorithm by using OID xref utilities instead.
9820 Before this you could only use some ECC ciphersuites with SHA1 only.
9824 * Add protection against ECDSA timing attacks as mentioned in the paper
9825 by Billy Bob Brumley and Nicola Tuveri, see:
9826 <http://eprint.iacr.org/2011/232.pdf>
9828 *Billy Bob Brumley and Nicola Tuveri*
9830 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9832 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9834 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9836 * Fix bug in string printing code: if *any* escaping is enabled we must
9837 escape the escape character (backslash) or the resulting string is
9842 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9844 * Disable code workaround for ancient and obsolete Netscape browsers
9845 and servers: an attacker can use it in a ciphersuite downgrade attack.
9846 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9850 * Fixed J-PAKE implementation error, originally discovered by
9851 Sebastien Martini, further info and confirmation from Stefan
9852 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9856 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9858 * Fix extension code to avoid race conditions which can result in a buffer
9859 overrun vulnerability: resumed sessions must not be modified as they can
9860 be shared by multiple threads. CVE-2010-3864
9864 * Fix WIN32 build system to correctly link an ENGINE directory into
9869 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9871 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9874 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9876 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9878 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9879 context. The operation can be customised via the ctrl mechanism in
9880 case ENGINEs want to include additional functionality.
9884 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9888 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9889 output hashes compatible with older versions of OpenSSL.
9891 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9893 * Fix compression algorithm handling: if resuming a session use the
9894 compression algorithm of the resumed session instead of determining
9895 it from client hello again. Don't allow server to change algorithm.
9899 * Add load_crls() function to commands tidying load_certs() too. Add option
9900 to verify utility to allow additional CRLs to be included.
9904 * Update OCSP request code to permit adding custom headers to the request:
9905 some responders need this.
9909 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9912 *Julia Lawall <julia@diku.dk>*
9914 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9915 needlessly dereferenced structures, used obsolete functions and
9916 didn't handle all updated verify codes correctly.
9920 * Disable MD2 in the default configuration.
9924 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9925 indicate the initial BIO being pushed or popped. This makes it possible
9926 to determine whether the BIO is the one explicitly called or as a result
9927 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9928 it handles reference counts correctly and doesn't zero out the I/O bio
9929 when it is not being explicitly popped. WARNING: applications which
9930 included workarounds for the old buggy behaviour will need to be modified
9931 or they could free up already freed BIOs.
9935 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9936 renaming to all platforms (within the 0.9.8 branch, this was
9937 done conditionally on Netware platforms to avoid a name clash).
9939 *Guenter <lists@gknw.net>*
9941 * Add ECDHE and PSK support to DTLS.
9943 *Michael Tuexen <tuexen@fh-muenster.de>*
9945 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9950 * Add "missing" function EVP_MD_flags() (without this the only way to
9951 retrieve a digest flags is by accessing the structure directly. Update
9952 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9953 or cipher is registered as in the "from" argument. Print out all
9954 registered digests in the dgst usage message instead of manually
9955 attempting to work them out.
9959 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9960 this allows the use of compression and extensions. Change default cipher
9961 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9962 by default unless an application cipher string requests it.
9966 * Alter match criteria in PKCS12_parse(). It used to try to use local
9967 key ids to find matching certificates and keys but some PKCS#12 files
9968 don't follow the (somewhat unwritten) rules and this strategy fails.
9969 Now just gather all certificates together and the first private key
9970 then look for the first certificate that matches the key.
9974 * Support use of registered digest and cipher names for dgst and cipher
9975 commands instead of having to add each one as a special case. So now
9982 openssl dgst -sha256 foo
9984 and this works for ENGINE based algorithms too.
9988 * Update Gost ENGINE to support parameter files.
9990 *Victor B. Wagner <vitus@cryptocom.ru>*
9992 * Support GeneralizedTime in ca utility.
9994 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9996 * Enhance the hash format used for certificate directory links. The new
9997 form uses the canonical encoding (meaning equivalent names will work
9998 even if they aren't identical) and uses SHA1 instead of MD5. This form
9999 is incompatible with the older format and as a result c_rehash should
10000 be used to rebuild symbolic links.
10004 * Make PKCS#8 the default write format for private keys, replacing the
10005 traditional format. This form is standardised, more secure and doesn't
10006 include an implicit MD5 dependency.
10010 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
10011 committed to OpenSSL should pass this lot as a minimum.
10015 * Add session ticket override functionality for use by EAP-FAST.
10017 *Jouni Malinen <j@w1.fi>*
10019 * Modify HMAC functions to return a value. Since these can be implemented
10020 in an ENGINE errors can occur.
10024 * Type-checked OBJ_bsearch_ex.
10028 * Type-checked OBJ_bsearch. Also some constification necessitated
10029 by type-checking. Still to come: TXT_DB, bsearch(?),
10030 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
10035 * New function OPENSSL_gmtime_adj() to add a specific number of days and
10036 seconds to a tm structure directly, instead of going through OS
10037 specific date routines. This avoids any issues with OS routines such
10038 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
10039 and X509_time_adj_ex() to cover the extended range. The existing
10040 X509_time_adj() is still usable and will no longer have any date issues.
10044 * Delta CRL support. New use deltas option which will attempt to locate
10045 and search any appropriate delta CRLs available.
10047 This work was sponsored by Google.
10051 * Support for CRLs partitioned by reason code. Reorganise CRL processing
10052 code and add additional score elements. Validate alternate CRL paths
10053 as part of the CRL checking and indicate a new error "CRL path validation
10054 error" in this case. Applications wanting additional details can use
10055 the verify callback and check the new "parent" field. If this is not
10056 NULL CRL path validation is taking place. Existing applications won't
10057 see this because it requires extended CRL support which is off by
10060 This work was sponsored by Google.
10064 * Support for freshest CRL extension.
10066 This work was sponsored by Google.
10070 * Initial indirect CRL support. Currently only supported in the CRLs
10071 passed directly and not via lookup. Process certificate issuer
10072 CRL entry extension and lookup CRL entries by bother issuer name
10073 and serial number. Check and process CRL issuer entry in IDP extension.
10075 This work was sponsored by Google.
10079 * Add support for distinct certificate and CRL paths. The CRL issuer
10080 certificate is validated separately in this case. Only enabled if
10081 an extended CRL support flag is set: this flag will enable additional
10082 CRL functionality in future.
10084 This work was sponsored by Google.
10088 * Add support for policy mappings extension.
10090 This work was sponsored by Google.
10094 * Fixes to pathlength constraint, self issued certificate handling,
10095 policy processing to align with RFC3280 and PKITS tests.
10097 This work was sponsored by Google.
10101 * Support for name constraints certificate extension. DN, email, DNS
10102 and URI types are currently supported.
10104 This work was sponsored by Google.
10108 * To cater for systems that provide a pointer-based thread ID rather
10109 than numeric, deprecate the current numeric thread ID mechanism and
10110 replace it with a structure and associated callback type. This
10111 mechanism allows a numeric "hash" to be extracted from a thread ID in
10112 either case, and on platforms where pointers are larger than 'long',
10113 mixing is done to help ensure the numeric 'hash' is usable even if it
10114 can't be guaranteed unique. The default mechanism is to use "&errno"
10115 as a pointer-based thread ID to distinguish between threads.
10117 Applications that want to provide their own thread IDs should now use
10118 CRYPTO_THREADID_set_callback() to register a callback that will call
10119 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
10121 Note that ERR_remove_state() is now deprecated, because it is tied
10122 to the assumption that thread IDs are numeric. ERR_remove_state(0)
10123 to free the current thread's error state should be replaced by
10124 ERR_remove_thread_state(NULL).
10126 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
10127 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
10128 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
10129 application was previously providing a numeric thread callback that
10130 was inappropriate for distinguishing threads, then uniqueness might
10131 have been obtained with &errno that happened immediately in the
10132 intermediate development versions of OpenSSL; this is no longer the
10133 case, the numeric thread callback will now override the automatic use
10136 *Geoff Thorpe, with help from Bodo Moeller*
10138 * Initial support for different CRL issuing certificates. This covers a
10139 simple case where the self issued certificates in the chain exist and
10140 the real CRL issuer is higher in the existing chain.
10142 This work was sponsored by Google.
10146 * Removed effectively defunct crypto/store from the build.
10150 * Revamp of STACK to provide stronger type-checking. Still to come:
10151 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
10152 ASN1_STRING, CONF_VALUE.
10156 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
10157 RAM on SSL connections. This option can save about 34k per idle SSL.
10161 * Revamp of LHASH to provide stronger type-checking. Still to come:
10162 STACK, TXT_DB, bsearch, qsort.
10166 * Initial support for Cryptographic Message Syntax (aka CMS) based
10167 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
10168 support for data, signedData, compressedData, digestedData and
10169 encryptedData, envelopedData types included. Scripts to check against
10170 RFC4134 examples draft and interop and consistency checks of many
10171 content types and variants.
10175 * Add options to enc utility to support use of zlib compression BIO.
10179 * Extend mk1mf to support importing of options and assembly language
10180 files from Configure script, currently only included in VC-WIN32.
10181 The assembly language rules can now optionally generate the source
10182 files from the associated perl scripts.
10186 * Implement remaining functionality needed to support GOST ciphersuites.
10187 Interop testing has been performed using CryptoPro implementations.
10189 *Victor B. Wagner <vitus@cryptocom.ru>*
10191 * s390x assembler pack.
10195 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10200 * Implement Opaque PRF Input TLS extension as specified in
10201 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10202 official specification yet and no extension type assignment by
10203 IANA exists, this extension (for now) will have to be explicitly
10204 enabled when building OpenSSL by providing the extension number
10205 to use. For example, specify an option
10207 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10209 to the "config" or "Configure" script to enable the extension,
10210 assuming extension number 0x9527 (which is a completely arbitrary
10211 and unofficial assignment based on the MD5 hash of the Internet
10212 Draft). Note that by doing so, you potentially lose
10213 interoperability with other TLS implementations since these might
10214 be using the same extension number for other purposes.
10216 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10217 opaque PRF input value to use in the handshake. This will create
10218 an internal copy of the length-'len' string at 'src', and will
10219 return non-zero for success.
10221 To get more control and flexibility, provide a callback function
10224 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10225 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10229 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10232 Callback function 'cb' will be called in handshakes, and is
10233 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10234 Argument 'arg' is for application purposes (the value as given to
10235 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10236 be provided to the callback function). The callback function
10237 has to return non-zero to report success: usually 1 to use opaque
10238 PRF input just if possible, or 2 to enforce use of the opaque PRF
10239 input. In the latter case, the library will abort the handshake
10240 if opaque PRF input is not successfully negotiated.
10242 Arguments 'peerinput' and 'len' given to the callback function
10243 will always be NULL and 0 in the case of a client. A server will
10244 see the client's opaque PRF input through these variables if
10245 available (NULL and 0 otherwise). Note that if the server
10246 provides an opaque PRF input, the length must be the same as the
10247 length of the client's opaque PRF input.
10249 Note that the callback function will only be called when creating
10250 a new session (session resumption can resume whatever was
10251 previously negotiated), and will not be called in SSL 2.0
10252 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10253 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10254 for applications that need to enforce opaque PRF input.
10258 * Update ssl code to support digests other than SHA1+MD5 for handshake
10261 *Victor B. Wagner <vitus@cryptocom.ru>*
10263 * Add RFC4507 support to OpenSSL. This includes the corrections in
10264 RFC4507bis. The encrypted ticket format is an encrypted encoded
10265 SSL_SESSION structure, that way new session features are automatically
10268 If a client application caches session in an SSL_SESSION structure
10269 support is transparent because tickets are now stored in the encoded
10272 The SSL_CTX structure automatically generates keys for ticket
10273 protection in servers so again support should be possible
10274 with no application modification.
10276 If a client or server wishes to disable RFC4507 support then the option
10277 SSL_OP_NO_TICKET can be set.
10279 Add a TLS extension debugging callback to allow the contents of any client
10280 or server extensions to be examined.
10282 This work was sponsored by Google.
10286 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10287 OpenSSL should now compile cleanly on gcc 4.2
10289 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10291 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10292 support including streaming MAC support: this is required for GOST
10293 ciphersuite support.
10295 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10297 * Add option -stream to use PKCS#7 streaming in smime utility. New
10298 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10299 to output in BER and PEM format.
10303 * Experimental support for use of HMAC via EVP_PKEY interface. This
10304 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10305 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10306 ENGINE support for HMAC keys which are unextractable. New -mac and
10307 -macopt options to dgst utility.
10311 * New option -sigopt to dgst utility. Update dgst to use
10312 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10313 alternative signing parameters such as X9.31 or PSS in the dgst
10318 * Change ssl_cipher_apply_rule(), the internal function that does
10319 the work each time a ciphersuite string requests enabling
10320 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10321 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10322 the order of disabled ciphersuites such that those ciphersuites
10323 that most recently went from enabled to disabled not only stay
10324 in order with respect to each other, but also have higher priority
10325 than other disabled ciphersuites the next time ciphersuites are
10328 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10329 the same ciphersuites as with "HIGH" alone, but in a specific
10330 order where the PSK ciphersuites come first (since they are the
10331 most recently disabled ciphersuites when "HIGH" is parsed).
10333 Also, change ssl_create_cipher_list() (using this new
10334 functionality) such that between otherwise identical
10335 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10340 * Change ssl_create_cipher_list() so that it automatically
10341 arranges the ciphersuites in reasonable order before starting
10342 to process the rule string. Thus, the definition for "DEFAULT"
10343 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10344 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10345 This makes it much easier to arrive at a reasonable default order
10346 in applications for which anonymous ciphers are OK (meaning
10347 that you can't actually use DEFAULT).
10349 *Bodo Moeller; suggested by Victor Duchovni*
10351 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10352 processing) into multiple integers instead of setting
10353 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10354 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10355 (These masks as well as the individual bit definitions are hidden
10356 away into the non-exported interface ssl/ssl_locl.h, so this
10357 change to the definition of the SSL_CIPHER structure shouldn't
10358 affect applications.) This give us more bits for each of these
10359 categories, so there is no longer a need to coagulate AES128 and
10360 AES256 into a single algorithm bit, and to coagulate Camellia128
10361 and Camellia256 into a single algorithm bit, which has led to all
10364 Thus, among other things, the kludge introduced in 0.9.7m and
10365 0.9.8e for masking out AES256 independently of AES128 or masking
10366 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10368 With the change, we also introduce new ciphersuite aliases that
10369 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10374 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10375 Use the leftmost N bytes of the signature input if the input is
10376 larger than the prime q (with N being the size in bytes of q).
10380 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10381 it yet and it is largely untested.
10385 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10389 * Initial incomplete changes to avoid need for function casts in OpenSSL
10390 some compilers (gcc 4.2 and later) reject their use. Safestack is
10391 reimplemented. Update ASN1 to avoid use of legacy functions.
10395 * Win32/64 targets are linked with Winsock2.
10399 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10400 to external functions. This can be used to increase CRL handling
10401 efficiency especially when CRLs are very large by (for example) storing
10402 the CRL revoked certificates in a database.
10406 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10407 new CRLs added to a directory can be used. New command line option
10408 -verify_return_error to s_client and s_server. This causes real errors
10409 to be returned by the verify callback instead of carrying on no matter
10410 what. This reflects the way a "real world" verify callback would behave.
10414 * GOST engine, supporting several GOST algorithms and public key formats.
10415 Kindly donated by Cryptocom.
10419 * Partial support for Issuing Distribution Point CRL extension. CRLs
10420 partitioned by DP are handled but no indirect CRL or reason partitioning
10421 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10422 selected via a scoring technique which handles IDP and AKID in CRLs.
10426 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10427 will ultimately be used for all verify operations: this will remove the
10428 X509_STORE dependency on certificate verification and allow alternative
10429 lookup methods. X509_STORE based implementations of these two callbacks.
10433 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10434 Modify get_crl() to find a valid (unexpired) CRL if possible.
10438 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10439 this would be called X509_CRL_cmp() but that name is already used by
10440 a function that just compares CRL issuer names. Cache several CRL
10441 extensions in X509_CRL structure and cache CRLDP in X509.
10445 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10446 this maps equivalent X509_NAME structures into a consistent structure.
10447 Name comparison can then be performed rapidly using memcmp().
10451 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10456 * Allow digests to supply their own micalg string for S/MIME type using
10457 the ctrl EVP_MD_CTRL_MICALG.
10461 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10462 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10463 ctrl. It can then customise the structure before and/or after signing
10468 * New function OBJ_add_sigid() to allow application defined signature OIDs
10469 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10470 to free up any added signature OIDs.
10474 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10475 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10476 digest and cipher tables. New options added to openssl utility:
10477 list-message-digest-algorithms and list-cipher-algorithms.
10481 * Change the array representation of binary polynomials: the list
10482 of degrees of non-zero coefficients is now terminated with -1.
10483 Previously it was terminated with 0, which was also part of the
10484 value; thus, the array representation was not applicable to
10485 polynomials where t^0 has coefficient zero. This change makes
10486 the array representation useful in a more general context.
10490 * Various modifications and fixes to SSL/TLS cipher string
10491 handling. For ECC, the code now distinguishes between fixed ECDH
10492 with RSA certificates on the one hand and with ECDSA certificates
10493 on the other hand, since these are separate ciphersuites. The
10494 unused code for Fortezza ciphersuites has been removed.
10496 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10497 (not "ECDHE"). For consistency with the code for DH
10498 certificates, use of ECDH certificates is now considered ECDH
10499 authentication, not RSA or ECDSA authentication (the latter is
10500 merely the CA's signing algorithm and not actively used in the
10503 The temporary ciphersuite alias "ECCdraft" is no longer
10504 available, and ECC ciphersuites are no longer excluded from "ALL"
10505 and "DEFAULT". The following aliases now exist for RFC 4492
10506 ciphersuites, most of these by analogy with the DH case:
10508 kECDHr - ECDH cert, signed with RSA
10509 kECDHe - ECDH cert, signed with ECDSA
10510 kECDH - ECDH cert (signed with either RSA or ECDSA)
10511 kEECDH - ephemeral ECDH
10512 ECDH - ECDH cert or ephemeral ECDH
10515 aECDSA - ECDSA cert
10518 AECDH - anonymous ECDH
10519 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10523 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10524 Use correct micalg parameters depending on digest(s) in signed message.
10528 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10529 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10533 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10534 an engine to register a method. Add ENGINE lookups for methods and
10535 functional reference processing.
10539 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10540 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10545 * New -resign option to smime utility. This adds one or more signers
10546 to an existing PKCS#7 signedData structure. Also -md option to use an
10547 alternative message digest algorithm for signing.
10551 * Tidy up PKCS#7 routines and add new functions to make it easier to
10552 create PKCS7 structures containing multiple signers. Update smime
10553 application to support multiple signers.
10557 * New -macalg option to pkcs12 utility to allow setting of an alternative
10562 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10563 Reorganize PBE internals to lookup from a static table using NIDs,
10564 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10565 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10566 PRF which will be automatically used with PBES2.
10570 * Replace the algorithm specific calls to generate keys in "req" with the
10575 * Update PKCS#7 enveloped data routines to use new API. This is now
10576 supported by any public key method supporting the encrypt operation. A
10577 ctrl is added to allow the public key algorithm to examine or modify
10578 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10583 * Add a ctrl to asn1 method to allow a public key algorithm to express
10584 a default digest type to use. In most cases this will be SHA1 but some
10585 algorithms (such as GOST) need to specify an alternative digest. The
10586 return value indicates how strong the preference is 1 means optional and
10587 2 is mandatory (that is it is the only supported type). Modify
10588 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10589 use the default md. Update openssl utilities to use the default digest
10590 type for signing if it is not explicitly indicated.
10594 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10595 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10596 signing method from the key type. This effectively removes the link
10597 between digests and public key types.
10601 * Add an OID cross reference table and utility functions. Its purpose is to
10602 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10603 rsaEncryption. This will allow some of the algorithm specific hackery
10604 needed to use the correct OID to be removed.
10608 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10609 structures for PKCS7_sign(). They are now set up by the relevant public
10614 * Add provisional EC pkey method with support for ECDSA and ECDH.
10618 * Add support for key derivation (agreement) in the API, DH method and
10623 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10624 public and private key formats. As a side effect these add additional
10625 command line functionality not previously available: DSA signatures can be
10626 generated and verified using pkeyutl and DH key support and generation in
10633 *Oliver Tappe <zooey@hirschkaefer.de>*
10635 * New make target "install_html_docs" installs HTML renditions of the
10638 *Oliver Tappe <zooey@hirschkaefer.de>*
10640 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10641 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10642 support key and parameter generation and add initial key generation
10643 functionality for RSA.
10647 * Add functions for main EVP_PKEY_method operations. The undocumented
10648 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10649 `EVP_PKEY_{encrypt,decrypt}_old`.
10653 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10654 key API, doesn't do much yet.
10658 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10659 public key algorithms. New option to openssl utility:
10660 "list-public-key-algorithms" to print out info.
10664 * Implement the Supported Elliptic Curves Extension for
10665 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10669 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10670 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10674 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10675 utilities such as rsa, dsa, dsaparam etc except they process any key
10680 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10681 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10682 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10687 * Initial support for pluggable public key ASN1.
10688 De-spaghettify the public key ASN1 handling. Move public and private
10689 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10690 algorithm specific handling to a single module within the relevant
10691 algorithm directory. Add functions to allow (near) opaque processing
10692 of public and private key structures.
10696 * Implement the Supported Point Formats Extension for
10697 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10701 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10702 for the psk identity [hint] and the psk callback functions to the
10703 SSL_SESSION, SSL and SSL_CTX structure.
10706 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10710 SSL_CTX_use_psk_identity_hint
10711 SSL_get_psk_identity_hint
10712 SSL_get_psk_identity
10713 SSL_use_psk_identity_hint
10715 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10717 * Add RFC 3161 compliant time stamp request creation, response generation
10718 and response verification functionality.
10720 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10722 * Add initial support for TLS extensions, specifically for the server_name
10723 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10724 have new members for a hostname. The SSL data structure has an
10725 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10726 stored in that context to allow for session resumption, even after the
10727 SSL has been switched to a new SSL_CTX in reaction to a client's
10728 server_name extension.
10730 New functions (subject to change):
10732 SSL_get_servername()
10733 SSL_get_servername_type()
10736 New CTRL codes and macros (subject to change):
10738 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10739 - SSL_CTX_set_tlsext_servername_callback()
10740 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10741 - SSL_CTX_set_tlsext_servername_arg()
10742 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10744 openssl s_client has a new '-servername ...' option.
10746 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10747 '-key2 ...', '-servername_fatal' (subject to change). This allows
10748 testing the HostName extension for a specific single hostname ('-cert'
10749 and '-key' remain fallbacks for handshakes without HostName
10750 negotiation). If the unrecognized_name alert has to be sent, this by
10751 default is a warning; it becomes fatal with the '-servername_fatal'
10754 *Peter Sylvester, Remy Allais, Christophe Renou*
10756 * Whirlpool hash implementation is added.
10760 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10761 bn(64,32). Because of instruction set limitations it doesn't have
10762 any negative impact on performance. This was done mostly in order
10763 to make it possible to share assembler modules, such as bn_mul_mont
10764 implementations, between 32- and 64-bit builds without hassle.
10768 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10769 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10774 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10775 dedicated Montgomery multiplication procedure, is introduced.
10776 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10777 "64-bit" performance on certain 32-bit targets.
10781 * New option SSL_OP_NO_COMP to disable use of compression selectively
10782 in SSL structures. New SSL ctrl to set maximum send fragment size.
10783 Save memory by setting the I/O buffer sizes dynamically instead of
10784 using the maximum available value.
10788 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10789 in addition to the text details.
10793 * Very, very preliminary EXPERIMENTAL support for printing of general
10794 ASN1 structures. This currently produces rather ugly output and doesn't
10795 handle several customised structures at all.
10799 * Integrated support for PVK file format and some related formats such
10800 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10801 these in the 'rsa' and 'dsa' utilities.
10805 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10809 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10810 place for the (very old) "NETSCAPE" format certificates which are now
10811 handled using new ASN1 code equivalents.
10815 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10816 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10817 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10821 * Modify CRL distribution points extension code to print out previously
10822 unsupported fields. Enhance extension setting code to allow setting of
10827 * Add print and set support for Issuing Distribution Point CRL extension.
10831 * Change 'Configure' script to enable Camellia by default.
10838 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10840 * When rejecting SSL/TLS records due to an incorrect version number, never
10841 update s->server with a new major version number. As of
10842 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10843 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10844 the previous behavior could result in a read attempt at NULL when
10845 receiving specific incorrect SSL/TLS records once record payload
10846 protection is active. ([CVE-2010-0740])
10848 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10850 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10851 could be crashed if the relevant tables were not present (e.g. chrooted).
10853 *Tomas Hoger <thoger@redhat.com>*
10855 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10857 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10859 *Martin Olsson, Neel Mehta*
10861 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10862 accommodate for stack sorting, always a write lock!).
10866 * On some versions of WIN32 Heap32Next is very slow. This can cause
10867 excessive delays in the RAND_poll(): over a minute. As a workaround
10868 include a time check in the inner Heap32Next loop too.
10872 * The code that handled flushing of data in SSL/TLS originally used the
10873 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10874 the problem outlined in PR#1949. The fix suggested there however can
10875 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10876 of Apache). So instead simplify the code to flush unconditionally.
10877 This should be fine since flushing with no data to flush is a no op.
10881 * Handle TLS versions 2.0 and later properly and correctly use the
10882 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10883 off ancient servers have a habit of sticking around for a while...
10887 * Modify compression code so it frees up structures without using the
10888 ex_data callbacks. This works around a problem where some applications
10889 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10890 restarting) then use compression (e.g. SSL with compression) later.
10891 This results in significant per-connection memory leaks and
10892 has caused some security issues including CVE-2008-1678 and
10897 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10898 change when encrypting or decrypting.
10902 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10903 connect and renegotiate with servers which do not support RI.
10904 Until RI is more widely deployed this option is enabled by default.
10908 * Add "missing" ssl ctrls to clear options and mode.
10912 * If client attempts to renegotiate and doesn't support RI respond with
10913 a no_renegotiation alert as required by RFC5746. Some renegotiating
10914 TLS clients will continue a connection gracefully when they receive
10915 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10916 waiting for a server hello which it will never receive. Now we treat a
10917 received no_renegotiation alert as a fatal error. This is because
10918 applications requesting a renegotiation might well expect it to succeed
10919 and would have no code in place to handle the server denying it so the
10920 only safe thing to do is to terminate the connection.
10924 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10925 peer supports secure renegotiation and 0 otherwise. Print out peer
10926 renegotiation support in s_client/s_server.
10930 * Replace the highly broken and deprecated SPKAC certification method with
10931 the updated NID creation version. This should correctly handle UTF8.
10935 * Implement RFC5746. Re-enable renegotiation but require the extension
10936 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10937 turns out to be a bad idea. It has been replaced by
10938 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10939 SSL_CTX_set_options(). This is really not recommended unless you
10940 know what you are doing.
10942 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10944 * Fixes to stateless session resumption handling. Use initial_ctx when
10945 issuing and attempting to decrypt tickets in case it has changed during
10946 servername handling. Use a non-zero length session ID when attempting
10947 stateless session resumption: this makes it possible to determine if
10948 a resumption has occurred immediately after receiving server hello
10949 (several places in OpenSSL subtly assume this) instead of later in
10954 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10955 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10956 fixes for a few places where the return code is not checked
10959 *Julia Lawall <julia@diku.dk>*
10961 * Add --strict-warnings option to Configure script to include devteam
10962 warnings in other configurations.
10966 * Add support for --libdir option and LIBDIR variable in makefiles. This
10967 makes it possible to install openssl libraries in locations which
10968 have names other than "lib", for example "/usr/lib64" which some
10971 *Steve Henson, based on patch from Jeremy Utley*
10973 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10974 X690 8.9.12 and can produce some misleading textual output of OIDs.
10976 *Steve Henson, reported by Dan Kaminsky*
10978 * Delete MD2 from algorithm tables. This follows the recommendation in
10979 several standards that it is not used in new applications due to
10980 several cryptographic weaknesses. For binary compatibility reasons
10981 the MD2 API is still compiled in by default.
10985 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10990 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10991 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10994 *Guenter <lists@gknw.net>*
10996 * Fix the server certificate chain building code to use X509_verify_cert(),
10997 it used to have an ad-hoc builder which was unable to cope with anything
10998 other than a simple chain.
11000 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
11002 * Don't check self signed certificate signatures in X509_verify_cert()
11003 by default (a flag can override this): it just wastes time without
11004 adding any security. As a useful side effect self signed root CAs
11005 with non-FIPS digests are now usable in FIPS mode.
11009 * In dtls1_process_out_of_seq_message() the check if the current message
11010 is already buffered was missing. For every new message was memory
11011 allocated, allowing an attacker to perform an denial of service attack
11012 with sending out of seq handshake messages until there is no memory
11013 left. Additionally every future message was buffered, even if the
11014 sequence number made no sense and would be part of another handshake.
11015 So only messages with sequence numbers less than 10 in advance will be
11016 buffered. ([CVE-2009-1378])
11018 *Robin Seggelmann, discovered by Daniel Mentz*
11020 * Records are buffered if they arrive with a future epoch to be
11021 processed after finishing the corresponding handshake. There is
11022 currently no limitation to this buffer allowing an attacker to perform
11023 a DOS attack with sending records with future epochs until there is no
11024 memory left. This patch adds the pqueue_size() function to determine
11025 the size of a buffer and limits the record buffer to 100 entries.
11028 *Robin Seggelmann, discovered by Daniel Mentz*
11030 * Keep a copy of frag->msg_header.frag_len so it can be used after the
11031 parent structure is freed. ([CVE-2009-1379])
11035 * Handle non-blocking I/O properly in SSL_shutdown() call.
11037 *Darryl Miles <darryl-mailinglists@netbauds.net>*
11039 * Add `2.5.4.*` OIDs
11041 *Ilya O. <vrghost@gmail.com>*
11043 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
11045 * Disable renegotiation completely - this fixes a severe security
11046 problem ([CVE-2009-3555]) at the cost of breaking all
11047 renegotiation. Renegotiation can be re-enabled by setting
11048 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
11049 run-time. This is really not recommended unless you know what
11054 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
11056 * Don't set val to NULL when freeing up structures, it is freed up by
11057 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
11058 zeroing past the valid field. ([CVE-2009-0789])
11060 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
11062 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
11063 checked correctly. This would allow some invalid signed attributes to
11064 appear to verify correctly. ([CVE-2009-0591])
11066 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11068 * Reject UniversalString and BMPString types with invalid lengths. This
11069 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
11070 a legal length. ([CVE-2009-0590])
11074 * Set S/MIME signing as the default purpose rather than setting it
11075 unconditionally. This allows applications to override it at the store
11080 * Permit restricted recursion of ASN1 strings. This is needed in practice
11081 to handle some structures.
11085 * Improve efficiency of mem_gets: don't search whole buffer each time
11088 *Jeremy Shapiro <jnshapir@us.ibm.com>*
11090 * New -hex option for openssl rand.
11094 * Print out UTF8String and NumericString when parsing ASN1.
11098 * Support NumericString type for name components.
11102 * Allow CC in the environment to override the automatically chosen
11103 compiler. Note that nothing is done to ensure flags work with the
11108 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
11110 * Properly check EVP_VerifyFinal() and similar return values
11113 *Ben Laurie, Bodo Moeller, Google Security Team*
11115 * Enable TLS extensions by default.
11119 * Allow the CHIL engine to be loaded, whether the application is
11120 multithreaded or not. (This does not release the developer from the
11121 obligation to set up the dynamic locking callbacks.)
11123 *Sander Temme <sander@temme.net>*
11125 * Use correct exit code if there is an error in dgst command.
11127 *Steve Henson; problem pointed out by Roland Dirlewanger*
11129 * Tweak Configure so that you need to say "experimental-jpake" to enable
11130 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
11134 * Add experimental JPAKE support, including demo authentication in
11135 s_client and s_server.
11139 * Set the comparison function in v3_addr_canonize().
11141 *Rob Austein <sra@hactrn.net>*
11143 * Add support for XMPP STARTTLS in s_client.
11145 *Philip Paeps <philip@freebsd.org>*
11147 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
11148 to ensure that even with this option, only ciphersuites in the
11149 server's preference list will be accepted. (Note that the option
11150 applies only when resuming a session, so the earlier behavior was
11151 just about the algorithm choice for symmetric cryptography.)
11155 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
11157 * Fix NULL pointer dereference if a DTLS server received
11158 ChangeCipherSpec as first record ([CVE-2009-1386]).
11162 * Fix a state transition in s3_srvr.c and d1_srvr.c
11163 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
11165 *Nagendra Modadugu*
11167 * The fix in 0.9.8c that supposedly got rid of unsafe
11168 double-checked locking was incomplete for RSA blinding,
11169 addressing just one layer of what turns out to have been
11170 doubly unsafe triple-checked locking.
11172 So now fix this for real by retiring the MONT_HELPER macro
11173 in crypto/rsa/rsa_eay.c.
11175 *Bodo Moeller; problem pointed out by Marius Schilder*
11177 * Various precautionary measures:
11179 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
11181 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
11182 (NB: This would require knowledge of the secret session ticket key
11183 to exploit, in which case you'd be SOL either way.)
11185 - Change bn_nist.c so that it will properly handle input BIGNUMs
11186 outside the expected range.
11188 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11191 *Neel Mehta, Bodo Moeller*
11193 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11194 the load fails. Useful for distros.
11196 *Ben Laurie and the FreeBSD team*
11198 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11202 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11206 * Expand ENGINE to support engine supplied SSL client certificate functions.
11208 This work was sponsored by Logica.
11212 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11213 keystores. Support for SSL/TLS client authentication too.
11214 Not compiled unless enable-capieng specified to Configure.
11216 This work was sponsored by Logica.
11220 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11221 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11222 attribute creation routines such as certificate requests and PKCS#12
11227 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11229 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11230 handshake which could lead to a client crash as found using the
11231 Codenomicon TLS test suite ([CVE-2008-1672])
11233 *Steve Henson, Mark Cox*
11235 * Fix double free in TLS server name extensions which could lead to
11236 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11240 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11242 Clear the error queue to ensure that error entries left from
11243 older function calls do not interfere with the correct operation.
11245 *Lutz Jaenicke, Erik de Castro Lopo*
11247 * Remove root CA certificates of commercial CAs:
11249 The OpenSSL project does not recommend any specific CA and does not
11250 have any policy with respect to including or excluding any CA.
11251 Therefore, it does not make any sense to ship an arbitrary selection
11252 of root CA certificates with the OpenSSL software.
11256 * RSA OAEP patches to fix two separate invalid memory reads.
11257 The first one involves inputs when 'lzero' is greater than
11258 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11259 before the beginning of from). The second one involves inputs where
11260 the 'db' section contains nothing but zeroes (there is a one-byte
11261 invalid read after the end of 'db').
11263 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11265 * Partial backport from 0.9.9-dev:
11267 Introduce bn_mul_mont (dedicated Montgomery multiplication
11268 procedure) as a candidate for BIGNUM assembler implementation.
11269 While 0.9.9-dev uses assembler for various architectures, only
11270 x86_64 is available by default here in the 0.9.8 branch, and
11271 32-bit x86 is available through a compile-time setting.
11273 To try the 32-bit x86 assembler implementation, use Configure
11274 option "enable-montasm" (which exists only for this backport).
11276 As "enable-montasm" for 32-bit x86 disclaims code stability
11277 anyway, in this constellation we activate additional code
11278 backported from 0.9.9-dev for further performance improvements,
11279 namely BN_from_montgomery_word. (To enable this otherwise,
11280 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11282 *Andy Polyakov (backport partially by Bodo Moeller)*
11284 * Add TLS session ticket callback. This allows an application to set
11285 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11286 values. This is useful for key rollover for example where several key
11287 sets may exist with different names.
11291 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11292 This was broken until now in 0.9.8 releases, such that the only way
11293 a registered ENGINE could be used (assuming it initialises
11294 successfully on the host) was to explicitly set it as the default
11295 for the relevant algorithms. This is in contradiction with 0.9.7
11296 behaviour and the documentation. With this fix, when an ENGINE is
11297 registered into a given algorithm's table of implementations, the
11298 'uptodate' flag is reset so that auto-discovery will be used next
11299 time a new context for that algorithm attempts to select an
11302 *Ian Lister (tweaked by Geoff Thorpe)*
11304 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11305 implementation in the following ways:
11307 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11310 Lack of BER streaming support means one pass streaming processing is
11311 only supported if data is detached: setting the streaming flag is
11312 ignored for embedded content.
11314 CMS support is disabled by default and must be explicitly enabled
11315 with the enable-cms configuration option.
11319 * Update the GMP engine glue to do direct copies between BIGNUM and
11320 mpz_t when openssl and GMP use the same limb size. Otherwise the
11321 existing "conversion via a text string export" trick is still used.
11323 *Paul Sheer <paulsheer@gmail.com>*
11325 * Zlib compression BIO. This is a filter BIO which compressed and
11326 uncompresses any data passed through it.
11330 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11331 RFC3394 compatible AES key wrapping.
11335 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11336 sets string data without copying. X509_ALGOR_set0() and
11337 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11338 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11339 from an X509_ATTRIBUTE structure optionally checking it occurs only
11340 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11345 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11346 to get the expected BN_FLG_CONSTTIME behavior.
11348 *Bodo Moeller (Google)*
11352 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11353 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11354 - added some more tests to do_tests.pl
11355 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11356 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11357 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11358 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11359 - various changes to netware.pl to enable gcc-cross builds on Win32
11361 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11362 - various changes to fix missing prototype warnings
11363 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11364 - added AES, WHIRLPOOL and CPUID assembler code to build files
11365 - added missing AES assembler make rules to mk1mf.pl
11366 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11368 *Guenter Knauf <eflash@gmx.net>*
11370 * Implement certificate status request TLS extension defined in RFC3546.
11371 A client can set the appropriate parameters and receive the encoded
11372 OCSP response via a callback. A server can query the supplied parameters
11373 and set the encoded OCSP response in the callback. Add simplified examples
11374 to s_client and s_server.
11378 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11380 * Fix various bugs:
11381 + Binary incompatibility of ssl_ctx_st structure
11382 + DTLS interoperation with non-compliant servers
11383 + Don't call get_session_cb() without proposed session
11384 + Fix ia64 assembler code
11386 *Andy Polyakov, Steve Henson*
11388 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11390 * DTLS Handshake overhaul. There were longstanding issues with
11391 OpenSSL DTLS implementation, which were making it impossible for
11392 RFC 4347 compliant client to communicate with OpenSSL server.
11393 Unfortunately just fixing these incompatibilities would "cut off"
11394 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11395 server keeps tolerating non RFC compliant syntax. The opposite is
11396 not true, 0.9.8f client can not communicate with earlier server.
11397 This update even addresses CVE-2007-4995.
11401 * Changes to avoid need for function casts in OpenSSL: some compilers
11402 (gcc 4.2 and later) reject their use.
11403 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11406 * Add RFC4507 support to OpenSSL. This includes the corrections in
11407 RFC4507bis. The encrypted ticket format is an encrypted encoded
11408 SSL_SESSION structure, that way new session features are automatically
11411 If a client application caches session in an SSL_SESSION structure
11412 support is transparent because tickets are now stored in the encoded
11415 The SSL_CTX structure automatically generates keys for ticket
11416 protection in servers so again support should be possible
11417 with no application modification.
11419 If a client or server wishes to disable RFC4507 support then the option
11420 SSL_OP_NO_TICKET can be set.
11422 Add a TLS extension debugging callback to allow the contents of any client
11423 or server extensions to be examined.
11425 This work was sponsored by Google.
11429 * Add initial support for TLS extensions, specifically for the server_name
11430 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11431 have new members for a hostname. The SSL data structure has an
11432 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11433 stored in that context to allow for session resumption, even after the
11434 SSL has been switched to a new SSL_CTX in reaction to a client's
11435 server_name extension.
11437 New functions (subject to change):
11439 SSL_get_servername()
11440 SSL_get_servername_type()
11443 New CTRL codes and macros (subject to change):
11445 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11446 - SSL_CTX_set_tlsext_servername_callback()
11447 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11448 - SSL_CTX_set_tlsext_servername_arg()
11449 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11451 openssl s_client has a new '-servername ...' option.
11453 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11454 '-key2 ...', '-servername_fatal' (subject to change). This allows
11455 testing the HostName extension for a specific single hostname ('-cert'
11456 and '-key' remain fallbacks for handshakes without HostName
11457 negotiation). If the unrecognized_name alert has to be sent, this by
11458 default is a warning; it becomes fatal with the '-servername_fatal'
11461 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11463 * Add AES and SSE2 assembly language support to VC++ build.
11467 * Mitigate attack on final subtraction in Montgomery reduction.
11471 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11472 (which previously caused an internal error).
11476 * Squeeze another 10% out of IGE mode when in != out.
11480 * AES IGE mode speedup.
11482 *Dean Gaudet (Google)*
11484 * Add the Korean symmetric 128-bit cipher SEED (see
11485 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11486 add SEED ciphersuites from RFC 4162:
11488 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11489 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11490 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11491 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11493 To minimize changes between patchlevels in the OpenSSL 0.9.8
11494 series, SEED remains excluded from compilation unless OpenSSL
11495 is configured with 'enable-seed'.
11497 *KISA, Bodo Moeller*
11499 * Mitigate branch prediction attacks, which can be practical if a
11500 single processor is shared, allowing a spy process to extract
11501 information. For detailed background information, see
11502 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11503 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11504 and Necessary Software Countermeasures"). The core of the change
11505 are new versions BN_div_no_branch() and
11506 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11507 respectively, which are slower, but avoid the security-relevant
11508 conditional branches. These are automatically called by BN_div()
11509 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11510 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11511 remove a conditional branch.
11513 BN_FLG_CONSTTIME is the new name for the previous
11514 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11515 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11516 in the exponent causes BN_mod_exp_mont() to use the alternative
11517 implementation in BN_mod_exp_mont_consttime().) The old name
11518 remains as a deprecated alias.
11520 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11521 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11522 constant-time implementations for more than just exponentiation.
11523 Here too the old name is kept as a deprecated alias.
11525 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11526 the BN_BLINDING structure gets an independent copy of the
11527 modulus. This means that the previous `BIGNUM *m` argument to
11528 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11529 essentially becomes `const BIGNUM *m`, although we can't actually
11530 change this in the header file before 0.9.9. It allows
11531 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11532 enable BN_FLG_CONSTTIME.
11534 *Matthew D Wood (Intel Corp)*
11536 * In the SSL/TLS server implementation, be strict about session ID
11537 context matching (which matters if an application uses a single
11538 external cache for different purposes). Previously,
11539 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11540 set. This did ensure strict client verification, but meant that,
11541 with applications using a single external cache for quite
11542 different requirements, clients could circumvent ciphersuite
11543 restrictions for a given session ID context by starting a session
11544 in a different context.
11548 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11549 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11550 authentication-only ciphersuites.
11554 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11555 not complete and could lead to a possible single byte overflow
11556 ([CVE-2007-5135]) [Ben Laurie]
11558 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11560 * Since AES128 and AES256 (and similarly Camellia128 and
11561 Camellia256) share a single mask bit in the logic of
11562 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11563 kludge to work properly if AES128 is available and AES256 isn't
11564 (or if Camellia128 is available and Camellia256 isn't).
11568 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11569 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11570 When a point or a seed is encoded in a BIT STRING, we need to
11571 prevent the removal of trailing zero bits to get the proper DER
11572 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11573 of a NamedBitList, for which trailing 0 bits need to be removed.)
11577 * Have SSL/TLS server implementation tolerate "mismatched" record
11578 protocol version while receiving ClientHello even if the
11579 ClientHello is fragmented. (The server can't insist on the
11580 particular protocol version it has chosen before the ServerHello
11581 message has informed the client about his choice.)
11585 * Add RFC 3779 support.
11587 *Rob Austein for ARIN, Ben Laurie*
11589 * Load error codes if they are not already present instead of using a
11590 static variable. This allows them to be cleanly unloaded and reloaded.
11591 Improve header file function name parsing.
11595 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11596 or CAPABILITY handshake as required by RFCs.
11598 *Goetz Babin-Ebell*
11600 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11602 * Introduce limits to prevent malicious keys being able to
11603 cause a denial of service. ([CVE-2006-2940])
11605 *Steve Henson, Bodo Moeller*
11607 * Fix ASN.1 parsing of certain invalid structures that can result
11608 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11610 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11611 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11613 * Fix SSL client code which could crash if connecting to a
11614 malicious SSLv2 server. ([CVE-2006-4343])
11616 *Tavis Ormandy and Will Drewry, Google Security Team*
11618 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11619 match only those. Before that, "AES256-SHA" would be interpreted
11620 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11621 the same strength classification in 0.9.7h) as we currently only
11622 have a single AES bit in the ciphersuite description bitmap.
11623 That change, however, also applied to ciphersuite strings such as
11624 "RC4-MD5" that intentionally matched multiple ciphersuites --
11625 namely, SSL 2.0 ciphersuites in addition to the more common ones
11626 from SSL 3.0/TLS 1.0.
11628 So we change the selection algorithm again: Naming an explicit
11629 ciphersuite selects this one ciphersuite, and any other similar
11630 ciphersuite (same bitmap) from *other* protocol versions.
11631 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11632 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11634 Since SSL 2.0 does not have any ciphersuites for which the
11635 128/256 bit distinction would be relevant, this works for now.
11636 The proper fix will be to use different bits for AES128 and
11637 AES256, which would have avoided the problems from the beginning;
11638 however, bits are scarce, so we can only do this in a new release
11639 (not just a patchlevel) when we can change the SSL_CIPHER
11640 definition to split the single 'unsigned long mask' bitmap into
11641 multiple values to extend the available space.
11645 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11647 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11648 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11650 * Add AES IGE and biIGE modes.
11654 * Change the Unix randomness entropy gathering to use poll() when
11655 possible instead of select(), since the latter has some
11656 undesirable limitations.
11658 *Darryl Miles via Richard Levitte and Bodo Moeller*
11660 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11661 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11662 cannot be implicitly activated as part of, e.g., the "AES" alias.
11663 However, please upgrade to OpenSSL 0.9.9[-dev] for
11664 non-experimental use of the ECC ciphersuites to get TLS extension
11665 support, which is required for curve and point format negotiation
11666 to avoid potential handshake problems.
11670 * Disable rogue ciphersuites:
11672 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11673 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11674 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11676 The latter two were purportedly from
11677 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11680 Also deactivate the remaining ciphersuites from
11681 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11682 unofficial, and the ID has long expired.
11686 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11687 dual-core machines) and other potential thread-safety issues.
11691 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11692 versions), which is now available for royalty-free use
11693 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11694 Also, add Camellia TLS ciphersuites from RFC 4132.
11696 To minimize changes between patchlevels in the OpenSSL 0.9.8
11697 series, Camellia remains excluded from compilation unless OpenSSL
11698 is configured with 'enable-camellia'.
11702 * Disable the padding bug check when compression is in use. The padding
11703 bug check assumes the first packet is of even length, this is not
11704 necessarily true if compression is enabled and can result in false
11705 positives causing handshake failure. The actual bug test is ancient
11706 code so it is hoped that implementations will either have fixed it by
11707 now or any which still have the bug do not support compression.
11711 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11713 * When applying a cipher rule check to see if string match is an explicit
11714 cipher suite and only match that one cipher suite if it is.
11718 * Link in manifests for VC++ if needed.
11720 *Austin Ziegler <halostatue@gmail.com>*
11722 * Update support for ECC-based TLS ciphersuites according to
11723 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11724 TLS extensions, which are supported starting with the 0.9.9
11725 branch, not in the OpenSSL 0.9.8 branch).
11729 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11730 opaque EVP_CIPHER_CTX handling.
11734 * Fixes and enhancements to zlib compression code. We now only use
11735 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11736 to conform with the standards mentioned here:
11737 <http://www.zlib.net/DLL_FAQ.txt>
11738 Static zlib linking now works on Windows and the new --with-zlib-include
11739 --with-zlib-lib options to Configure can be used to supply the location
11740 of the headers and library. Gracefully handle case where zlib library
11745 * Several fixes and enhancements to the OID generation code. The old code
11746 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11747 handle numbers larger than ULONG_MAX, truncated printing and had a
11748 non standard OBJ_obj2txt() behaviour.
11752 * Add support for building of engines under engine/ as shared libraries
11753 under VC++ build system.
11757 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11758 Hopefully, we will not see any false combination of paths any more.
11762 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11764 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11765 (part of SSL_OP_ALL). This option used to disable the
11766 countermeasure against man-in-the-middle protocol-version
11767 rollback in the SSL 2.0 server implementation, which is a bad
11768 idea. ([CVE-2005-2969])
11770 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11771 for Information Security, National Institute of Advanced Industrial
11772 Science and Technology [AIST], Japan)*
11774 * Add two function to clear and return the verify parameter flags.
11778 * Keep cipherlists sorted in the source instead of sorting them at
11779 runtime, thus removing the need for a lock.
11783 * Avoid some small subgroup attacks in Diffie-Hellman.
11785 *Nick Mathewson and Ben Laurie*
11787 * Add functions for well-known primes.
11791 * Extended Windows CE support.
11793 *Satoshi Nakamura and Andy Polyakov*
11795 * Initialize SSL_METHOD structures at compile time instead of during
11796 runtime, thus removing the need for a lock.
11800 * Make PKCS7_decrypt() work even if no certificate is supplied by
11801 attempting to decrypt each encrypted key in turn. Add support to
11806 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11808 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11811 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11815 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11816 key into the same file any more.
11820 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11824 * Add -utf8 command line and config file option to 'ca'.
11826 *Stefan <stf@udoma.org*
11828 * Removed the macro des_crypt(), as it seems to conflict with some
11829 libraries. Use DES_crypt().
11833 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11834 involves renaming the source and generated shared-libs for
11835 both. The engines will accept the corrected or legacy ids
11836 ('ncipher' and '4758_cca' respectively) when binding. NB,
11837 this only applies when building 'shared'.
11839 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11841 * Add attribute functions to EVP_PKEY structure. Modify
11842 PKCS12_create() to recognize a CSP name attribute and
11843 use it. Make -CSP option work again in pkcs12 utility.
11847 * Add new functionality to the bn blinding code:
11848 - automatic re-creation of the BN_BLINDING parameters after
11849 a fixed number of uses (currently 32)
11850 - add new function for parameter creation
11851 - introduce flags to control the update behaviour of the
11852 BN_BLINDING parameters
11853 - hide BN_BLINDING structure
11854 Add a second BN_BLINDING slot to the RSA structure to improve
11855 performance when a single RSA object is shared among several
11860 * Add support for DTLS.
11862 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11864 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11865 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11869 * Remove buggy and incomplete DH cert support from
11870 ssl/ssl_rsa.c and ssl/s3_both.c
11874 * Use SHA-1 instead of MD5 as the default digest algorithm for
11875 the `apps/openssl` commands.
11879 * Compile clean with "-Wall -Wmissing-prototypes
11880 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11881 DEBUG_SAFESTACK must also be set.
11885 * Change ./Configure so that certain algorithms can be disabled by default.
11886 The new counterpiece to "no-xxx" is "enable-xxx".
11888 The patented RC5 and MDC2 algorithms will now be disabled unless
11889 "enable-rc5" and "enable-mdc2", respectively, are specified.
11891 (IDEA remains enabled despite being patented. This is because IDEA
11892 is frequently required for interoperability, and there is no license
11893 fee for non-commercial use. As before, "no-idea" can be used to
11894 avoid this algorithm.)
11898 * Add processing of proxy certificates (see RFC 3820). This work was
11899 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11900 EGEE (Enabling Grids for E-science in Europe).
11904 * RC4 performance overhaul on modern architectures/implementations, such
11905 as Intel P4, IA-64 and AMD64.
11909 * New utility extract-section.pl. This can be used specify an alternative
11910 section number in a pod file instead of having to treat each file as
11911 a separate case in Makefile. This can be done by adding two lines to the
11914 =for comment openssl_section:XXX
11916 The blank line is mandatory.
11920 * New arguments -certform, -keyform and -pass for s_client and s_server
11921 to allow alternative format key and certificate files and passphrase
11926 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11927 update associated structures and add various utility functions.
11929 Add new policy related verify parameters, include policy checking in
11930 standard verify code. Enhance 'smime' application with extra parameters
11931 to support policy checking and print out.
11935 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11936 Nehemiah processors. These extensions support AES encryption in hardware
11937 as well as RNG (though RNG support is currently disabled).
11939 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11941 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11945 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11947 *Andy Polyakov and a number of other people*
11949 * Improved PowerPC platform support. Most notably BIGNUM assembler
11950 implementation contributed by IBM.
11952 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11954 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11955 exponent rather than 'unsigned long'. There is a corresponding change to
11956 the new 'rsa_keygen' element of the RSA_METHOD structure.
11958 *Jelte Jansen, Geoff Thorpe*
11960 * Functionality for creating the initial serial number file is now
11961 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11963 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11964 number file to 1, which is bound to cause problems. To avoid
11965 the problems while respecting compatibility between different 0.9.7
11966 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11967 CA.pl for serial number initialization. With the new release 0.9.8,
11968 we can fix the problem directly in the 'ca' utility.)
11972 * Reduced header interdependencies by declaring more opaque objects in
11973 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11974 give fewer recursive includes, which could break lazy source code - so
11975 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11976 developers should define this symbol when building and using openssl to
11977 ensure they track the recommended behaviour, interfaces, [etc], but
11978 backwards-compatible behaviour prevails when this isn't defined.
11982 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11986 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11987 This will generate a random key of the appropriate length based on the
11988 cipher context. The EVP_CIPHER can provide its own random key generation
11989 routine to support keys of a specific form. This is used in the des and
11990 3des routines to generate a key of the correct parity. Update S/MIME
11991 code to use new functions and hence generate correct parity DES keys.
11992 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11993 valid (weak or incorrect parity).
11997 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11998 as looking them up. This is useful when the verified structure may contain
11999 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
12000 present unless the new PKCS7_NO_CRL flag is asserted.
12004 * Extend ASN1 oid configuration module. It now additionally accepts the
12007 shortName = some long name, 1.2.3.4
12011 * Reimplemented the BN_CTX implementation. There is now no more static
12012 limitation on the number of variables it can handle nor the depth of the
12013 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
12014 information can now expand as required, and rather than having a single
12015 static array of bignums, BN_CTX now uses a linked-list of such arrays
12016 allowing it to expand on demand whilst maintaining the usefulness of
12017 BN_CTX's "bundling".
12021 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
12022 to allow all RSA operations to function using a single BN_CTX.
12026 * Preliminary support for certificate policy evaluation and checking. This
12027 is initially intended to pass the tests outlined in "Conformance Testing
12028 of Relying Party Client Certificate Path Processing Logic" v1.07.
12032 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
12033 remained unused and not that useful. A variety of other little bignum
12034 tweaks and fixes have also been made continuing on from the audit (see
12039 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
12040 associated ASN1, EVP and SSL functions and old ASN1 macros.
12044 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
12045 and this should never fail. So the return value from the use of
12046 BN_set_word() (which can fail due to needless expansion) is now deprecated;
12047 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
12051 * BN_CTX_get() should return zero-valued bignums, providing the same
12052 initialised value as BN_new().
12054 *Geoff Thorpe, suggested by Ulf Möller*
12056 * Support for inhibitAnyPolicy certificate extension.
12060 * An audit of the BIGNUM code is underway, for which debugging code is
12061 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
12062 is considered valid when processing BIGNUMs, and causes execution to
12063 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
12064 further steps are taken to deliberately pollute unused data in BIGNUM
12065 structures to try and expose faulty code further on. For now, openssl will
12066 (in its default mode of operation) continue to tolerate the inconsistent
12067 forms that it has tolerated in the past, but authors and packagers should
12068 consider trying openssl and their own applications when compiled with
12069 these debugging symbols defined. It will help highlight potential bugs in
12070 their own code, and will improve the test coverage for OpenSSL itself. At
12071 some point, these tighter rules will become openssl's default to improve
12072 maintainability, though the assert()s and other overheads will remain only
12073 in debugging configurations. See bn.h for more details.
12075 *Geoff Thorpe, Nils Larsch, Ulf Möller*
12077 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
12078 that can only be obtained through BN_CTX_new() (which implicitly
12079 initialises it). The presence of this function only made it possible
12080 to overwrite an existing structure (and cause memory leaks).
12084 * Because of the callback-based approach for implementing LHASH as a
12085 template type, lh_insert() adds opaque objects to hash-tables and
12086 lh_doall() or lh_doall_arg() are typically used with a destructor callback
12087 to clean up those corresponding objects before destroying the hash table
12088 (and losing the object pointers). So some over-zealous constifications in
12089 LHASH have been relaxed so that lh_insert() does not take (nor store) the
12090 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
12091 prototyped to have "const" restrictions on the object pointers they are
12092 given (and so aren't required to cast them away any more).
12096 * The tmdiff.h API was so ugly and minimal that our own timing utility
12097 (speed) prefers to use its own implementation. The two implementations
12098 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
12099 its object type properly exposed (MS_TM) instead of casting to/from
12100 `char *`. This may still change yet if someone realises MS_TM and
12102 aren't necessarily the greatest nomenclatures - but this is what was used
12103 internally to the implementation so I've used that for now.
12107 * Ensure that deprecated functions do not get compiled when
12108 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
12109 the self-tests were still using deprecated key-generation functions so
12110 these have been updated also.
12114 * Reorganise PKCS#7 code to separate the digest location functionality
12115 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
12116 New function PKCS7_set_digest() to set the digest type for PKCS#7
12117 digestedData type. Add additional code to correctly generate the
12118 digestedData type and add support for this type in PKCS7 initialization
12123 * New function PKCS7_set0_type_other() this initializes a PKCS7
12124 structure of type "other".
12128 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
12129 sure the loop does correctly stop and breaking ("division by zero")
12130 modulus operations are not performed. The (pre-generated) prime
12131 table crypto/bn/bn_prime.h was already correct, but it could not be
12132 re-generated on some platforms because of the "division by zero"
12133 situation in the script.
12135 *Ralf S. Engelschall*
12137 * Update support for ECC-based TLS ciphersuites according to
12138 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
12139 SHA-1 now is only used for "small" curves (where the
12140 representation of a field element takes up to 24 bytes); for
12141 larger curves, the field element resulting from ECDH is directly
12142 used as premaster secret.
12144 *Douglas Stebila (Sun Microsystems Laboratories)*
12146 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
12147 curve secp160r1 to the tests.
12149 *Douglas Stebila (Sun Microsystems Laboratories)*
12151 * Add the possibility to load symbols globally with DSO.
12153 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
12155 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
12156 control of the error stack.
12160 * Add support for STORE in ENGINE.
12164 * Add the STORE type. The intention is to provide a common interface
12165 to certificate and key stores, be they simple file-based stores, or
12166 HSM-type store, or LDAP stores, or...
12167 NOTE: The code is currently UNTESTED and isn't really used anywhere.
12171 * Add a generic structure called OPENSSL_ITEM. This can be used to
12172 pass a list of arguments to any function as well as provide a way
12173 for a function to pass data back to the caller.
12177 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
12178 works like BUF_strdup() but can be used to duplicate a portion of
12179 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
12184 * Add the function sk_find_ex() which works like sk_find(), but will
12185 return an index to an element even if an exact match couldn't be
12186 found. The index is guaranteed to point at the element where the
12187 searched-for key would be inserted to preserve sorting order.
12191 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12192 takes an extra flags argument for optional functionality. Currently,
12193 the following flags are defined:
12195 OBJ_BSEARCH_VALUE_ON_NOMATCH
12196 This one gets OBJ_bsearch_ex() to return a pointer to the first
12197 element where the comparing function returns a negative or zero
12200 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12201 This one gets OBJ_bsearch_ex() to return a pointer to the first
12202 element where the comparing function returns zero. This is useful
12203 if there are more than one element where the comparing function
12208 * Make it possible to create self-signed certificates with 'openssl ca'
12209 in such a way that the self-signed certificate becomes part of the
12210 CA database and uses the same mechanisms for serial number generation
12211 as all other certificate signing. The new flag '-selfsign' enables
12212 this functionality. Adapt CA.sh and CA.pl.in.
12216 * Add functionality to check the public key of a certificate request
12217 against a given private. This is useful to check that a certificate
12218 request can be signed by that key (self-signing).
12222 * Make it possible to have multiple active certificates with the same
12223 subject in the CA index file. This is done only if the keyword
12224 'unique_subject' is set to 'no' in the main CA section (default
12225 if 'CA_default') of the configuration file. The value is saved
12226 with the database itself in a separate index attribute file,
12227 named like the index file with '.attr' appended to the name.
12231 * Generate multi-valued AVAs using '+' notation in config files for
12236 * Support for nameConstraints certificate extension.
12240 * Support for policyConstraints certificate extension.
12244 * Support for policyMappings certificate extension.
12248 * Make sure the default DSA_METHOD implementation only uses its
12249 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12250 and change its own handlers to be NULL so as to remove unnecessary
12251 indirection. This lets alternative implementations fallback to the
12252 default implementation more easily.
12256 * Support for directoryName in GeneralName related extensions
12261 * Make it possible to link applications using Makefile.shared.
12262 Make that possible even when linking against static libraries!
12266 * Support for single pass processing for S/MIME signing. This now
12267 means that S/MIME signing can be done from a pipe, in addition
12268 cleartext signing (multipart/signed type) is effectively streaming
12269 and the signed data does not need to be all held in memory.
12271 This is done with a new flag PKCS7_STREAM. When this flag is set
12272 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12273 is done after the data is output (and digests calculated) in
12274 SMIME_write_PKCS7().
12278 * Add full support for -rpath/-R, both in shared libraries and
12279 applications, at least on the platforms where it's known how
12284 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12285 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12286 will now compute a table of multiples of the generator that
12287 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12288 faster (notably in the case of a single point multiplication,
12289 scalar * generator).
12291 *Nils Larsch, Bodo Moeller*
12293 * IPv6 support for certificate extensions. The various extensions
12294 which use the IP:a.b.c.d can now take IPv6 addresses using the
12295 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12300 * Added an ENGINE that implements RSA by performing private key
12301 exponentiations with the GMP library. The conversions to and from
12302 GMP's mpz_t format aren't optimised nor are any montgomery forms
12303 cached, and on x86 it appears OpenSSL's own performance has caught up.
12304 However there are likely to be other architectures where GMP could
12305 provide a boost. This ENGINE is not built in by default, but it can be
12306 specified at Configure time and should be accompanied by the necessary
12307 linker additions, eg;
12308 ./config -DOPENSSL_USE_GMP -lgmp
12312 * "openssl engine" will not display ENGINE/DSO load failure errors when
12313 testing availability of engines with "-t" - the old behaviour is
12314 produced by increasing the feature's verbosity with "-tt".
12318 * ECDSA routines: under certain error conditions uninitialized BN objects
12319 could be freed. Solution: make sure initialization is performed early
12320 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12325 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12326 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12327 software implementations. For DSA and DH, parameter generation can
12328 also be overridden by providing the appropriate method callbacks.
12332 * Change the "progress" mechanism used in key-generation and
12333 primality testing to functions that take a new BN_GENCB pointer in
12334 place of callback/argument pairs. The new API functions have `_ex`
12335 postfixes and the older functions are reimplemented as wrappers for
12336 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12337 declarations of the old functions to help (graceful) attempts to
12338 migrate to the new functions. Also, the new key-generation API
12339 functions operate on a caller-supplied key-structure and return
12340 success/failure rather than returning a key or NULL - this is to
12341 help make "keygen" another member function of RSA_METHOD etc.
12343 Example for using the new callback interface:
12345 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12346 void *my_arg = ...;
12349 BN_GENCB_set(&my_cb, my_callback, my_arg);
12351 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12352 /* For the meaning of a, b in calls to my_callback(), see the
12353 * documentation of the function that calls the callback.
12354 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12355 * my_callback should return 1 if it wants BN_is_prime_ex()
12356 * to continue, or 0 to stop.
12361 * Change the ZLIB compression method to be stateful, and make it
12362 available to TLS with the number defined in
12363 draft-ietf-tls-compression-04.txt.
12367 * Add the ASN.1 structures and functions for CertificatePair, which
12368 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12370 CertificatePair ::= SEQUENCE {
12371 forward [0] Certificate OPTIONAL,
12372 reverse [1] Certificate OPTIONAL,
12373 -- at least one of the pair shall be present -- }
12375 Also implement the PEM functions to read and write certificate
12376 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12378 This needed to be defined, mostly for the sake of the LDAP
12379 attribute crossCertificatePair, but may prove useful elsewhere as
12384 * Make it possible to inhibit symlinking of shared libraries in
12385 Makefile.shared, for Cygwin's sake.
12389 * Extend the BIGNUM API by creating a function
12390 void BN_set_negative(BIGNUM *a, int neg);
12391 and a macro that behave like
12392 int BN_is_negative(const BIGNUM *a);
12394 to avoid the need to access 'a->neg' directly in applications.
12398 * Implement fast modular reduction for pseudo-Mersenne primes
12399 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12400 EC_GROUP_new_curve_GFp() will now automatically use this
12403 *Nils Larsch <nla@trustcenter.de>*
12405 * Add new lock type (CRYPTO_LOCK_BN).
12409 * Change the ENGINE framework to automatically load engines
12410 dynamically from specific directories unless they could be
12411 found to already be built in or loaded. Move all the
12412 current engines except for the cryptodev one to a new
12413 directory engines/.
12414 The engines in engines/ are built as shared libraries if
12415 the "shared" options was given to ./Configure or ./config.
12416 Otherwise, they are inserted in libcrypto.a.
12417 /usr/local/ssl/engines is the default directory for dynamic
12418 engines, but that can be overridden at configure time through
12419 the usual use of --prefix and/or --openssldir, and at run
12420 time with the environment variable OPENSSL_ENGINES.
12422 *Geoff Thorpe and Richard Levitte*
12424 * Add Makefile.shared, a helper makefile to build shared
12425 libraries. Adapt Makefile.org.
12429 * Add version info to Win32 DLLs.
12431 *Peter 'Luna' Runestig" <peter@runestig.com>*
12433 * Add new 'medium level' PKCS#12 API. Certificates and keys
12434 can be added using this API to created arbitrary PKCS#12
12435 files while avoiding the low-level API.
12437 New options to PKCS12_create(), key or cert can be NULL and
12438 will then be omitted from the output file. The encryption
12439 algorithm NIDs can be set to -1 for no encryption, the mac
12440 iteration count can be set to 0 to omit the mac.
12442 Enhance pkcs12 utility by making the -nokeys and -nocerts
12443 options work when creating a PKCS#12 file. New option -nomac
12444 to omit the mac, NONE can be set for an encryption algorithm.
12445 New code is modified to use the enhanced PKCS12_create()
12446 instead of the low-level API.
12450 * Extend ASN1 encoder to support indefinite length constructed
12451 encoding. This can output sequences tags and octet strings in
12452 this form. Modify pk7_asn1.c to support indefinite length
12453 encoding. This is experimental and needs additional code to
12454 be useful, such as an ASN1 bio and some enhanced streaming
12457 Extend template encode functionality so that tagging is passed
12458 down to the template encoder.
12462 * Let 'openssl req' fail if an argument to '-newkey' is not
12463 recognized instead of using RSA as a default.
12467 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12468 As these are not official, they are not included in "ALL";
12469 the "ECCdraft" ciphersuite group alias can be used to select them.
12471 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12473 * Add ECDH engine support.
12475 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12477 * Add ECDH in new directory crypto/ecdh/.
12479 *Douglas Stebila (Sun Microsystems Laboratories)*
12481 * Let BN_rand_range() abort with an error after 100 iterations
12482 without success (which indicates a broken PRNG).
12486 * Change BN_mod_sqrt() so that it verifies that the input value
12487 is really the square of the return value. (Previously,
12488 BN_mod_sqrt would show GIGO behaviour.)
12492 * Add named elliptic curves over binary fields from X9.62, SECG,
12493 and WAP/WTLS; add OIDs that were still missing.
12495 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12497 * Extend the EC library for elliptic curves over binary fields
12498 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12501 EC_GF2m_simple_method
12505 EC_GROUP_new_curve_GF2m
12506 EC_GROUP_set_curve_GF2m
12507 EC_GROUP_get_curve_GF2m
12508 EC_POINT_set_affine_coordinates_GF2m
12509 EC_POINT_get_affine_coordinates_GF2m
12510 EC_POINT_set_compressed_coordinates_GF2m
12512 Point compression for binary fields is disabled by default for
12513 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12516 As binary polynomials are represented as BIGNUMs, various members
12517 of the EC_GROUP and EC_POINT data structures can be shared
12518 between the implementations for prime fields and binary fields;
12519 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12520 are essentially identical to their `..._GFp` counterparts.
12521 (For simplicity, the `..._GFp` prefix has been dropped from
12522 various internal method names.)
12524 An internal 'field_div' method (similar to 'field_mul' and
12525 'field_sqr') has been added; this is used only for binary fields.
12527 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12529 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12530 through methods ('mul', 'precompute_mult').
12532 The generic implementations (now internally called 'ec_wNAF_mul'
12533 and 'ec_wNAF_precomputed_mult') remain the default if these
12534 methods are undefined.
12536 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12538 * New function EC_GROUP_get_degree, which is defined through
12539 EC_METHOD. For curves over prime fields, this returns the bit
12540 length of the modulus.
12542 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12544 * New functions EC_GROUP_dup, EC_POINT_dup.
12545 (These simply call ..._new and ..._copy).
12547 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12549 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12550 Polynomials are represented as BIGNUMs (where the sign bit is not
12551 used) in the following functions [macros]:
12554 BN_GF2m_sub [= BN_GF2m_add]
12555 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12556 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12557 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12559 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12560 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12561 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12562 BN_GF2m_cmp [= BN_ucmp]
12564 (Note that only the 'mod' functions are actually for fields GF(2^m).
12565 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12567 For some functions, an the irreducible polynomial defining a
12568 field can be given as an 'unsigned int[]' with strictly
12569 decreasing elements giving the indices of those bits that are set;
12570 i.e., p[] represents the polynomial
12571 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12573 p[0] > p[1] > ... > p[k] = 0.
12574 This applies to the following functions:
12577 BN_GF2m_mod_mul_arr
12578 BN_GF2m_mod_sqr_arr
12579 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12580 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12581 BN_GF2m_mod_exp_arr
12582 BN_GF2m_mod_sqrt_arr
12583 BN_GF2m_mod_solve_quad_arr
12587 Conversion can be performed by the following functions:
12592 bntest.c has additional tests for binary polynomial arithmetic.
12594 Two implementations for BN_GF2m_mod_div() are available.
12595 The default algorithm simply uses BN_GF2m_mod_inv() and
12596 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12597 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12598 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12600 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12602 * Add new error code 'ERR_R_DISABLED' that can be used when some
12603 functionality is disabled at compile-time.
12605 *Douglas Stebila <douglas.stebila@sun.com>*
12607 * Change default behaviour of 'openssl asn1parse' so that more
12608 information is visible when viewing, e.g., a certificate:
12610 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12611 mode the content of non-printable OCTET STRINGs is output in a
12612 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12613 avoid the appearance of a printable string.
12615 *Nils Larsch <nla@trustcenter.de>*
12617 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12619 EC_GROUP_set_asn1_flag()
12620 EC_GROUP_get_asn1_flag()
12621 EC_GROUP_set_point_conversion_form()
12622 EC_GROUP_get_point_conversion_form()
12623 These control ASN1 encoding details:
12624 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12625 has been set to OPENSSL_EC_NAMED_CURVE.
12626 - Points are encoded in uncompressed form by default; options for
12627 asn1_for are as for point2oct, namely
12628 POINT_CONVERSION_COMPRESSED
12629 POINT_CONVERSION_UNCOMPRESSED
12630 POINT_CONVERSION_HYBRID
12632 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12634 EC_GROUP_set_seed()
12635 EC_GROUP_get0_seed()
12636 EC_GROUP_get_seed_len()
12637 This is used only for ASN1 purposes (so far).
12639 *Nils Larsch <nla@trustcenter.de>*
12641 * Add 'field_type' member to EC_METHOD, which holds the NID
12642 of the appropriate field type OID. The new function
12643 EC_METHOD_get_field_type() returns this value.
12645 *Nils Larsch <nla@trustcenter.de>*
12648 EC_POINT_point2bn()
12649 EC_POINT_bn2point()
12650 EC_POINT_point2hex()
12651 EC_POINT_hex2point()
12652 providing useful interfaces to EC_POINT_point2oct() and
12653 EC_POINT_oct2point().
12655 *Nils Larsch <nla@trustcenter.de>*
12657 * Change internals of the EC library so that the functions
12658 EC_GROUP_set_generator()
12659 EC_GROUP_get_generator()
12660 EC_GROUP_get_order()
12661 EC_GROUP_get_cofactor()
12662 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12663 to methods, which would lead to unnecessary code duplication when
12664 adding different types of curves.
12666 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12668 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12669 arithmetic, and such that modified wNAFs are generated
12670 (which avoid length expansion in many cases).
12674 * Add a function EC_GROUP_check_discriminant() (defined via
12675 EC_METHOD) that verifies that the curve discriminant is non-zero.
12677 Add a function EC_GROUP_check() that makes some sanity tests
12678 on a EC_GROUP, its generator and order. This includes
12679 EC_GROUP_check_discriminant().
12681 *Nils Larsch <nla@trustcenter.de>*
12683 * Add ECDSA in new directory crypto/ecdsa/.
12685 Add applications 'openssl ecparam' and 'openssl ecdsa'
12686 (these are based on 'openssl dsaparam' and 'openssl dsa').
12688 ECDSA support is also included in various other files across the
12689 library. Most notably,
12690 - 'openssl req' now has a '-newkey ecdsa:file' option;
12691 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12692 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12693 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12694 them suitable for ECDSA where domain parameters must be
12695 extracted before the specific public key;
12696 - ECDSA engine support has been added.
12698 *Nils Larsch <nla@trustcenter.de>*
12700 * Include some named elliptic curves, and add OIDs from X9.62,
12701 SECG, and WAP/WTLS. Each curve can be obtained from the new
12703 EC_GROUP_new_by_curve_name(),
12704 and the list of available named curves can be obtained with
12705 EC_get_builtin_curves().
12706 Also add a 'curve_name' member to EC_GROUP objects, which can be
12708 EC_GROUP_set_curve_name()
12709 EC_GROUP_get_curve_name()
12711 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12713 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12714 was actually never needed) and in BN_mul(). The removal in BN_mul()
12715 required a small change in bn_mul_part_recursive() and the addition
12716 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12717 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12718 bn_sub_words() and bn_add_words() except they take arrays with
12723 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12725 * Cleanse PEM buffers before freeing them since they may contain
12728 *Benjamin Bennett <ben@psc.edu>*
12730 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12731 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12732 authentication-only ciphersuites.
12736 * Since AES128 and AES256 share a single mask bit in the logic of
12737 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12738 kludge to work properly if AES128 is available and AES256 isn't.
12742 * Expand security boundary to match 1.1.1 module.
12746 * Remove redundant features: hash file source, editing of test vectors
12747 modify fipsld to use external fips_premain.c signature.
12751 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12752 run algorithm test programs.
12756 * Make algorithm test programs more tolerant of whitespace.
12760 * Have SSL/TLS server implementation tolerate "mismatched" record
12761 protocol version while receiving ClientHello even if the
12762 ClientHello is fragmented. (The server can't insist on the
12763 particular protocol version it has chosen before the ServerHello
12764 message has informed the client about his choice.)
12768 * Load error codes if they are not already present instead of using a
12769 static variable. This allows them to be cleanly unloaded and reloaded.
12773 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12775 * Introduce limits to prevent malicious keys being able to
12776 cause a denial of service. ([CVE-2006-2940])
12778 *Steve Henson, Bodo Moeller*
12780 * Fix ASN.1 parsing of certain invalid structures that can result
12781 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12783 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12784 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12786 * Fix SSL client code which could crash if connecting to a
12787 malicious SSLv2 server. ([CVE-2006-4343])
12789 *Tavis Ormandy and Will Drewry, Google Security Team*
12791 * Change ciphersuite string processing so that an explicit
12792 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12793 will no longer include "AES128-SHA"), and any other similar
12794 ciphersuite (same bitmap) from *other* protocol versions (so that
12795 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12796 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12797 changes from 0.9.8b and 0.9.8d.
12801 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12803 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12804 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12806 * Change the Unix randomness entropy gathering to use poll() when
12807 possible instead of select(), since the latter has some
12808 undesirable limitations.
12810 *Darryl Miles via Richard Levitte and Bodo Moeller*
12812 * Disable rogue ciphersuites:
12814 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12815 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12816 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12818 The latter two were purportedly from
12819 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12822 Also deactivate the remaining ciphersuites from
12823 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12824 unofficial, and the ID has long expired.
12828 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12829 dual-core machines) and other potential thread-safety issues.
12833 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12835 * Adapt fipsld and the build system to link against the validated FIPS
12836 module in FIPS mode.
12840 * Fixes for VC++ 2005 build under Windows.
12844 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12845 from a Windows bash shell such as MSYS. It is autodetected from the
12846 "config" script when run from a VC++ environment. Modify standard VC++
12847 build to use fipscanister.o from the GNU make build.
12851 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12853 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12854 The value now differs depending on if you build for FIPS or not.
12855 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12856 safely run with a non-FIPSed libcrypto, as it may crash because of
12857 the difference induced by this change.
12861 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12863 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12864 (part of SSL_OP_ALL). This option used to disable the
12865 countermeasure against man-in-the-middle protocol-version
12866 rollback in the SSL 2.0 server implementation, which is a bad
12867 idea. ([CVE-2005-2969])
12869 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12870 for Information Security, National Institute of Advanced Industrial
12871 Science and Technology [AIST, Japan)]*
12873 * Minimal support for X9.31 signatures and PSS padding modes. This is
12874 mainly for FIPS compliance and not fully integrated at this stage.
12878 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12879 the exponentiation using a fixed-length exponent. (Otherwise,
12880 the information leaked through timing could expose the secret key
12881 after many signatures; cf. Bleichenbacher's attack on DSA with
12886 * Make a new fixed-window mod_exp implementation the default for
12887 RSA, DSA, and DH private-key operations so that the sequence of
12888 squares and multiplies and the memory access pattern are
12889 independent of the particular secret key. This will mitigate
12890 cache-timing and potential related attacks.
12892 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12893 and this is automatically used by BN_mod_exp_mont() if the new flag
12894 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12895 will use this BN flag for private exponents unless the flag
12896 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12897 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12899 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12901 * Change the client implementation for SSLv23_method() and
12902 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12903 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12904 (Previously, the SSL 2.0 backwards compatible Client Hello
12905 message format would be used even with SSL_OP_NO_SSLv2.)
12909 * Add support for smime-type MIME parameter in S/MIME messages which some
12914 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12915 a threadsafe manner. Modify rsa code to use new function and add calls
12916 to dsa and dh code (which had race conditions before).
12920 * Include the fixed error library code in the C error file definitions
12921 instead of fixing them up at runtime. This keeps the error code
12922 structures constant.
12926 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12928 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12931 * Fixes for newer kerberos headers. NB: the casts are needed because
12932 the 'length' field is signed on one version and unsigned on another
12933 with no (?) obvious way to tell the difference, without these VC++
12934 complains. Also the "definition" of FAR (blank) is no longer included
12935 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12936 some needed definitions.
12940 * Undo Cygwin change.
12944 * Added support for proxy certificates according to RFC 3820.
12945 Because they may be a security thread to unaware applications,
12946 they must be explicitly allowed in run-time. See
12947 docs/HOWTO/proxy_certificates.txt for further information.
12951 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12953 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12954 server and client random values. Previously
12955 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12956 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12958 This change has negligible security impact because:
12960 1. Server and client random values still have 24 bytes of pseudo random
12963 2. Server and client random values are sent in the clear in the initial
12966 3. The master secret is derived using the premaster secret (48 bytes in
12967 size for static RSA ciphersuites) as well as client server and random
12970 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12973 *Stephen Henson, reported by UK NISCC*
12975 * Use Windows randomness collection on Cygwin.
12979 * Fix hang in EGD/PRNGD query when communication socket is closed
12980 prematurely by EGD/PRNGD.
12982 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12984 * Prompt for pass phrases when appropriate for PKCS12 input format.
12988 * Back-port of selected performance improvements from development
12989 branch, as well as improved support for PowerPC platforms.
12993 * Add lots of checks for memory allocation failure, error codes to indicate
12994 failure and freeing up memory if a failure occurs.
12996 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12998 * Add new -passin argument to dgst.
13002 * Perform some character comparisons of different types in X509_NAME_cmp:
13003 this is needed for some certificates that re-encode DNs into UTF8Strings
13004 (in violation of RFC3280) and can't or won't issue name rollover
13009 * Make an explicit check during certificate validation to see that
13010 the CA setting in each certificate on the chain is correct. As a
13011 side effect always do the following basic checks on extensions,
13012 not just when there's an associated purpose to the check:
13014 - if there is an unhandled critical extension (unless the user
13015 has chosen to ignore this fault)
13016 - if the path length has been exceeded (if one is set at all)
13017 - that certain extensions fit the associated purpose (if one has
13022 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
13024 * Avoid a race condition when CRLs are checked in a multi threaded
13025 environment. This would happen due to the reordering of the revoked
13026 entries during signature checking and serial number lookup. Now the
13027 encoding is cached and the serial number sort performed under a lock.
13028 Add new STACK function sk_is_sorted().
13032 * Add Delta CRL to the extension code.
13036 * Various fixes to s3_pkt.c so alerts are sent properly.
13038 *David Holmes <d.holmes@f5.com>*
13040 * Reduce the chances of duplicate issuer name and serial numbers (in
13041 violation of RFC3280) using the OpenSSL certificate creation utilities.
13042 This is done by creating a random 64 bit value for the initial serial
13043 number when a serial number file is created or when a self signed
13044 certificate is created using 'openssl req -x509'. The initial serial
13045 number file is created using 'openssl x509 -next_serial' in CA.pl
13046 rather than being initialized to 1.
13050 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
13052 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13053 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13055 *Joe Orton, Steve Henson*
13057 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
13060 *Joe Orton, Steve Henson*
13062 * Make it possible to have multiple active certificates with the same
13063 subject in the CA index file. This is done only if the keyword
13064 'unique_subject' is set to 'no' in the main CA section (default
13065 if 'CA_default') of the configuration file. The value is saved
13066 with the database itself in a separate index attribute file,
13067 named like the index file with '.attr' appended to the name.
13071 * X509 verify fixes. Disable broken certificate workarounds when
13072 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
13073 keyUsage extension present. Don't accept CRLs with unhandled critical
13074 extensions: since verify currently doesn't process CRL extensions this
13075 rejects a CRL with *any* critical extensions. Add new verify error codes
13080 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
13081 A clarification of RFC2560 will require the use of OCTET STRINGs and
13082 some implementations cannot handle the current raw format. Since OpenSSL
13083 copies and compares OCSP nonces as opaque blobs without any attempt at
13084 parsing them this should not create any compatibility issues.
13088 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
13089 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
13090 this HMAC (and other) operations are several times slower than OpenSSL
13095 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
13097 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13099 * Use the correct content when signing type "other".
13103 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
13105 * Fix various bugs revealed by running the NISCC test suite:
13107 Stop out of bounds reads in the ASN1 code when presented with
13108 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13110 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
13112 If verify callback ignores invalid public key errors don't try to check
13113 certificate signature with the NULL public key.
13117 * New -ignore_err option in ocsp application to stop the server
13118 exiting on the first error in a request.
13122 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13123 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13128 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13129 extra data after the compression methods not only for TLS 1.0
13130 but also for SSL 3.0 (as required by the specification).
13132 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13134 * Change X509_certificate_type() to mark the key as exported/exportable
13135 when it's 512 *bits* long, not 512 bytes.
13139 * Change AES_cbc_encrypt() so it outputs exact multiple of
13140 blocks during encryption.
13144 * Various fixes to base64 BIO and non blocking I/O. On write
13145 flushes were not handled properly if the BIO retried. On read
13146 data was not being buffered properly and had various logic bugs.
13147 This also affects blocking I/O when the data being decoded is a
13152 * Various S/MIME bugfixes and compatibility changes:
13153 output correct application/pkcs7 MIME type if
13154 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
13155 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
13156 of files as .eml work). Correctly handle very long lines in MIME
13161 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
13163 * Countermeasure against the Klima-Pokorny-Rosa extension of
13164 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13165 a protocol version number mismatch like a decryption error
13166 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13170 * Turn on RSA blinding by default in the default implementation
13171 to avoid a timing attack. Applications that don't want it can call
13172 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13173 They would be ill-advised to do so in most cases.
13175 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13177 * Change RSA blinding code so that it works when the PRNG is not
13178 seeded (in this case, the secret RSA exponent is abused as
13179 an unpredictable seed -- if it is not unpredictable, there
13180 is no point in blinding anyway). Make RSA blinding thread-safe
13181 by remembering the creator's thread ID in rsa->blinding and
13182 having all other threads use local one-time blinding factors
13183 (this requires more computation than sharing rsa->blinding, but
13184 avoids excessive locking; and if an RSA object is not shared
13185 between threads, blinding will still be very fast).
13189 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13190 ENGINE as defaults for all supported algorithms irrespective of
13191 the 'flags' parameter. 'flags' is now honoured, so applications
13192 should make sure they are passing it correctly.
13196 * Target "mingw" now allows native Windows code to be generated in
13197 the Cygwin environment as well as with the MinGW compiler.
13201 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13203 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13204 via timing by performing a MAC computation even if incorrect
13205 block cipher padding has been found. This is a countermeasure
13206 against active attacks where the attacker has to distinguish
13207 between bad padding and a MAC verification error. ([CVE-2003-0078])
13209 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13210 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13211 Martin Vuagnoux (EPFL, Ilion)*
13213 * Make the no-err option work as intended. The intention with no-err
13214 is not to have the whole error stack handling routines removed from
13215 libcrypto, it's only intended to remove all the function name and
13216 reason texts, thereby removing some of the footprint that may not
13217 be interesting if those errors aren't displayed anyway.
13219 NOTE: it's still possible for any application or module to have its
13220 own set of error texts inserted. The routines are there, just not
13221 used by default when no-err is given.
13225 * Add support for FreeBSD on IA64.
13227 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13229 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13230 Kerberos function mit_des_cbc_cksum(). Before this change,
13231 the value returned by DES_cbc_cksum() was like the one from
13232 mit_des_cbc_cksum(), except the bytes were swapped.
13234 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13236 * Allow an application to disable the automatic SSL chain building.
13237 Before this a rather primitive chain build was always performed in
13238 ssl3_output_cert_chain(): an application had no way to send the
13239 correct chain if the automatic operation produced an incorrect result.
13241 Now the chain builder is disabled if either:
13243 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13245 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13247 The reasoning behind this is that an application would not want the
13248 auto chain building to take place if extra chain certificates are
13249 present and it might also want a means of sending no additional
13250 certificates (for example the chain has two certificates and the
13255 * Add the possibility to build without the ENGINE framework.
13257 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13259 * Under Win32 gmtime() can return NULL: check return value in
13260 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13264 * DSA routines: under certain error conditions uninitialized BN objects
13265 could be freed. Solution: make sure initialization is performed early
13266 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13267 Nils Larsch <nla@trustcenter.de> via PR#459)
13271 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13272 checked on reconnect on the client side, therefore session resumption
13273 could still fail with a "ssl session id is different" error. This
13274 behaviour is masked when SSL_OP_ALL is used due to
13275 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13276 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13277 followup to PR #377.
13281 * IA-32 assembler support enhancements: unified ELF targets, support
13282 for SCO/Caldera platforms, fix for Cygwin shared build.
13286 * Add support for FreeBSD on sparc64. As a consequence, support for
13287 FreeBSD on non-x86 processors is separate from x86 processors on
13288 the config script, much like the NetBSD support.
13290 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13292 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13294 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13297 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13298 code (06) was taken as the first octet of the session ID and the last
13299 octet was ignored consequently. As a result SSLv2 client side session
13300 caching could not have worked due to the session ID mismatch between
13302 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13307 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13308 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13313 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13314 seems that in spite of existing for more than a year, many application
13315 author have done nothing to provide the necessary callbacks, which
13316 means that this particular engine will not work properly anywhere.
13317 This is a very unfortunate situation which forces us, in the name
13318 of usability, to give the hw_ncipher.c a static lock, which is part
13320 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13321 appear in 0.9.8 or later. We EXPECT application authors to have
13322 dealt properly with this when 0.9.8 is released (unless we actually
13323 make such changes in the libcrypto locking code that changes will
13324 have to be made anyway).
13328 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13329 octets have been read, EOF or an error occurs. Without this change
13330 some truncated ASN1 structures will not produce an error.
13334 * Disable Heimdal support, since it hasn't been fully implemented.
13335 Still give the possibility to force the use of Heimdal, but with
13336 warnings and a request that patches get sent to openssl-dev.
13340 * Add the VC-CE target, introduce the WINCE sysname, and add
13341 INSTALL.WCE and appropriate conditionals to make it build.
13343 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13345 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13346 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13347 edit numbers of the version.
13349 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13351 * Introduce safe string copy and catenation functions
13352 (BUF_strlcpy() and BUF_strlcat()).
13354 *Ben Laurie (CHATS) and Richard Levitte*
13356 * Avoid using fixed-size buffers for one-line DNs.
13358 *Ben Laurie (CHATS)*
13360 * Add BUF_MEM_grow_clean() to avoid information leakage when
13361 resizing buffers containing secrets, and use where appropriate.
13363 *Ben Laurie (CHATS)*
13365 * Avoid using fixed size buffers for configuration file location.
13367 *Ben Laurie (CHATS)*
13369 * Avoid filename truncation for various CA files.
13371 *Ben Laurie (CHATS)*
13373 * Use sizeof in preference to magic numbers.
13375 *Ben Laurie (CHATS)*
13377 * Avoid filename truncation in cert requests.
13379 *Ben Laurie (CHATS)*
13381 * Add assertions to check for (supposedly impossible) buffer
13384 *Ben Laurie (CHATS)*
13386 * Don't cache truncated DNS entries in the local cache (this could
13387 potentially lead to a spoofing attack).
13389 *Ben Laurie (CHATS)*
13391 * Fix various buffers to be large enough for hex/decimal
13392 representations in a platform independent manner.
13394 *Ben Laurie (CHATS)*
13396 * Add CRYPTO_realloc_clean() to avoid information leakage when
13397 resizing buffers containing secrets, and use where appropriate.
13399 *Ben Laurie (CHATS)*
13401 * Add BIO_indent() to avoid much slightly worrying code to do
13404 *Ben Laurie (CHATS)*
13406 * Convert sprintf()/BIO_puts() to BIO_printf().
13408 *Ben Laurie (CHATS)*
13410 * buffer_gets() could terminate with the buffer only half
13413 *Ben Laurie (CHATS)*
13415 * Add assertions to prevent user-supplied crypto functions from
13416 overflowing internal buffers by having large block sizes, etc.
13418 *Ben Laurie (CHATS)*
13420 * New OPENSSL_assert() macro (similar to assert(), but enabled
13423 *Ben Laurie (CHATS)*
13425 * Eliminate unused copy of key in RC4.
13427 *Ben Laurie (CHATS)*
13429 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13431 *Ben Laurie (CHATS)*
13433 * Fix off-by-one error in EGD path.
13435 *Ben Laurie (CHATS)*
13437 * If RANDFILE path is too long, ignore instead of truncating.
13439 *Ben Laurie (CHATS)*
13441 * Eliminate unused and incorrectly sized X.509 structure
13444 *Ben Laurie (CHATS)*
13446 * Eliminate unused and dangerous function knumber().
13448 *Ben Laurie (CHATS)*
13450 * Eliminate unused and dangerous structure, KSSL_ERR.
13452 *Ben Laurie (CHATS)*
13454 * Protect against overlong session ID context length in an encoded
13455 session object. Since these are local, this does not appear to be
13458 *Ben Laurie (CHATS)*
13460 * Change from security patch (see 0.9.6e below) that did not affect
13461 the 0.9.6 release series:
13463 Remote buffer overflow in SSL3 protocol - an attacker could
13464 supply an oversized master key in Kerberos-enabled versions.
13467 *Ben Laurie (CHATS)*
13469 * Change the SSL kerb5 codes to match RFC 2712.
13473 * Make -nameopt work fully for req and add -reqopt switch.
13475 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13477 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13479 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13481 * Make sure tests can be performed even if the corresponding algorithms
13482 have been removed entirely. This was also the last step to make
13483 OpenSSL compilable with DJGPP under all reasonable conditions.
13485 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13487 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13488 to allow version independent disabling of normally unselected ciphers,
13489 which may be activated as a side-effect of selecting a single cipher.
13491 (E.g., cipher list string "RSA" enables ciphersuites that are left
13492 out of "ALL" because they do not provide symmetric encryption.
13493 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13495 *Lutz Jaenicke, Bodo Moeller*
13497 * Add appropriate support for separate platform-dependent build
13498 directories. The recommended way to make a platform-dependent
13499 build directory is the following (tested on Linux), maybe with
13502 # Place yourself outside of the OpenSSL source tree. In
13503 # this example, the environment variable OPENSSL_SOURCE
13504 # is assumed to contain the absolute OpenSSL source directory.
13505 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13506 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13507 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13508 mkdir -p `dirname $F`
13509 ln -s $OPENSSL_SOURCE/$F $F
13512 To be absolutely sure not to disturb the source tree, a "make clean"
13513 is a good thing. If it isn't successful, don't worry about it,
13514 it probably means the source directory is very clean.
13518 * Make sure any ENGINE control commands make local copies of string
13519 pointers passed to them whenever necessary. Otherwise it is possible
13520 the caller may have overwritten (or deallocated) the original string
13521 data when a later ENGINE operation tries to use the stored values.
13523 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13525 * Improve diagnostics in file reading and command-line digests.
13527 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13529 * Add AES modes CFB and OFB to the object database. Correct an
13530 error in AES-CFB decryption.
13534 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13535 allows existing EVP_CIPHER_CTX structures to be reused after
13536 calling `EVP_*Final()`. This behaviour is used by encryption
13537 BIOs and some applications. This has the side effect that
13538 applications must explicitly clean up cipher contexts with
13539 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13543 * Check the values of dna and dnb in bn_mul_recursive before calling
13544 bn_mul_comba (a non zero value means the a or b arrays do not contain
13545 n2 elements) and fallback to bn_mul_normal if either is not zero.
13549 * Fix escaping of non-ASCII characters when using the -subj option
13550 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13554 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13555 form for "surname", serialNumber has no short form.
13556 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13557 therefore remove "mail" short name for "internet 7".
13558 The OID for unique identifiers in X509 certificates is
13559 x500UniqueIdentifier, not uniqueIdentifier.
13560 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13564 * Add an "init" command to the ENGINE config module and auto initialize
13565 ENGINEs. Without any "init" command the ENGINE will be initialized
13566 after all ctrl commands have been executed on it. If init=1 the
13567 ENGINE is initialized at that point (ctrls before that point are run
13568 on the uninitialized ENGINE and after on the initialized one). If
13569 init=0 then the ENGINE will not be initialized at all.
13573 * Fix the 'app_verify_callback' interface so that the user-defined
13574 argument is actually passed to the callback: In the
13575 SSL_CTX_set_cert_verify_callback() prototype, the callback
13576 declaration has been changed from
13579 int (*cb)(X509_STORE_CTX *,void *);
13580 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13581 i=s->ctx->app_verify_callback(&ctx)
13582 has been changed into
13583 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13585 To update applications using SSL_CTX_set_cert_verify_callback(),
13586 a dummy argument can be added to their callback functions.
13588 *D. K. Smetters <smetters@parc.xerox.com>*
13590 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13592 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13594 * Add and OPENSSL_LOAD_CONF define which will cause
13595 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13596 This allows older applications to transparently support certain
13597 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13598 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13599 load the config file and OPENSSL_add_all_algorithms_conf() which will
13600 always load it have also been added.
13604 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13605 Adjust NIDs and EVP layer.
13607 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13609 * Config modules support in openssl utility.
13611 Most commands now load modules from the config file,
13612 though in a few (such as version) this isn't done
13613 because it couldn't be used for anything.
13615 In the case of ca and req the config file used is
13616 the same as the utility itself: that is the -config
13617 command line option can be used to specify an
13622 * Move default behaviour from OPENSSL_config(). If appname is NULL
13623 use "openssl_conf" if filename is NULL use default openssl config file.
13627 * Add an argument to OPENSSL_config() to allow the use of an alternative
13628 config section name. Add a new flag to tolerate a missing config file
13629 and move code to CONF_modules_load_file().
13633 * Support for crypto accelerator cards from Accelerated Encryption
13634 Processing, www.aep.ie. (Use engine 'aep')
13635 The support was copied from 0.9.6c [engine] and adapted/corrected
13636 to work with the new engine framework.
13638 *AEP Inc. and Richard Levitte*
13640 * Support for SureWare crypto accelerator cards from Baltimore
13641 Technologies. (Use engine 'sureware')
13642 The support was copied from 0.9.6c [engine] and adapted
13643 to work with the new engine framework.
13647 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13648 make the newer ENGINE framework commands for the CHIL engine work.
13650 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13652 * Make it possible to produce shared libraries on ReliantUNIX.
13654 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13656 * Add the configuration target debug-linux-ppro.
13657 Make 'openssl rsa' use the general key loading routines
13658 implemented in `apps.c`, and make those routines able to
13659 handle the key format FORMAT_NETSCAPE and the variant
13662 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13664 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13666 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13668 * Add -keyform to rsautl, and document -engine.
13670 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13672 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13673 BIO_R_NO_SUCH_FILE error code rather than the generic
13674 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13678 * Add new functions
13679 ERR_peek_last_error
13680 ERR_peek_last_error_line
13681 ERR_peek_last_error_line_data.
13682 These are similar to
13684 ERR_peek_error_line
13685 ERR_peek_error_line_data,
13686 but report on the latest error recorded rather than the first one
13687 still in the error queue.
13689 *Ben Laurie, Bodo Moeller*
13691 * default_algorithms option in ENGINE config module. This allows things
13693 default_algorithms = ALL
13694 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13698 * Preliminary ENGINE config module.
13702 * New experimental application configuration code.
13706 * Change the AES code to follow the same name structure as all other
13707 symmetric ciphers, and behave the same way. Move everything to
13708 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13710 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13712 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13714 *Ben Laurie and Theo de Raadt*
13716 * Add option to output public keys in req command.
13718 *Massimiliano Pala madwolf@openca.org*
13720 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13721 (up to about 10% better than before for P-192 and P-224).
13725 * New functions/macros
13727 SSL_CTX_set_msg_callback(ctx, cb)
13728 SSL_CTX_set_msg_callback_arg(ctx, arg)
13729 SSL_set_msg_callback(ssl, cb)
13730 SSL_set_msg_callback_arg(ssl, arg)
13732 to request calling a callback function
13734 void cb(int write_p, int version, int content_type,
13735 const void *buf, size_t len, SSL *ssl, void *arg)
13737 whenever a protocol message has been completely received
13738 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13739 protocol version according to which the SSL library interprets
13740 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13741 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13742 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13743 specification (change_cipher_spec(20), alert(21), handshake(22)).
13744 'buf' and 'len' point to the actual message, 'ssl' to the
13745 SSL object, and 'arg' is the application-defined value set by
13746 SSL[_CTX]_set_msg_callback_arg().
13748 'openssl s_client' and 'openssl s_server' have new '-msg' options
13749 to enable a callback that displays all protocol messages.
13753 * Change the shared library support so shared libraries are built as
13754 soon as the corresponding static library is finished, and thereby get
13755 openssl and the test programs linked against the shared library.
13756 This still only happens when the keyword "shard" has been given to
13757 the configuration scripts.
13759 NOTE: shared library support is still an experimental thing, and
13760 backward binary compatibility is still not guaranteed.
13762 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13764 * Add support for Subject Information Access extension.
13766 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13768 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13769 additional bytes when new memory had to be allocated, not just
13770 when reusing an existing buffer.
13774 * New command line and configuration option 'utf8' for the req command.
13775 This allows field values to be specified as UTF8 strings.
13779 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13780 runs for the former and machine-readable output for the latter.
13784 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13785 of the e-mail address in the DN (i.e., it will go into a certificate
13786 extension only). The new configuration file option 'email_in_dn = no'
13787 has the same effect.
13789 *Massimiliano Pala madwolf@openca.org*
13791 * Change all functions with names starting with `des_` to be starting
13792 with `DES_` instead. Add wrappers that are compatible with libdes,
13793 but are named `_ossl_old_des_*`. Finally, add macros that map the
13794 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13795 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13796 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13799 Since we provide two compatibility mappings, the user needs to
13800 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13801 compatibility is desired. The default (i.e., when that macro
13802 isn't defined) is OpenSSL 0.9.6c compatibility.
13804 There are also macros that enable and disable the support of old
13805 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13806 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13807 are defined, the default will apply: to support the old des routines.
13809 In either case, one must include openssl/des.h to get the correct
13810 definitions. Do not try to just include openssl/des_old.h, that
13813 NOTE: This is a major break of an old API into a new one. Software
13814 authors are encouraged to switch to the `DES_` style functions. Some
13815 time in the future, des_old.h and the libdes compatibility functions
13816 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13817 default), and then completely removed.
13821 * Test for certificates which contain unsupported critical extensions.
13822 If such a certificate is found during a verify operation it is
13823 rejected by default: this behaviour can be overridden by either
13824 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13825 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13826 X509_supported_extension() has also been added which returns 1 if a
13827 particular extension is supported.
13831 * Modify the behaviour of EVP cipher functions in similar way to digests
13832 to retain compatibility with existing code.
13836 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13837 compatibility with existing code. In particular the 'ctx' parameter does
13838 not have to be to be initialized before the call to EVP_DigestInit() and
13839 it is tidied up after a call to EVP_DigestFinal(). New function
13840 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13841 EVP_MD_CTX_copy() changed to not require the destination to be
13842 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13843 requires the destination to be valid.
13845 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13846 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13850 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13851 so that complete 'Handshake' protocol structures are kept in memory
13852 instead of overwriting 'msg_type' and 'length' with 'body' data.
13856 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13858 *Massimo Santin via Richard Levitte*
13860 * Major restructuring to the underlying ENGINE code. This includes
13861 reduction of linker bloat, separation of pure "ENGINE" manipulation
13862 (initialisation, etc) from functionality dealing with implementations
13863 of specific crypto interfaces. This change also introduces integrated
13864 support for symmetric ciphers and digest implementations - so ENGINEs
13865 can now accelerate these by providing EVP_CIPHER and EVP_MD
13866 implementations of their own. This is detailed in
13867 [crypto/engine/README.md](crypto/engine/README.md)
13868 as it couldn't be adequately described here. However, there are a few
13869 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13870 were changed in the original introduction of ENGINE code have now
13871 reverted back - the hooking from this code to ENGINE is now a good
13872 deal more passive and at run-time, operations deal directly with
13873 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13874 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13875 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13876 they were not being used by the framework as there is no concept of a
13877 BIGNUM_METHOD and they could not be generalised to the new
13878 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13879 ENGINE_cpy() has been removed as it cannot be consistently defined in
13884 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13888 * Change mkdef.pl to sort symbols that get the same entry number,
13889 and make sure the automatically generated functions `ERR_load_*`
13890 become part of libeay.num as well.
13894 * New function SSL_renegotiate_pending(). This returns true once
13895 renegotiation has been requested (either SSL_renegotiate() call
13896 or HelloRequest/ClientHello received from the peer) and becomes
13897 false once a handshake has been completed.
13898 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13899 sends a HelloRequest, but does not ensure that a handshake takes
13900 place. SSL_renegotiate_pending() is useful for checking if the
13901 client has followed the request.)
13905 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13906 By default, clients may request session resumption even during
13907 renegotiation (if session ID contexts permit); with this option,
13908 session resumption is possible only in the first handshake.
13910 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13911 more bits available for options that should not be part of
13912 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13916 * Add some demos for certificate and certificate request creation.
13920 * Make maximum certificate chain size accepted from the peer application
13921 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13922 "Douglas E. Engert" <deengert@anl.gov>.
13926 * Add support for shared libraries for Unixware-7
13927 (Boyd Lynn Gerber <gerberb@zenez.com>).
13931 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13932 be done prior to destruction. Use this to unload error strings from
13933 ENGINEs that load their own error strings. NB: This adds two new API
13934 functions to "get" and "set" this destroy handler in an ENGINE.
13938 * Alter all existing ENGINE implementations (except "openssl" and
13939 "openbsd") to dynamically instantiate their own error strings. This
13940 makes them more flexible to be built both as statically-linked ENGINEs
13941 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13942 Also, add stub code to each that makes building them as self-contained
13943 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13947 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13948 implementations into applications that are completely implemented in
13949 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13950 commands that can be used to configure what shared-library to load and
13951 to control aspects of the way it is handled. Also, made an update to
13952 the [README-Engine.md](README-Engine.md) file
13953 that brings its information up-to-date and
13954 provides some information and instructions on the "dynamic" ENGINE
13955 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13959 * Make it possible to unload ranges of ERR strings with a new
13960 "ERR_unload_strings" function.
13964 * Add a copy() function to EVP_MD.
13968 * Make EVP_MD routines take a context pointer instead of just the
13969 md_data void pointer.
13973 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13974 that the digest can only process a single chunk of data
13975 (typically because it is provided by a piece of
13976 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13977 is only going to provide a single chunk of data, and hence the
13978 framework needn't accumulate the data for oneshot drivers.
13982 * As with "ERR", make it possible to replace the underlying "ex_data"
13983 functions. This change also alters the storage and management of global
13984 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13985 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13986 index counters. The API functions that use this state have been changed
13987 to take a "class_index" rather than pointers to the class's local STACK
13988 and counter, and there is now an API function to dynamically create new
13989 classes. This centralisation allows us to (a) plug a lot of the
13990 thread-safety problems that existed, and (b) makes it possible to clean
13991 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13992 such data would previously have always leaked in application code and
13993 workarounds were in place to make the memory debugging turn a blind eye
13994 to it. Application code that doesn't use this new function will still
13995 leak as before, but their memory debugging output will announce it now
13996 rather than letting it slide.
13998 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13999 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
14000 has a return value to indicate success or failure.
14004 * Make it possible to replace the underlying "ERR" functions such that the
14005 global state (2 LHASH tables and 2 locks) is only used by the "default"
14006 implementation. This change also adds two functions to "get" and "set"
14007 the implementation prior to it being automatically set the first time
14008 any other ERR function takes place. Ie. an application can call "get",
14009 pass the return value to a module it has just loaded, and that module
14010 can call its own "set" function using that value. This means the
14011 module's "ERR" operations will use (and modify) the error state in the
14012 application and not in its own statically linked copy of OpenSSL code.
14016 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
14017 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
14018 the operation, and provides a more encapsulated way for external code
14019 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
14020 to use these functions rather than manually incrementing the counts.
14022 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
14026 * Add EVP test program.
14030 * Add symmetric cipher support to ENGINE. Expect the API to change!
14034 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
14035 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
14036 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
14037 These allow a CRL to be built without having to access X509_CRL fields
14038 directly. Modify 'ca' application to use new functions.
14042 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
14043 bug workarounds. Rollback attack detection is a security feature.
14044 The problem will only arise on OpenSSL servers when TLSv1 is not
14045 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
14046 Software authors not wanting to support TLSv1 will have special reasons
14047 for their choice and can explicitly enable this option.
14049 *Bodo Moeller, Lutz Jaenicke*
14051 * Rationalise EVP so it can be extended: don't include a union of
14052 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
14053 (similar to those existing for EVP_CIPHER_CTX).
14058 EVP_MD_CTX_init(&md); /* new function call */
14059 EVP_DigestInit(&md, EVP_sha1());
14060 EVP_DigestUpdate(&md, in, len);
14061 EVP_DigestFinal(&md, out, NULL);
14062 EVP_MD_CTX_cleanup(&md); /* new function call */
14066 * Make DES key schedule conform to the usual scheme, as well as
14067 correcting its structure. This means that calls to DES functions
14068 now have to pass a pointer to a des_key_schedule instead of a
14069 plain des_key_schedule (which was actually always a pointer
14072 des_key_schedule ks;
14074 des_set_key_checked(..., &ks);
14075 des_ncbc_encrypt(..., &ks, ...);
14077 (Note that a later change renames 'des_...' into 'DES_...'.)
14081 * Initial reduction of linker bloat: the use of some functions, such as
14082 PEM causes large amounts of unused functions to be linked in due to
14083 poor organisation. For example pem_all.c contains every PEM function
14084 which has a knock on effect of linking in large amounts of (unused)
14085 ASN1 code. Grouping together similar functions and splitting unrelated
14086 functions prevents this.
14090 * Cleanup of EVP macros.
14094 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
14095 correct `_ecb suffix`.
14099 * Add initial OCSP responder support to ocsp application. The
14100 revocation information is handled using the text based index
14101 use by the ca application. The responder can either handle
14102 requests generated internally, supplied in files (for example
14103 via a CGI script) or using an internal minimal server.
14107 * Add configuration choices to get zlib compression for TLS.
14111 * Changes to Kerberos SSL for RFC 2712 compliance:
14112 1. Implemented real KerberosWrapper, instead of just using
14113 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
14114 2. Implemented optional authenticator field of KerberosWrapper.
14116 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
14117 and authenticator structs; see crypto/krb5/.
14119 Generalized Kerberos calls to support multiple Kerberos libraries.
14120 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
14121 via Richard Levitte*
14123 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
14124 already does with RSA. testdsa.h now has 'priv_key/pub_key'
14125 values for each of the key sizes rather than having just
14126 parameters (and 'speed' generating keys each time).
14130 * Speed up EVP routines.
14133 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
14134 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
14135 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
14136 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
14138 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
14139 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
14140 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
14143 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
14145 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
14149 * Added the OS2-EMX target.
14151 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
14153 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
14154 New functions to support `NCONF` routines in extension code.
14155 New function `CONF_set_nconf()`
14156 to allow functions which take an `NCONF` to also handle the old `LHASH`
14157 structure: this means that the old `CONF` compatible routines can be
14158 retained (in particular w.rt. extensions) without having to duplicate the
14159 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
14163 * Enhance the general user interface with mechanisms for inner control
14164 and with possibilities to have yes/no kind of prompts.
14168 * Change all calls to low-level digest routines in the library and
14169 applications to use EVP. Add missing calls to HMAC_cleanup() and
14170 don't assume HMAC_CTX can be copied using memcpy().
14172 *Verdon Walker <VWalker@novell.com>, Steve Henson*
14174 * Add the possibility to control engines through control names but with
14175 arbitrary arguments instead of just a string.
14176 Change the key loaders to take a UI_METHOD instead of a callback
14177 function pointer. NOTE: this breaks binary compatibility with earlier
14178 versions of OpenSSL [engine].
14179 Adapt the nCipher code for these new conditions and add a card insertion
14184 * Enhance the general user interface with mechanisms to better support
14185 dialog box interfaces, application-defined prompts, the possibility
14186 to use defaults (for example default passwords from somewhere else)
14187 and interrupts/cancellations.
14191 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14192 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14196 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14197 tidy up some unnecessarily weird code in 'sk_new()').
14199 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14201 * Change the key loading routines for ENGINEs to use the same kind
14202 callback (pem_password_cb) as all other routines that need this
14207 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14208 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14209 than this minimum value is recommended.
14213 * New random seeder for OpenVMS, using the system process statistics
14214 that are easily reachable.
14218 * Windows apparently can't transparently handle global
14219 variables defined in DLLs. Initialisations such as:
14221 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14223 won't compile. This is used by the any applications that need to
14224 declare their own ASN1 modules. This was fixed by adding the option
14225 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14226 needed for static libraries under Win32.
14230 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14231 setting of purpose and trust fields. New X509_STORE trust and
14232 purpose functions and tidy up setting in other SSL functions.
14236 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14237 structure. These are inherited by X509_STORE_CTX when it is
14238 initialised. This allows various defaults to be set in the
14239 X509_STORE structure (such as flags for CRL checking and custom
14240 purpose or trust settings) for functions which only use X509_STORE_CTX
14241 internally such as S/MIME.
14243 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14244 trust settings if they are not set in X509_STORE. This allows X509_STORE
14245 purposes and trust (in S/MIME for example) to override any set by default.
14247 Add command line options for CRL checking to smime, s_client and s_server
14252 * Initial CRL based revocation checking. If the CRL checking flag(s)
14253 are set then the CRL is looked up in the X509_STORE structure and
14254 its validity and signature checked, then if the certificate is found
14255 in the CRL the verify fails with a revoked error.
14257 Various new CRL related callbacks added to X509_STORE_CTX structure.
14259 Command line options added to 'verify' application to support this.
14261 This needs some additional work, such as being able to handle multiple
14262 CRLs with different times, extension based lookup (rather than just
14263 by subject name) and ultimately more complete V2 CRL extension
14268 * Add a general user interface API (crypto/ui/). This is designed
14269 to replace things like des_read_password and friends (backward
14270 compatibility functions using this new API are provided).
14271 The purpose is to remove prompting functions from the DES code
14272 section as well as provide for prompting through dialog boxes in
14273 a window system and the like.
14277 * Add "ex_data" support to ENGINE so implementations can add state at a
14278 per-structure level rather than having to store it globally.
14282 * Make it possible for ENGINE structures to be copied when retrieved by
14283 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14284 This causes the "original" ENGINE structure to act like a template,
14285 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14286 operational state can be localised to each ENGINE structure, despite the
14287 fact they all share the same "methods". New ENGINE structures returned in
14288 this case have no functional references and the return value is the single
14289 structural reference. This matches the single structural reference returned
14290 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14295 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14296 needs to match any other type at all we need to manually clear the
14301 * Changes to the "openssl engine" utility to include;
14302 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14303 about an ENGINE's available control commands.
14304 - executing control commands from command line arguments using the
14305 '-pre' and '-post' switches. '-post' is only used if '-t' is
14306 specified and the ENGINE is successfully initialised. The syntax for
14307 the individual commands are colon-separated, for example;
14308 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14312 * New dynamic control command support for ENGINEs. ENGINEs can now
14313 declare their own commands (numbers), names (strings), descriptions,
14314 and input types for run-time discovery by calling applications. A
14315 subset of these commands are implicitly classed as "executable"
14316 depending on their input type, and only these can be invoked through
14317 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14318 can be based on user input, config files, etc). The distinction is
14319 that "executable" commands cannot return anything other than a boolean
14320 result and can only support numeric or string input, whereas some
14321 discoverable commands may only be for direct use through
14322 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14323 pointers, or other custom uses. The "executable" commands are to
14324 support parameterisations of ENGINE behaviour that can be
14325 unambiguously defined by ENGINEs and used consistently across any
14326 OpenSSL-based application. Commands have been added to all the
14327 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14328 control over shared-library paths without source code alterations.
14332 * Changed all ENGINE implementations to dynamically allocate their
14333 ENGINEs rather than declaring them statically. Apart from this being
14334 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14335 this also allows the implementations to compile without using the
14336 internal engine_int.h header.
14340 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14341 'const' value. Any code that should be able to modify a RAND_METHOD
14342 should already have non-const pointers to it (ie. they should only
14343 modify their own ones).
14347 * Made a variety of little tweaks to the ENGINE code.
14348 - "atalla" and "ubsec" string definitions were moved from header files
14349 to C code. "nuron" string definitions were placed in variables
14350 rather than hard-coded - allowing parameterisation of these values
14351 later on via ctrl() commands.
14352 - Removed unused "#if 0"'d code.
14353 - Fixed engine list iteration code so it uses ENGINE_free() to release
14354 structural references.
14355 - Constified the RAND_METHOD element of ENGINE structures.
14356 - Constified various get/set functions as appropriate and added
14357 missing functions (including a catch-all ENGINE_cpy that duplicates
14358 all ENGINE values onto a new ENGINE except reference counts/state).
14359 - Removed NULL parameter checks in get/set functions. Setting a method
14360 or function to NULL is a way of cancelling out a previously set
14361 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14362 and doesn't justify the extra error symbols and code.
14363 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14364 flags from engine_int.h to engine.h.
14365 - Changed prototypes for ENGINE handler functions (init(), finish(),
14366 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14370 * Implement binary inversion algorithm for BN_mod_inverse in addition
14371 to the algorithm using long division. The binary algorithm can be
14372 used only if the modulus is odd. On 32-bit systems, it is faster
14373 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14374 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14375 up to 450 bits. In 64-bit environments, the binary algorithm
14376 appears to be advantageous for much longer moduli; here we use it
14377 for moduli up to 2048 bits.
14381 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14382 could not support the combine flag in choice fields.
14386 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14387 extensions from a certificate request to the certificate.
14391 * Allow multiple 'certopt' and 'nameopt' options to be separated
14392 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14393 file: this allows the display of the certificate about to be
14394 signed to be customised, to allow certain fields to be included
14395 or excluded and extension details. The old system didn't display
14396 multicharacter strings properly, omitted fields not in the policy
14397 and couldn't display additional details such as extensions.
14401 * Function EC_POINTs_mul for multiple scalar multiplication
14402 of an arbitrary number of elliptic curve points
14403 \sum scalars[i]*points[i],
14404 optionally including the generator defined for the EC_GROUP:
14405 scalar*generator + \sum scalars[i]*points[i].
14407 EC_POINT_mul is a simple wrapper function for the typical case
14408 that the point list has just one item (besides the optional
14413 * First EC_METHODs for curves over GF(p):
14415 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14416 operations and provides various method functions that can also
14417 operate with faster implementations of modular arithmetic.
14419 EC_GFp_mont_method() reuses most functions that are part of
14420 EC_GFp_simple_method, but uses Montgomery arithmetic.
14422 *Bodo Moeller; point addition and point doubling
14423 implementation directly derived from source code provided by
14424 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14426 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14427 crypto/ec/ec_lib.c):
14429 Curves are EC_GROUP objects (with an optional group generator)
14430 based on EC_METHODs that are built into the library.
14432 Points are EC_POINT objects based on EC_GROUP objects.
14434 Most of the framework would be able to handle curves over arbitrary
14435 finite fields, but as there are no obvious types for fields other
14436 than GF(p), some functions are limited to that for now.
14440 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14441 that the file contains a complete HTTP response.
14445 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14446 change the def and num file printf format specifier from "%-40sXXX"
14447 to "%-39s XXX". The latter will always guarantee a space after the
14448 field while the former will cause them to run together if the field
14449 is 40 of more characters long.
14453 * Constify the cipher and digest 'method' functions and structures
14454 and modify related functions to take constant EVP_MD and EVP_CIPHER
14459 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14460 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14464 * Modify `EVP_Digest*()` routines so they now return values. Although the
14465 internal software routines can never fail additional hardware versions
14470 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14472 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14473 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14476 ERR_R_NESTED_ASN1_ERROR
14478 ERR_R_MISSING_ASN1_EOS
14479 were 4 .. 9, conflicting with
14480 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14482 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14483 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14485 Add new error code 'ERR_R_INTERNAL_ERROR'.
14489 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14494 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14495 sets the subject name for a new request or supersedes the
14496 subject name in a given request. Formats that can be parsed are
14497 'CN=Some Name, OU=myOU, C=IT'
14499 'CN=Some Name/OU=myOU/C=IT'.
14501 Add options '-batch' and '-verbose' to 'openssl req'.
14503 *Massimiliano Pala <madwolf@hackmasters.net>*
14505 * Introduce the possibility to access global variables through
14506 functions on platform were that's the best way to handle exporting
14507 global variables in shared libraries. To enable this functionality,
14508 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14509 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14510 is normally done by Configure or something similar).
14512 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14513 in the source file (foo.c) like this:
14515 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14516 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14518 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14519 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14521 OPENSSL_DECLARE_GLOBAL(int,foo);
14522 #define foo OPENSSL_GLOBAL_REF(foo)
14523 OPENSSL_DECLARE_GLOBAL(double,bar);
14524 #define bar OPENSSL_GLOBAL_REF(bar)
14526 The #defines are very important, and therefore so is including the
14527 header file everywhere where the defined globals are used.
14529 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14530 of ASN.1 items, but that structure is a bit different.
14532 The largest change is in util/mkdef.pl which has been enhanced with
14533 better and easier to understand logic to choose which symbols should
14534 go into the Windows .def files as well as a number of fixes and code
14535 cleanup (among others, algorithm keywords are now sorted
14536 lexicographically to avoid constant rewrites).
14540 * In BN_div() keep a copy of the sign of 'num' before writing the
14541 result to 'rm' because if rm==num the value will be overwritten
14542 and produce the wrong result if 'num' is negative: this caused
14543 problems with BN_mod() and BN_nnmod().
14547 * Function OCSP_request_verify(). This checks the signature on an
14548 OCSP request and verifies the signer certificate. The signer
14549 certificate is just checked for a generic purpose and OCSP request
14554 * Add OCSP_check_validity() function to check the validity of OCSP
14555 responses. OCSP responses are prepared in real time and may only
14556 be a few seconds old. Simply checking that the current time lies
14557 between thisUpdate and nextUpdate max reject otherwise valid responses
14558 caused by either OCSP responder or client clock inaccuracy. Instead
14559 we allow thisUpdate and nextUpdate to fall within a certain period of
14560 the current time. The age of the response can also optionally be
14561 checked. Two new options -validity_period and -status_age added to
14566 * If signature or public key algorithm is unrecognized print out its
14567 OID rather that just UNKNOWN.
14571 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14572 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14573 ID to be generated from the issuer certificate alone which can then be
14574 passed to OCSP_id_issuer_cmp().
14578 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14579 ASN1 modules to export functions returning ASN1_ITEM pointers
14580 instead of the ASN1_ITEM structures themselves. This adds several
14581 new macros which allow the underlying ASN1 function/structure to
14582 be accessed transparently. As a result code should not use ASN1_ITEM
14583 references directly (such as &X509_it) but instead use the relevant
14584 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14585 use of the new ASN1 code on platforms where exporting structures
14586 is problematical (for example in shared libraries) but exporting
14587 functions returning pointers to structures is not.
14591 * Add support for overriding the generation of SSL/TLS session IDs.
14592 These callbacks can be registered either in an SSL_CTX or per SSL.
14593 The purpose of this is to allow applications to control, if they wish,
14594 the arbitrary values chosen for use as session IDs, particularly as it
14595 can be useful for session caching in multiple-server environments. A
14596 command-line switch for testing this (and any client code that wishes
14597 to use such a feature) has been added to "s_server".
14599 *Geoff Thorpe, Lutz Jaenicke*
14601 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14602 of the form `#if defined(...) || defined(...) || ...` and
14603 `#if !defined(...) && !defined(...) && ...`. This also avoids
14604 the growing number of special cases it was previously handling.
14608 * Make all configuration macros available for application by making
14609 sure they are available in opensslconf.h, by giving them names starting
14610 with `OPENSSL_` to avoid conflicts with other packages and by making
14611 sure e_os2.h will cover all platform-specific cases together with
14613 Additionally, it is now possible to define configuration/platform-
14614 specific names (called "system identities"). In the C code, these
14615 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14616 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14617 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14622 * New option -set_serial to 'req' and 'x509' this allows the serial
14623 number to use to be specified on the command line. Previously self
14624 signed certificates were hard coded with serial number 0 and the
14625 CA options of 'x509' had to use a serial number in a file which was
14630 * New options to 'ca' utility to support V2 CRL entry extensions.
14631 Currently CRL reason, invalidity date and hold instruction are
14632 supported. Add new CRL extensions to V3 code and some new objects.
14636 * New function EVP_CIPHER_CTX_set_padding() this is used to
14637 disable standard block padding (aka PKCS#5 padding) in the EVP
14638 API, which was previously mandatory. This means that the data is
14639 not padded in any way and so the total length much be a multiple
14640 of the block size, otherwise an error occurs.
14644 * Initial (incomplete) OCSP SSL support.
14648 * New function OCSP_parse_url(). This splits up a URL into its host,
14649 port and path components: primarily to parse OCSP URLs. New -url
14650 option to ocsp utility.
14654 * New nonce behavior. The return value of OCSP_check_nonce() now
14655 reflects the various checks performed. Applications can decide
14656 whether to tolerate certain situations such as an absent nonce
14657 in a response when one was present in a request: the ocsp application
14658 just prints out a warning. New function OCSP_add1_basic_nonce()
14659 this is to allow responders to include a nonce in a response even if
14660 the request is nonce-less.
14664 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14665 skipped when using openssl x509 multiple times on a single input file,
14666 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14670 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14671 set string type: to handle setting ASN1_TIME structures. Fix ca
14672 utility to correctly initialize revocation date of CRLs.
14676 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14677 the clients preferred ciphersuites and rather use its own preferences.
14678 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14679 Internet Explorer by ensuring unchanged hash method during stepup.
14680 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14684 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14685 to aes and add a new 'exist' option to print out symbols that don't
14690 * Additional options to ocsp utility to allow flags to be set and
14691 additional certificates supplied.
14695 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14696 OCSP client a number of certificate to only verify the response
14701 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14702 handle the new API. Currently only ECB, CBC modes supported. Add new
14705 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14706 Encryption Standard (AES) Ciphersuites for Transport Layer
14707 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14708 not enabled by default and were not part of the "ALL" ciphersuite
14709 alias because they were not yet official; they could be
14710 explicitly requested by specifying the "AESdraft" ciphersuite
14711 group alias. In the final release of OpenSSL 0.9.7, the group
14712 alias is called "AES" and is part of "ALL".)
14714 *Ben Laurie, Steve Henson, Bodo Moeller*
14716 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14717 request to response.
14721 * Functions for OCSP responders. OCSP_request_onereq_count(),
14722 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14723 extract information from a certificate request. OCSP_response_create()
14724 creates a response and optionally adds a basic response structure.
14725 OCSP_basic_add1_status() adds a complete single response to a basic
14726 response and returns the OCSP_SINGLERESP structure just added (to allow
14727 extensions to be included for example). OCSP_basic_add1_cert() adds a
14728 certificate to a basic response and OCSP_basic_sign() signs a basic
14729 response with various flags. New helper functions ASN1_TIME_check()
14730 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14731 (converts ASN1_TIME to GeneralizedTime).
14735 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14736 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14737 structure from a certificate. X509_pubkey_digest() digests the public_key
14738 contents: this is used in various key identifiers.
14742 * Make sk_sort() tolerate a NULL argument.
14744 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14746 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14747 passed by the function are trusted implicitly. If any of them signed the
14748 response then it is assumed to be valid and is not verified.
14752 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14753 to data. This was previously part of the PKCS7 ASN1 code. This
14754 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14755 *Steve Henson, reported by Kenneth R. Robinette
14756 <support@securenetterm.com>*
14758 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14759 routines: without these tracing memory leaks is very painful.
14760 Fix leaks in PKCS12 and PKCS7 routines.
14764 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14765 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14766 effectively meant GeneralizedTime would never be used. Now it
14767 is initialised to -1 but X509_time_adj() now has to check the value
14768 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14769 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14770 *Steve Henson, reported by Kenneth R. Robinette
14771 <support@securenetterm.com>*
14773 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14774 result in a zero length in the ASN1_INTEGER structure which was
14775 not consistent with the structure when d2i_ASN1_INTEGER() was used
14776 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14777 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14778 where it did not print out a minus for negative ASN1_INTEGER.
14782 * Add summary printout to ocsp utility. The various functions which
14783 convert status values to strings have been renamed to:
14784 OCSP_response_status_str(), OCSP_cert_status_str() and
14785 OCSP_crl_reason_str() and are no longer static. New options
14786 to verify nonce values and to disable verification. OCSP response
14787 printout format cleaned up.
14791 * Add additional OCSP certificate checks. These are those specified
14792 in RFC2560. This consists of two separate checks: the CA of the
14793 certificate being checked must either be the OCSP signer certificate
14794 or the issuer of the OCSP signer certificate. In the latter case the
14795 OCSP signer certificate must contain the OCSP signing extended key
14796 usage. This check is performed by attempting to match the OCSP
14797 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14798 in the OCSP_CERTID structures of the response.
14802 * Initial OCSP certificate verification added to OCSP_basic_verify()
14803 and related routines. This uses the standard OpenSSL certificate
14804 verify routines to perform initial checks (just CA validity) and
14805 to obtain the certificate chain. Then additional checks will be
14806 performed on the chain. Currently the root CA is checked to see
14807 if it is explicitly trusted for OCSP signing. This is used to set
14808 a root CA as a global signing root: that is any certificate that
14809 chains to that CA is an acceptable OCSP signing certificate.
14813 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14814 extensions from a separate configuration file.
14815 As when reading extensions from the main configuration file,
14816 the '-extensions ...' option may be used for specifying the
14819 *Massimiliano Pala <madwolf@comune.modena.it>*
14821 * New OCSP utility. Allows OCSP requests to be generated or
14822 read. The request can be sent to a responder and the output
14823 parsed, outputted or printed in text form. Not complete yet:
14824 still needs to check the OCSP response validity.
14828 * New subcommands for 'openssl ca':
14829 `openssl ca -status <serial>` prints the status of the cert with
14830 the given serial number (according to the index file).
14831 `openssl ca -updatedb` updates the expiry status of certificates
14834 *Massimiliano Pala <madwolf@comune.modena.it>*
14836 * New '-newreq-nodes' command option to CA.pl. This is like
14837 '-newreq', but calls 'openssl req' with the '-nodes' option
14838 so that the resulting key is not encrypted.
14840 *Damien Miller <djm@mindrot.org>*
14842 * New configuration for the GNU Hurd.
14844 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14846 * Initial code to implement OCSP basic response verify. This
14847 is currently incomplete. Currently just finds the signer's
14848 certificate and verifies the signature on the response.
14852 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14853 value of OPENSSLDIR. This is available via the new '-d' option
14854 to 'openssl version', and is also included in 'openssl version -a'.
14858 * Allowing defining memory allocation callbacks that will be given
14859 file name and line number information in additional arguments
14860 (a `const char*` and an int). The basic functionality remains, as
14861 well as the original possibility to just replace malloc(),
14862 realloc() and free() by functions that do not know about these
14863 additional arguments. To register and find out the current
14864 settings for extended allocation functions, the following
14865 functions are provided:
14867 CRYPTO_set_mem_ex_functions
14868 CRYPTO_set_locked_mem_ex_functions
14869 CRYPTO_get_mem_ex_functions
14870 CRYPTO_get_locked_mem_ex_functions
14872 These work the same way as CRYPTO_set_mem_functions and friends.
14873 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14874 extended allocation function is enabled.
14875 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14876 a conventional allocation function is enabled.
14878 *Richard Levitte, Bodo Moeller*
14880 * Finish off removing the remaining LHASH function pointer casts.
14881 There should no longer be any prototype-casting required when using
14882 the LHASH abstraction, and any casts that remain are "bugs". See
14883 the callback types and macros at the head of lhash.h for details
14884 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14888 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14889 If /dev/[u]random devices are not available or do not return enough
14890 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14892 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14893 /etc/entropy will be queried once each in this sequence, querying stops
14894 when enough entropy was collected without querying more sockets.
14898 * Change the Unix RAND_poll() variant to be able to poll several
14899 random devices, as specified by DEVRANDOM, until a sufficient amount
14900 of data has been collected. We spend at most 10 ms on each file
14901 (select timeout) and read in non-blocking mode. DEVRANDOM now
14902 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14903 (previously it was just the string "/dev/urandom"), so on typical
14904 platforms the 10 ms delay will never occur.
14905 Also separate out the Unix variant to its own file, rand_unix.c.
14906 For VMS, there's a currently-empty rand_vms.c.
14910 * Move OCSP client related routines to ocsp_cl.c. These
14911 provide utility functions which an application needing
14912 to issue a request to an OCSP responder and analyse the
14913 response will typically need: as opposed to those which an
14914 OCSP responder itself would need which will be added later.
14916 OCSP_request_sign() signs an OCSP request with an API similar
14917 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14918 response. OCSP_response_get1_basic() extracts basic response
14919 from response. OCSP_resp_find_status(): finds and extracts status
14920 information from an OCSP_CERTID structure (which will be created
14921 when the request structure is built). These are built from lower
14922 level functions which work on OCSP_SINGLERESP structures but
14923 won't normally be used unless the application wishes to examine
14924 extensions in the OCSP response for example.
14926 Replace nonce routines with a pair of functions.
14927 OCSP_request_add1_nonce() adds a nonce value and optionally
14928 generates a random value. OCSP_check_nonce() checks the
14929 validity of the nonce in an OCSP response.
14933 * Change function OCSP_request_add() to OCSP_request_add0_id().
14934 This doesn't copy the supplied OCSP_CERTID and avoids the
14935 need to free up the newly created id. Change return type
14936 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14937 This can then be used to add extensions to the request.
14938 Deleted OCSP_request_new(), since most of its functionality
14939 is now in OCSP_REQUEST_new() (and the case insensitive name
14940 clash) apart from the ability to set the request name which
14941 will be added elsewhere.
14945 * Update OCSP API. Remove obsolete extensions argument from
14946 various functions. Extensions are now handled using the new
14947 OCSP extension code. New simple OCSP HTTP function which
14948 can be used to send requests and parse the response.
14952 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14953 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14954 uses the special reorder version of SET OF to sort the attributes
14955 and reorder them to match the encoded order. This resolves a long
14956 standing problem: a verify on a PKCS7 structure just after signing
14957 it used to fail because the attribute order did not match the
14958 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14959 it uses the received order. This is necessary to tolerate some broken
14960 software that does not order SET OF. This is handled by encoding
14961 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14962 to produce the required SET OF.
14966 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14967 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14968 files to get correct declarations of the ASN.1 item variables.
14972 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14973 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14974 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14975 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14976 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14977 ASN1_ITEM and no wrapper functions.
14981 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14982 replace the old function pointer based I/O routines. Change most of
14983 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14987 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14988 lines, recognize more "algorithms" that can be deselected, and make
14989 it complain about algorithm deselection that isn't recognised.
14993 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14994 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14995 to use new functions. Add NO_ASN1_OLD which can be set to remove
14996 some old style ASN1 functions: this can be used to determine if old
14997 code will still work when these eventually go away.
15001 * New extension functions for OCSP structures, these follow the
15002 same conventions as certificates and CRLs.
15006 * New function X509V3_add1_i2d(). This automatically encodes and
15007 adds an extension. Its behaviour can be customised with various
15008 flags to append, replace or delete. Various wrappers added for
15009 certificates and CRLs.
15013 * Fix to avoid calling the underlying ASN1 print routine when
15014 an extension cannot be parsed. Correct a typo in the
15015 OCSP_SERVICELOC extension. Tidy up print OCSP format.
15019 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
15020 entries for variables.
15024 * Add functionality to `apps/openssl.c` for detecting locking
15025 problems: As the program is single-threaded, all we have
15026 to do is register a locking callback using an array for
15027 storing which locks are currently held by the program.
15031 * Use a lock around the call to CRYPTO_get_ex_new_index() in
15032 SSL_get_ex_data_X509_STORE_idx(), which is used in
15033 ssl_verify_cert_chain() and thus can be called at any time
15034 during TLS/SSL handshakes so that thread-safety is essential.
15035 Unfortunately, the ex_data design is not at all suited
15036 for multi-threaded use, so it probably should be abolished.
15040 * Added Broadcom "ubsec" ENGINE to OpenSSL.
15042 *Broadcom, tweaked and integrated by Geoff Thorpe*
15044 * Move common extension printing code to new function
15045 X509V3_print_extensions(). Reorganise OCSP print routines and
15046 implement some needed OCSP ASN1 functions. Add OCSP extensions.
15050 * New function X509_signature_print() to remove duplication in some
15055 * Add a special meaning when SET OF and SEQUENCE OF flags are both
15056 set (this was treated exactly the same as SET OF previously). This
15057 is used to reorder the STACK representing the structure to match the
15058 encoding. This will be used to get round a problem where a PKCS7
15059 structure which was signed could not be verified because the STACK
15060 order did not reflect the encoded order.
15064 * Reimplement the OCSP ASN1 module using the new code.
15068 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
15069 for its ASN1 operations. The old style function pointers still exist
15070 for now but they will eventually go away.
15074 * Merge in replacement ASN1 code from the ASN1 branch. This almost
15075 completely replaces the old ASN1 functionality with a table driven
15076 encoder and decoder which interprets an ASN1_ITEM structure describing
15077 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
15078 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
15079 has also been converted to the new form.
15083 * Change BN_mod_exp_recp so that negative moduli are tolerated
15084 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
15085 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
15086 for negative moduli.
15090 * Fix BN_uadd and BN_usub: Always return non-negative results instead
15091 of not touching the result's sign bit.
15095 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
15100 * Changed the LHASH code to use prototypes for callbacks, and created
15101 macros to declare and implement thin (optionally static) functions
15102 that provide type-safety and avoid function pointer casting for the
15103 type-specific callbacks.
15107 * Added Kerberos Cipher Suites to be used with TLS, as written in
15109 *Veers Staats <staatsvr@asc.hpc.mil>,
15110 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
15112 * Reformat the FAQ so the different questions and answers can be divided
15113 in sections depending on the subject.
15117 * Have the zlib compression code load ZLIB.DLL dynamically under
15122 * New function BN_mod_sqrt for computing square roots modulo a prime
15123 (using the probabilistic Tonelli-Shanks algorithm unless
15124 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
15125 be handled deterministically).
15127 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15129 * Make BN_mod_inverse faster by explicitly handling small quotients
15130 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
15131 512 bits], about 30% for larger ones [1024 or 2048 bits].)
15135 * New function BN_kronecker.
15139 * Fix BN_gcd so that it works on negative inputs; the result is
15140 positive unless both parameters are zero.
15141 Previously something reasonably close to an infinite loop was
15142 possible because numbers could be growing instead of shrinking
15143 in the implementation of Euclid's algorithm.
15147 * Fix BN_is_word() and BN_is_one() macros to take into account the
15148 sign of the number in question.
15150 Fix BN_is_word(a,w) to work correctly for w == 0.
15152 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
15153 because its test if the absolute value of 'a' equals 'w'.
15154 Note that BN_abs_is_word does *not* handle w == 0 reliably;
15155 it exists mostly for use in the implementations of BN_is_zero(),
15156 BN_is_one(), and BN_is_word().
15160 * New function BN_swap.
15164 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
15165 the exponentiation functions are more likely to produce reasonable
15166 results on negative inputs.
15170 * Change BN_mod_mul so that the result is always non-negative.
15171 Previously, it could be negative if one of the factors was negative;
15172 I don't think anyone really wanted that behaviour.
15176 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
15177 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
15178 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
15179 and add new functions:
15188 BN_mod_lshift1_quick
15190 BN_mod_lshift_quick
15192 These functions always generate non-negative results.
15194 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15195 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15197 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15198 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15199 be reduced modulo `m`.
15201 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15204 The following entry accidentally appeared in the CHANGES file
15205 distributed with OpenSSL 0.9.7. The modifications described in
15206 it do *not* apply to OpenSSL 0.9.7.
15208 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15209 was actually never needed) and in BN_mul(). The removal in BN_mul()
15210 required a small change in bn_mul_part_recursive() and the addition
15211 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15212 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15213 bn_sub_words() and bn_add_words() except they take arrays with
15219 * In 'openssl passwd', verify passwords read from the terminal
15220 unless the '-salt' option is used (which usually means that
15221 verification would just waste user's time since the resulting
15222 hash is going to be compared with some given password hash)
15223 or the new '-noverify' option is used.
15225 This is an incompatible change, but it does not affect
15226 non-interactive use of 'openssl passwd' (passwords on the command
15227 line, '-stdin' option, '-in ...' option) and thus should not
15228 cause any problems.
15232 * Remove all references to RSAref, since there's no more need for it.
15236 * Make DSO load along a path given through an environment variable
15237 (SHLIB_PATH) with shl_load().
15241 * Constify the ENGINE code as a result of BIGNUM constification.
15242 Also constify the RSA code and most things related to it. In a
15243 few places, most notable in the depth of the ASN.1 code, ugly
15244 casts back to non-const were required (to be solved at a later
15249 * Make it so the openssl application has all engines loaded by default.
15253 * Constify the BIGNUM routines a little more.
15257 * Add the following functions:
15259 ENGINE_load_cswift()
15261 ENGINE_load_atalla()
15262 ENGINE_load_nuron()
15263 ENGINE_load_builtin_engines()
15265 That way, an application can itself choose if external engines that
15266 are built-in in OpenSSL shall ever be used or not. The benefit is
15267 that applications won't have to be linked with libdl or other dso
15268 libraries unless it's really needed.
15270 Changed 'openssl engine' to load all engines on demand.
15271 Changed the engine header files to avoid the duplication of some
15272 declarations (they differed!).
15276 * 'openssl engine' can now list capabilities.
15280 * Better error reporting in 'openssl engine'.
15284 * Never call load_dh_param(NULL) in s_server.
15288 * Add engine application. It can currently list engines by name and
15289 identity, and test if they are actually available.
15293 * Improve RPM specification file by forcing symbolic linking and making
15294 sure the installed documentation is also owned by root.root.
15296 *Damien Miller <djm@mindrot.org>*
15298 * Give the OpenSSL applications more possibilities to make use of
15299 keys (public as well as private) handled by engines.
15303 * Add OCSP code that comes from CertCo.
15307 * Add VMS support for the Rijndael code.
15311 * Added untested support for Nuron crypto accelerator.
15315 * Add support for external cryptographic devices. This code was
15316 previously distributed separately as the "engine" branch.
15318 *Geoff Thorpe, Richard Levitte*
15320 * Rework the filename-translation in the DSO code. It is now possible to
15321 have far greater control over how a "name" is turned into a filename
15322 depending on the operating environment and any oddities about the
15323 different shared library filenames on each system.
15327 * Support threads on FreeBSD-elf in Configure.
15331 * Fix for SHA1 assembly problem with MASM: it produces
15332 warnings about corrupt line number information when assembling
15333 with debugging information. This is caused by the overlapping
15336 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15339 NCONF_get_number() has no error checking at all. As a replacement,
15340 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15341 promoted strongly. The old NCONF_get_number is kept around for
15342 binary backward compatibility.
15343 Make it possible for methods to load from something other than a BIO,
15344 by providing a function pointer that is given a name instead of a BIO.
15345 For example, this could be used to load configuration data from an
15350 * Fix for non blocking accept BIOs. Added new I/O special reason
15351 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15352 with non blocking I/O was not possible because no retry code was
15353 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15358 * Added the beginnings of Rijndael support.
15362 * Fix for bug in DirectoryString mask setting. Add support for
15363 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15364 to allow certificate printing to more controllable, additional
15365 'certopt' option to 'x509' to allow new printing options to be
15370 * Clean old EAY MD5 hack from e_os.h.
15374 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15376 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15377 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15379 *Joe Orton, Steve Henson*
15381 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15383 * Fix additional bug revealed by the NISCC test suite:
15385 Stop bug triggering large recursion when presented with
15386 certain ASN.1 tags ([CVE-2003-0851])
15390 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15392 * Fix various bugs revealed by running the NISCC test suite:
15394 Stop out of bounds reads in the ASN1 code when presented with
15395 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15397 If verify callback ignores invalid public key errors don't try to check
15398 certificate signature with the NULL public key.
15402 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15403 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15408 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15409 extra data after the compression methods not only for TLS 1.0
15410 but also for SSL 3.0 (as required by the specification).
15412 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15414 * Change X509_certificate_type() to mark the key as exported/exportable
15415 when it's 512 *bits* long, not 512 bytes.
15419 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15421 * Countermeasure against the Klima-Pokorny-Rosa extension of
15422 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15423 a protocol version number mismatch like a decryption error
15424 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15428 * Turn on RSA blinding by default in the default implementation
15429 to avoid a timing attack. Applications that don't want it can call
15430 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15431 They would be ill-advised to do so in most cases.
15433 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15435 * Change RSA blinding code so that it works when the PRNG is not
15436 seeded (in this case, the secret RSA exponent is abused as
15437 an unpredictable seed -- if it is not unpredictable, there
15438 is no point in blinding anyway). Make RSA blinding thread-safe
15439 by remembering the creator's thread ID in rsa->blinding and
15440 having all other threads use local one-time blinding factors
15441 (this requires more computation than sharing rsa->blinding, but
15442 avoids excessive locking; and if an RSA object is not shared
15443 between threads, blinding will still be very fast).
15447 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15449 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15450 via timing by performing a MAC computation even if incorrect
15451 block cipher padding has been found. This is a countermeasure
15452 against active attacks where the attacker has to distinguish
15453 between bad padding and a MAC verification error. ([CVE-2003-0078])
15455 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15456 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15457 Martin Vuagnoux (EPFL, Ilion)*
15459 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15461 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15462 memory from its contents. This is done with a counter that will
15463 place alternating values in each byte. This can be used to solve
15464 two issues: 1) the removal of calls to memset() by highly optimizing
15465 compilers, and 2) cleansing with other values than 0, since those can
15466 be read through on certain media, for example a swap space on disk.
15470 * Bugfix: client side session caching did not work with external caching,
15471 because the session->cipher setting was not restored when reloading
15472 from the external cache. This problem was masked, when
15473 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15474 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15478 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15479 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15481 *Zeev Lieber <zeev-l@yahoo.com>*
15483 * Undo an undocumented change introduced in 0.9.6e which caused
15484 repeated calls to OpenSSL_add_all_ciphers() and
15485 OpenSSL_add_all_digests() to be ignored, even after calling
15490 * Change the default configuration reader to deal with last line not
15491 being properly terminated.
15495 * Change X509_NAME_cmp() so it applies the special rules on handling
15496 DN values that are of type PrintableString, as well as RDNs of type
15497 emailAddress where the value has the type ia5String.
15499 *stefank@valicert.com via Richard Levitte*
15501 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15502 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15503 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15504 the bitwise-OR of the two for use by the majority of applications
15505 wanting this behaviour, and update the docs. The documented
15506 behaviour and actual behaviour were inconsistent and had been
15507 changing anyway, so this is more a bug-fix than a behavioural
15510 *Geoff Thorpe, diagnosed by Nadav Har'El*
15512 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15513 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15517 * Fix initialization code race conditions in
15518 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15519 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15520 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15521 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15522 ssl2_get_cipher_by_char(),
15523 ssl3_get_cipher_by_char().
15525 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15527 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15528 the cached sessions are flushed, as the remove_cb() might use ex_data
15529 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15530 (see [openssl.org #212]).
15532 *Geoff Thorpe, Lutz Jaenicke*
15534 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15535 length, instead of the encoding length to d2i_ASN1_OBJECT.
15539 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15541 * [In 0.9.6g-engine release:]
15542 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15544 *Lynn Gazis <lgazis@rainbow.com>*
15546 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15548 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15549 and get fix the header length calculation.
15550 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15551 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15553 * Use proper error handling instead of 'assertions' in buffer
15554 overflow checks added in 0.9.6e. This prevents DoS (the
15555 assertions could call abort()).
15557 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15559 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15561 * Add various sanity checks to asn1_get_length() to reject
15562 the ASN1 length bytes if they exceed sizeof(long), will appear
15563 negative or the content length exceeds the length of the
15566 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15568 * Fix cipher selection routines: ciphers without encryption had no flags
15569 for the cipher strength set and where therefore not handled correctly
15570 by the selection routines (PR #130).
15574 * Fix EVP_dsa_sha macro.
15579 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15580 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15581 that was added in OpenSSL 0.9.6d.
15583 As the countermeasure turned out to be incompatible with some
15584 broken SSL implementations, the new option is part of SSL_OP_ALL.
15585 SSL_OP_ALL is usually employed when compatibility with weird SSL
15586 implementations is desired (e.g. '-bugs' option to 's_client' and
15587 's_server'), so the new option is automatically set in many
15592 * Changes in security patch:
15594 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15595 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15596 Air Force Materiel Command, USAF, under agreement number
15599 * Add various sanity checks to asn1_get_length() to reject
15600 the ASN1 length bytes if they exceed sizeof(long), will appear
15601 negative or the content length exceeds the length of the
15602 supplied buffer. ([CVE-2002-0659])
15604 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15606 * Assertions for various potential buffer overflows, not known to
15607 happen in practice.
15609 *Ben Laurie (CHATS)*
15611 * Various temporary buffers to hold ASCII versions of integers were
15612 too small for 64 bit platforms. ([CVE-2002-0655])
15613 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15615 * Remote buffer overflow in SSL3 protocol - an attacker could
15616 supply an oversized session ID to a client. ([CVE-2002-0656])
15618 *Ben Laurie (CHATS)*
15620 * Remote buffer overflow in SSL2 protocol - an attacker could
15621 supply an oversized client master key. ([CVE-2002-0656])
15623 *Ben Laurie (CHATS)*
15625 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15627 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15628 encoded as NULL) with id-dsa-with-sha1.
15630 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15632 * Check various `X509_...()` return values in `apps/req.c`.
15634 *Nils Larsch <nla@trustcenter.de>*
15636 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15637 an end-of-file condition would erroneously be flagged, when the CRLF
15638 was just at the end of a processed block. The bug was discovered when
15639 processing data through a buffering memory BIO handing the data to a
15640 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15641 <ptsekov@syntrex.com> and Nedelcho Stanev.
15645 * Implement a countermeasure against a vulnerability recently found
15646 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15647 before application data chunks to avoid the use of known IVs
15648 with data potentially chosen by the attacker.
15652 * Fix length checks in ssl3_get_client_hello().
15656 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15657 to prevent ssl3_read_internal() from incorrectly assuming that
15658 ssl3_read_bytes() found application data while handshake
15659 processing was enabled when in fact s->s3->in_read_app_data was
15660 merely automatically cleared during the initial handshake.
15662 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15664 * Fix object definitions for Private and Enterprise: they were not
15665 recognized in their shortname (=lowercase) representation. Extend
15666 obj_dat.pl to issue an error when using undefined keywords instead
15667 of silently ignoring the problem (Svenning Sorensen
15668 <sss@sss.dnsalias.net>).
15672 * Fix DH_generate_parameters() so that it works for 'non-standard'
15673 generators, i.e. generators other than 2 and 5. (Previously, the
15674 code did not properly initialise the 'add' and 'rem' values to
15675 BN_generate_prime().)
15677 In the new general case, we do not insist that 'generator' is
15678 actually a primitive root: This requirement is rather pointless;
15679 a generator of the order-q subgroup is just as good, if not
15684 * Map new X509 verification errors to alerts. Discovered and submitted by
15685 Tom Wu <tom@arcot.com>.
15689 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15690 returning non-zero before the data has been completely received
15691 when using non-blocking I/O.
15693 *Bodo Moeller; problem pointed out by John Hughes*
15695 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15697 *Ben Laurie, Lutz Jaenicke*
15699 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15700 Yoram Zahavi <YoramZ@gilian.com>).
15704 * Add information about CygWin 1.3 and on, and preserve proper
15705 configuration for the versions before that.
15707 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15709 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15710 check whether we deal with a copy of a session and do not delete from
15711 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15712 <izhar@checkpoint.com>.
15716 * Do not store session data into the internal session cache, if it
15717 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15718 flag is set). Proposed by Aslam <aslam@funk.com>.
15722 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15727 * [In 0.9.6d-engine release:]
15728 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15730 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15732 * Add the configuration target linux-s390x.
15734 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15736 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15737 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15738 variable as an indication that a ClientHello message has been
15739 received. As the flag value will be lost between multiple
15740 invocations of ssl3_accept when using non-blocking I/O, the
15741 function may not be aware that a handshake has actually taken
15742 place, thus preventing a new session from being added to the
15745 To avoid this problem, we now set s->new_session to 2 instead of
15746 using a local variable.
15748 *Lutz Jaenicke, Bodo Moeller*
15750 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15751 if the SSL_R_LENGTH_MISMATCH error is detected.
15753 *Geoff Thorpe, Bodo Moeller*
15755 * New 'shared_ldflag' column in Configure platform table.
15759 * Fix EVP_CIPHER_mode macro.
15761 *"Dan S. Camper" <dan@bti.net>*
15763 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15764 type, we must throw them away by setting rr->length to 0.
15766 *D P Chang <dpc@qualys.com>*
15768 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15770 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15771 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15772 worked incorrectly for those cases where range = `10..._2` and
15773 `3*range` is two bits longer than range.)
15777 * Only add signing time to PKCS7 structures if it is not already
15782 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15783 OBJ_ld_ce should be OBJ_id_ce.
15784 Also some ip-pda OIDs in crypto/objects/objects.txt were
15785 incorrect (cf. RFC 3039).
15787 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15789 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15790 returns early because it has nothing to do.
15792 *Andy Schneider <andy.schneider@bjss.co.uk>*
15794 * [In 0.9.6c-engine release:]
15795 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15797 *Andy Schneider <andy.schneider@bjss.co.uk>*
15799 * [In 0.9.6c-engine release:]
15800 Add support for Cryptographic Appliance's keyserver technology.
15801 (Use engine 'keyclient')
15803 *Cryptographic Appliances and Geoff Thorpe*
15805 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15806 is called via tools/c89.sh because arguments have to be
15807 rearranged (all '-L' options must appear before the first object
15810 *Richard Shapiro <rshapiro@abinitio.com>*
15812 * [In 0.9.6c-engine release:]
15813 Add support for Broadcom crypto accelerator cards, backported
15816 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15818 * [In 0.9.6c-engine release:]
15819 Add support for SureWare crypto accelerator cards from
15820 Baltimore Technologies. (Use engine 'sureware')
15822 *Baltimore Technologies and Mark Cox*
15824 * [In 0.9.6c-engine release:]
15825 Add support for crypto accelerator cards from Accelerated
15826 Encryption Processing, www.aep.ie. (Use engine 'aep')
15828 *AEP Inc. and Mark Cox*
15830 * Add a configuration entry for gcc on UnixWare.
15832 *Gary Benson <gbenson@redhat.com>*
15834 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15835 messages are stored in a single piece (fixed-length part and
15836 variable-length part combined) and fix various bugs found on the way.
15840 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15841 instead. BIO_gethostbyname() does not know what timeouts are
15842 appropriate, so entries would stay in cache even when they have
15844 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15846 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15847 faced with a pathologically small ClientHello fragment that does
15848 not contain client_version: Instead of aborting with an error,
15849 simply choose the highest available protocol version (i.e.,
15850 TLS 1.0 unless it is disabled). In practice, ClientHello
15851 messages are never sent like this, but this change gives us
15852 strictly correct behaviour at least for TLS.
15856 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15857 never resets s->method to s->ctx->method when called from within
15858 one of the SSL handshake functions.
15860 *Bodo Moeller; problem pointed out by Niko Baric*
15862 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15863 (sent using the client's version number) if client_version is
15864 smaller than the protocol version in use. Also change
15865 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15866 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15867 the client will at least see that alert.
15871 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15876 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15877 client receives HelloRequest while in a handshake.
15879 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15881 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15882 should end in 'break', not 'goto end' which circumvents various
15883 cleanups done in state SSL_ST_OK. But session related stuff
15884 must be disabled for SSL_ST_OK in the case that we just sent a
15887 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15888 before just sending a HelloRequest.
15890 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15892 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15893 reveal whether illegal block cipher padding was found or a MAC
15894 verification error occurred. (Neither SSLerr() codes nor alerts
15895 are directly visible to potential attackers, but the information
15896 may leak via logfiles.)
15898 Similar changes are not required for the SSL 2.0 implementation
15899 because the number of padding bytes is sent in clear for SSL 2.0,
15900 and the extra bytes are just ignored. However ssl/s2_pkt.c
15901 failed to verify that the purported number of padding bytes is in
15906 * Add OpenUNIX-8 support including shared libraries
15907 (Boyd Lynn Gerber <gerberb@zenez.com>).
15911 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15912 'wristwatch attack' using huge encoding parameters (cf.
15913 James H. Manger's CRYPTO 2001 paper). Note that the
15914 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15915 encoding parameters and hence was not vulnerable.
15919 * BN_sqr() bug fix.
15921 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15923 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15924 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15925 followed by modular reduction.
15927 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15929 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15930 equivalent based on BN_pseudo_rand() instead of BN_rand().
15934 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15935 This function was broken, as the check for a new client hello message
15936 to handle SGC did not allow these large messages.
15937 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15941 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15945 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15946 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15950 * Rework the configuration and shared library support for Tru64 Unix.
15951 The configuration part makes use of modern compiler features and
15952 still retains old compiler behavior for those that run older versions
15953 of the OS. The shared library support part includes a variant that
15954 uses the RPATH feature, and is available through the special
15955 configuration target "alpha-cc-rpath", which will never be selected
15958 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15960 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15961 with the same message size as in ssl3_get_certificate_request().
15962 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15963 messages might inadvertently be reject as too long.
15965 *Petr Lampa <lampa@fee.vutbr.cz>*
15967 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15971 * Modified SSL library such that the verify_callback that has been set
15972 specifically for an SSL object with SSL_set_verify() is actually being
15973 used. Before the change, a verify_callback set with this function was
15974 ignored and the verify_callback() set in the SSL_CTX at the time of
15975 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15976 to allow the necessary settings.
15980 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15981 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15982 done automatically (in contradiction to the requirements of the C
15983 standard). This made problems when used from OpenSSH.
15987 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15988 dh->length and always used
15990 BN_rand_range(priv_key, dh->p).
15992 BN_rand_range() is not necessary for Diffie-Hellman, and this
15993 specific range makes Diffie-Hellman unnecessarily inefficient if
15994 dh->length (recommended exponent length) is much smaller than the
15995 length of dh->p. We could use BN_rand_range() if the order of
15996 the subgroup was stored in the DH structure, but we only have
16001 BN_rand(priv_key, l, ...)
16003 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
16010 RSA_eay_public_encrypt
16011 RSA_eay_private_decrypt
16012 RSA_eay_private_encrypt (signing)
16013 RSA_eay_public_decrypt (signature verification)
16015 (default implementations for RSA_public_encrypt,
16016 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
16017 always reject numbers >= n.
16021 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
16022 to synchronize access to 'locking_thread'. This is necessary on
16023 systems where access to 'locking_thread' (an 'unsigned long'
16024 variable) is not atomic.
16028 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
16029 *before* setting the 'crypto_lock_rand' flag. The previous code had
16030 a race condition if 0 is a valid thread ID.
16032 *Travis Vitek <vitek@roguewave.com>*
16034 * Add support for shared libraries under Irix.
16036 *Albert Chin-A-Young <china@thewrittenword.com>*
16038 * Add configuration option to build on Linux on both big-endian and
16039 little-endian MIPS.
16041 *Ralf Baechle <ralf@uni-koblenz.de>*
16043 * Add the possibility to create shared libraries on HP-UX.
16047 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
16049 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
16050 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
16051 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
16052 PRNG state recovery was possible based on the output of
16053 one PRNG request appropriately sized to gain knowledge on
16054 'md' followed by enough consecutive 1-byte PRNG requests
16055 to traverse all of 'state'.
16057 1. When updating 'md_local' (the current thread's copy of 'md')
16058 during PRNG output generation, hash all of the previous
16059 'md_local' value, not just the half used for PRNG output.
16061 2. Make the number of bytes from 'state' included into the hash
16062 independent from the number of PRNG bytes requested.
16064 The first measure alone would be sufficient to avoid
16065 Markku-Juhani's attack. (Actually it had never occurred
16066 to me that the half of 'md_local' used for chaining was the
16067 half from which PRNG output bytes were taken -- I had always
16068 assumed that the secret half would be used.) The second
16069 measure makes sure that additional data from 'state' is never
16070 mixed into 'md_local' in small portions; this heuristically
16071 further strengthens the PRNG.
16075 * Fix crypto/bn/asm/mips3.s.
16079 * When only the key is given to "enc", the IV is undefined. Print out
16080 an error message in this case.
16084 * Handle special case when X509_NAME is empty in X509 printing routines.
16088 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
16089 positive and less than q.
16093 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
16094 used: it isn't thread safe and the add_lock_callback should handle
16097 *Paul Rose <Paul.Rose@bridge.com>*
16099 * Verify that incoming data obeys the block size in
16100 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
16106 *Ulf Möller, Bodo Möller*
16108 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
16109 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
16110 when fixing the server behaviour for backwards-compatible 'client
16111 hello' messages. (Note that the attack is impractical against
16112 SSL 3.0 and TLS 1.0 anyway because length and version checking
16113 means that the probability of guessing a valid ciphertext is
16114 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
16117 Before 0.9.5, the countermeasure (hide the error by generating a
16118 random 'decryption result') did not work properly because
16119 ERR_clear_error() was missing, meaning that SSL_get_error() would
16120 detect the supposedly ignored error.
16122 Both problems are now fixed.
16126 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
16127 (previously it was 1024).
16131 * Fix for compatibility mode trust settings: ignore trust settings
16132 unless some valid trust or reject settings are present.
16136 * Fix for blowfish EVP: its a variable length cipher.
16140 * Fix various bugs related to DSA S/MIME verification. Handle missing
16141 parameters in DSA public key structures and return an error in the
16142 DSA routines if parameters are absent.
16146 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
16147 in the current directory if neither $RANDFILE nor $HOME was set.
16148 RAND_file_name() in 0.9.6a returned NULL in this case. This has
16149 caused some confusion to Windows users who haven't defined $HOME.
16150 Thus RAND_file_name() is changed again: e_os.h can define a
16151 DEFAULT_HOME, which will be used if $HOME is not set.
16152 For Windows, we use "C:"; on other platforms, we still require
16153 environment variables.
16155 * Move 'if (!initialized) RAND_poll()' into regions protected by
16156 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
16157 having multiple threads call RAND_poll() concurrently.
16161 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
16162 combination of a flag and a thread ID variable.
16163 Otherwise while one thread is in ssleay_rand_bytes (which sets the
16164 flag), *other* threads can enter ssleay_add_bytes without obeying
16165 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
16166 that they do not hold after the first thread unsets add_do_not_lock).
16170 * Change bctest again: '-x' expressions are not available in all
16171 versions of 'test'.
16175 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
16177 * Fix a couple of memory leaks in PKCS7_dataDecode()
16179 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
16181 * Change Configure and Makefiles to provide EXE_EXT, which will contain
16182 the default extension for executables, if any. Also, make the perl
16183 scripts that use symlink() to test if it really exists and use "cp"
16184 if it doesn't. All this made OpenSSL compilable and installable in
16189 * Fix for asn1_GetSequence() for indefinite length constructed data.
16190 If SEQUENCE is length is indefinite just set c->slen to the total
16191 amount of data available.
16193 *Steve Henson, reported by shige@FreeBSD.org*
16195 *This change does not apply to 0.9.7.*
16197 * Change bctest to avoid here-documents inside command substitution
16198 (workaround for FreeBSD /bin/sh bug).
16199 For compatibility with Ultrix, avoid shell functions (introduced
16200 in the bctest version that searches along $PATH).
16204 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16205 with des_encrypt() defined on some operating systems, like Solaris
16210 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16211 On the Importance of Eliminating Errors in Cryptographic
16212 Computations, J. Cryptology 14 (2001) 2, 101-119,
16213 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16217 * MIPS assembler BIGNUM division bug fix.
16221 * Disabled incorrect Alpha assembler code.
16225 * Fix PKCS#7 decode routines so they correctly update the length
16226 after reading an EOC for the EXPLICIT tag.
16230 *This change does not apply to 0.9.7.*
16232 * Fix bug in PKCS#12 key generation routines. This was triggered
16233 if a 3DES key was generated with a 0 initial byte. Include
16234 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16235 (but broken) behaviour.
16239 * Enhance bctest to search for a working bc along $PATH and print
16242 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16244 * Fix memory leaks in err.c: free err_data string if necessary;
16245 don't write to the wrong index in ERR_set_error_data.
16249 * Implement ssl23_peek (analogous to ssl23_read), which previously
16254 * Replace rdtsc with `_emit` statements for VC++ version 5.
16256 *Jeremy Cooper <jeremy@baymoo.org>*
16258 * Make it possible to reuse SSLv2 sessions.
16262 * In copy_email() check for >= 0 as a return value for
16263 X509_NAME_get_index_by_NID() since 0 is a valid index.
16265 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16267 * Avoid coredump with unsupported or invalid public keys by checking if
16268 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16269 PKCS7_verify() fails with non detached data.
16273 * Don't use getenv in library functions when run as setuid/setgid.
16274 New function OPENSSL_issetugid().
16278 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16279 due to incorrect handling of multi-threading:
16281 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16283 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16285 3. Count how many times MemCheck_off() has been called so that
16286 nested use can be treated correctly. This also avoids
16287 inband-signalling in the previous code (which relied on the
16288 assumption that thread ID 0 is impossible).
16292 * Add "-rand" option also to s_client and s_server.
16296 * Fix CPU detection on Irix 6.x.
16297 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16298 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16300 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16305 *This change does not apply to 0.9.7.*
16307 * Use the cached encoding of an X509_NAME structure rather than
16308 copying it. This is apparently the reason for the libsafe "errors"
16309 but the code is actually correct.
16313 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16314 Bleichenbacher's DSA attack.
16315 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16316 to be set and top=0 forces the highest bit to be set; top=-1 is new
16317 and leaves the highest bit random.
16319 *Ulf Moeller, Bodo Moeller*
16321 * In the `NCONF_...`-based implementations for `CONF_...` queries
16322 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16323 a temporary CONF structure with the data component set to NULL
16324 (which gives segmentation faults in lh_retrieve).
16325 Instead, use NULL for the CONF pointer in CONF_get_string and
16326 CONF_get_number (which may use environment variables) and directly
16327 return NULL from CONF_get_section.
16331 * Fix potential buffer overrun for EBCDIC.
16335 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16336 keyUsage if basicConstraints absent for a CA.
16340 * Make SMIME_write_PKCS7() write mail header values with a format that
16341 is more generally accepted (no spaces before the semicolon), since
16342 some programs can't parse those values properly otherwise. Also make
16343 sure BIO's that break lines after each write do not create invalid
16348 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16349 macros previously used would not encode an empty SEQUENCE OF
16350 and break the signature.
16354 *This change does not apply to 0.9.7.*
16356 * Zero the premaster secret after deriving the master secret in
16361 * Add some EVP_add_digest_alias registrations (as found in
16362 OpenSSL_add_all_digests()) to SSL_library_init()
16363 aka OpenSSL_add_ssl_algorithms(). This provides improved
16364 compatibility with peers using X.509 certificates
16365 with unconventional AlgorithmIdentifier OIDs.
16369 * Fix for Irix with NO_ASM.
16371 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16373 * ./config script fixes.
16375 *Ulf Moeller, Richard Levitte*
16377 * Fix 'openssl passwd -1'.
16381 * Change PKCS12_key_gen_asc() so it can cope with non null
16382 terminated strings whose length is passed in the passlen
16383 parameter, for example from PEM callbacks. This was done
16384 by adding an extra length parameter to asc2uni().
16386 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16388 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16389 call failed, free the DSA structure.
16393 * Fix to uni2asc() to cope with zero length Unicode strings.
16394 These are present in some PKCS#12 files.
16398 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16399 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16400 when writing a 32767 byte record.
16402 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16404 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16405 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16407 (RSA objects have a reference count access to which is protected
16408 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16409 so they are meant to be shared between threads.)
16410 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16411 "Reddie, Steven" <Steven.Reddie@ca.com>*
16413 * Fix a deadlock in CRYPTO_mem_leaks().
16417 * Use better test patterns in bntest.
16421 * rand_win.c fix for Borland C.
16425 * BN_rshift bugfix for n == 0.
16429 * Add a 'bctest' script that checks for some known 'bc' bugs
16430 so that 'make test' does not abort just because 'bc' is broken.
16434 * Store verify_result within SSL_SESSION also for client side to
16435 avoid potential security hole. (Re-used sessions on the client side
16436 always resulted in verify_result==X509_V_OK, not using the original
16437 result of the server certificate verification.)
16441 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16442 SSL3_RT_APPLICATION_DATA, return 0.
16443 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16448 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16449 releases, have been re-implemented by renaming the previous
16450 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16451 and ssl3_read_internal, respectively, and adding 'peek' parameters
16452 to them. The new ssl[23]_{read,peek} functions are calls to
16453 ssl[23]_read_internal with the 'peek' flag set appropriately.
16454 A 'peek' parameter has also been added to ssl3_read_bytes, which
16455 does the actual work for ssl3_read_internal.
16459 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16460 the method-specific "init()" handler. Also clean up ex_data after
16461 calling the method-specific "finish()" handler. Previously, this was
16462 happening the other way round.
16466 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16467 The previous value, 12, was not always sufficient for BN_mod_exp().
16471 * Make sure that shared libraries get the internal name engine with
16472 the full version number and not just 0. This should mark the
16473 shared libraries as not backward compatible. Of course, this should
16474 be changed again when we can guarantee backward binary compatibility.
16478 * Fix typo in get_cert_by_subject() in by_dir.c
16480 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16482 * Rework the system to generate shared libraries:
16484 - Make note of the expected extension for the shared libraries and
16485 if there is a need for symbolic links from for example libcrypto.so.0
16486 to libcrypto.so.0.9.7. There is extended info in Configure for
16489 - Make as few rebuilds of the shared libraries as possible.
16491 - Still avoid linking the OpenSSL programs with the shared libraries.
16493 - When installing, install the shared libraries separately from the
16498 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16500 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16501 and not in SSL_clear because the latter is also used by the
16502 accept/connect functions; previously, the settings made by
16503 SSL_set_read_ahead would be lost during the handshake.
16505 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16507 * Correct util/mkdef.pl to be selective about disabled algorithms.
16508 Previously, it would create entries for disabled algorithms no
16513 * Added several new manual pages for SSL_* function.
16517 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16519 * In ssl23_get_client_hello, generate an error message when faced
16520 with an initial SSL 3.0/TLS record that is too small to contain the
16521 first two bytes of the ClientHello message, i.e. client_version.
16522 (Note that this is a pathologic case that probably has never happened
16523 in real life.) The previous approach was to use the version number
16524 from the record header as a substitute; but our protocol choice
16525 should not depend on that one because it is not authenticated
16526 by the Finished messages.
16530 * More robust randomness gathering functions for Windows.
16532 *Jeffrey Altman <jaltman@columbia.edu>*
16534 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16535 not set then we don't setup the error code for issuer check errors
16536 to avoid possibly overwriting other errors which the callback does
16537 handle. If an application does set the flag then we assume it knows
16538 what it is doing and can handle the new informational codes
16543 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16544 a general "ANY" type, as such it should be able to decode anything
16545 including tagged types. However it didn't check the class so it would
16546 wrongly interpret tagged types in the same way as their universal
16547 counterpart and unknown types were just rejected. Changed so that the
16548 tagged and unknown types are handled in the same way as a SEQUENCE:
16549 that is the encoding is stored intact. There is also a new type
16550 "V_ASN1_OTHER" which is used when the class is not universal, in this
16551 case we have no idea what the actual type is so we just lump them all
16556 * On VMS, stdout may very well lead to a file that is written to
16557 in a record-oriented fashion. That means that every write() will
16558 write a separate record, which will be read separately by the
16559 programs trying to read from it. This can be very confusing.
16561 The solution is to put a BIO filter in the way that will buffer
16562 text until a linefeed is reached, and then write everything a
16563 line at a time, so every record written will be an actual line,
16564 not chunks of lines and not (usually doesn't happen, but I've
16565 seen it once) several lines in one record. BIO_f_linebuffer() is
16568 Currently, it's a VMS-only method, because that's where it has
16569 been tested well enough.
16573 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16574 it can return incorrect results.
16575 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16576 but it was in 0.9.6-beta[12].)
16580 * Disable the check for content being present when verifying detached
16581 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16582 include zero length content when signing messages.
16586 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16587 BIO_ctrl (for BIO pairs).
16591 * Add DSO method for VMS.
16595 * Bug fix: Montgomery multiplication could produce results with the
16600 * Add RPM specification openssl.spec and modify it to build three
16601 packages. The default package contains applications, application
16602 documentation and run-time libraries. The devel package contains
16603 include files, static libraries and function documentation. The
16604 doc package contains the contents of the doc directory. The original
16605 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16609 * Add a large number of documentation files for many SSL routines.
16611 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16613 * Add a configuration entry for Sony News 4.
16615 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16617 * Don't set the two most significant bits to one when generating a
16618 random number < q in the DSA library.
16622 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16623 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16624 the underlying transport is blocking) if a handshake took place.
16625 (The default behaviour is needed by applications such as s_client
16626 and s_server that use select() to determine when to use SSL_read;
16627 but for applications that know in advance when to expect data, it
16628 just makes things more complicated.)
16632 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16637 * Add a few more EBCDIC conditionals that make `req` and `x509`
16638 work better on such systems.
16640 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16642 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16643 Update PKCS12_parse() so it copies the friendlyName and the
16644 keyid to the certificates aux info.
16648 * Fix bug in PKCS7_verify() which caused an infinite loop
16649 if there was more than one signature.
16651 *Sven Uszpelkat <su@celocom.de>*
16653 * Major change in util/mkdef.pl to include extra information
16654 about each symbol, as well as presenting variables as well
16655 as functions. This change means that there's n more need
16656 to rebuild the .num files when some algorithms are excluded.
16660 * Allow the verify time to be set by an application,
16661 rather than always using the current time.
16665 * Phase 2 verify code reorganisation. The certificate
16666 verify code now looks up an issuer certificate by a
16667 number of criteria: subject name, authority key id
16668 and key usage. It also verifies self signed certificates
16669 by the same criteria. The main comparison function is
16670 X509_check_issued() which performs these checks.
16672 Lot of changes were necessary in order to support this
16673 without completely rewriting the lookup code.
16675 Authority and subject key identifier are now cached.
16677 The LHASH 'certs' is X509_STORE has now been replaced
16678 by a STACK_OF(X509_OBJECT). This is mainly because an
16679 LHASH can't store or retrieve multiple objects with
16680 the same hash value.
16682 As a result various functions (which were all internal
16683 use only) have changed to handle the new X509_STORE
16684 structure. This will break anything that messed round
16685 with X509_STORE internally.
16687 The functions X509_STORE_add_cert() now checks for an
16688 exact match, rather than just subject name.
16690 The X509_STORE API doesn't directly support the retrieval
16691 of multiple certificates matching a given criteria, however
16692 this can be worked round by performing a lookup first
16693 (which will fill the cache with candidate certificates)
16694 and then examining the cache for matches. This is probably
16695 the best we can do without throwing out X509_LOOKUP
16696 entirely (maybe later...).
16698 The X509_VERIFY_CTX structure has been enhanced considerably.
16700 All certificate lookup operations now go via a get_issuer()
16701 callback. Although this currently uses an X509_STORE it
16702 can be replaced by custom lookups. This is a simple way
16703 to bypass the X509_STORE hackery necessary to make this
16704 work and makes it possible to use more efficient techniques
16705 in future. A very simple version which uses a simple
16706 STACK for its trusted certificate store is also provided
16707 using X509_STORE_CTX_trusted_stack().
16709 The verify_cb() and verify() callbacks now have equivalents
16710 in the X509_STORE_CTX structure.
16712 X509_STORE_CTX also has a 'flags' field which can be used
16713 to customise the verify behaviour.
16717 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16718 excludes S/MIME capabilities.
16722 * When a certificate request is read in keep a copy of the
16723 original encoding of the signed data and use it when outputting
16724 again. Signatures then use the original encoding rather than
16725 a decoded, encoded version which may cause problems if the
16726 request is improperly encoded.
16730 * For consistency with other BIO_puts implementations, call
16731 buffer_write(b, ...) directly in buffer_puts instead of calling
16734 In BIO_puts, increment b->num_write as in BIO_write.
16736 *Peter.Sylvester@EdelWeb.fr*
16738 * Fix BN_mul_word for the case where the word is 0. (We have to use
16739 BN_zero, we may not return a BIGNUM with an array consisting of
16740 words set to zero.)
16744 * Avoid calling abort() from within the library when problems are
16745 detected, except if preprocessor symbols have been defined
16746 (such as REF_CHECK, BN_DEBUG etc.).
16750 * New openssl application 'rsautl'. This utility can be
16751 used for low-level RSA operations. DER public key
16752 BIO/fp routines also added.
16756 * New Configure entry and patches for compiling on QNX 4.
16758 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16760 * A demo state-machine implementation was sponsored by
16761 Nuron (<http://www.nuron.com/>) and is now available in
16762 demos/state_machine.
16766 * New options added to the 'dgst' utility for signature
16767 generation and verification.
16771 * Unrecognized PKCS#7 content types are now handled via a
16772 catch all ASN1_TYPE structure. This allows unsupported
16773 types to be stored as a "blob" and an application can
16774 encode and decode it manually.
16778 * Fix various signed/unsigned issues to make a_strex.c
16779 compile under VC++.
16781 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16783 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16784 length if passed a buffer. ASN1_INTEGER_to_BN failed
16785 if passed a NULL BN and its argument was negative.
16787 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16789 * Modification to PKCS#7 encoding routines to output definite
16790 length encoding. Since currently the whole structures are in
16791 memory there's not real point in using indefinite length
16792 constructed encoding. However if OpenSSL is compiled with
16793 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16797 * Added BIO_vprintf() and BIO_vsnprintf().
16801 * Added more prefixes to parse for in the strings written
16802 through a logging bio, to cover all the levels that are available
16803 through syslog. The prefixes are now:
16805 PANIC, EMERG, EMR => LOG_EMERG
16806 ALERT, ALR => LOG_ALERT
16807 CRIT, CRI => LOG_CRIT
16808 ERROR, ERR => LOG_ERR
16809 WARNING, WARN, WAR => LOG_WARNING
16810 NOTICE, NOTE, NOT => LOG_NOTICE
16811 INFO, INF => LOG_INFO
16812 DEBUG, DBG => LOG_DEBUG
16814 and as before, if none of those prefixes are present at the
16815 beginning of the string, LOG_ERR is chosen.
16817 On Win32, the `LOG_*` levels are mapped according to this:
16819 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16820 LOG_WARNING => EVENTLOG_WARNING_TYPE
16821 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16825 * Made it possible to reconfigure with just the configuration
16826 argument "reconf" or "reconfigure". The command line arguments
16827 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16828 and are retrieved from there when reconfiguring.
16834 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16836 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16840 * The obj_dat.pl script was messing up the sorting of object
16841 names. The reason was that it compared the quoted version
16842 of strings as a result "OCSP" > "OCSP Signing" because
16843 " > SPACE. Changed script to store unquoted versions of
16844 names and add quotes on output. It was also omitting some
16845 names from the lookup table if they were given a default
16846 value (that is if SN is missing it is given the same
16847 value as LN and vice versa), these are now added on the
16848 grounds that if an object has a name we should be able to
16849 look it up. Finally added warning output when duplicate
16850 short or long names are found.
16854 * Changes needed for Tandem NSK.
16856 *Scott Uroff <scott@xypro.com>*
16858 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16859 RSA_padding_check_SSLv23(), special padding was never detected
16860 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16861 version rollback attacks was not effective.
16863 In s23_clnt.c, don't use special rollback-attack detection padding
16864 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16865 client; similarly, in s23_srvr.c, don't do the rollback check if
16866 SSL 2.0 is the only protocol enabled in the server.
16870 * Make it possible to get hexdumps of unprintable data with 'openssl
16871 asn1parse'. By implication, the functions ASN1_parse_dump() and
16872 BIO_dump_indent() are added.
16876 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16877 these print out strings and name structures based on various
16878 flags including RFC2253 support and proper handling of
16879 multibyte characters. Added options to the 'x509' utility
16880 to allow the various flags to be set.
16884 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16885 Also change the functions X509_cmp_current_time() and
16886 X509_gmtime_adj() work with an ASN1_TIME structure,
16887 this will enable certificates using GeneralizedTime in validity
16888 dates to be checked.
16892 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16893 negative public key encodings) on by default,
16894 NO_NEG_PUBKEY_BUG can be set to disable it.
16898 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16899 content octets. An i2c_ASN1_OBJECT is unnecessary because
16900 the encoding can be trivially obtained from the structure.
16904 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16905 not read locks (`CRYPTO_r_[un]lock`).
16909 * A first attempt at creating official support for shared
16910 libraries through configuration. I've kept it so the
16911 default is static libraries only, and the OpenSSL programs
16912 are always statically linked for now, but there are
16913 preparations for dynamic linking in place.
16914 This has been tested on Linux and Tru64.
16918 * Randomness polling function for Win9x, as described in:
16919 Peter Gutmann, Software Generation of Practically Strong
16924 * Fix so PRNG is seeded in req if using an already existing
16929 * New options to smime application. -inform and -outform
16930 allow alternative formats for the S/MIME message including
16931 PEM and DER. The -content option allows the content to be
16932 specified separately. This should allow things like Netscape
16933 form signing output easier to verify.
16937 * Fix the ASN1 encoding of tags using the 'long form'.
16941 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16942 STRING types. These convert content octets to and from the
16943 underlying type. The actual tag and length octets are
16944 already assumed to have been read in and checked. These
16945 are needed because all other string types have virtually
16946 identical handling apart from the tag. By having versions
16947 of the ASN1 functions that just operate on content octets
16948 IMPLICIT tagging can be handled properly. It also allows
16949 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16950 and ASN1_INTEGER are identical apart from the tag.
16954 * Change the handling of OID objects as follows:
16956 - New object identifiers are inserted in objects.txt, following
16957 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16958 - objects.pl is used to process obj_mac.num and create a new
16960 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16963 This is currently kind of a hack, and the perl code in objects.pl
16964 isn't very elegant, but it works as I intended. The simplest way
16965 to check that it worked correctly is to look in obj_dat.h and
16966 check the array nid_objs and make sure the objects haven't moved
16967 around (this is important!). Additions are OK, as well as
16968 consistent name changes.
16972 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16976 * Addition of the command line parameter '-rand file' to 'openssl req'.
16977 The given file adds to whatever has already been seeded into the
16978 random pool through the RANDFILE configuration file option or
16979 environment variable, or the default random state file.
16983 * mkstack.pl now sorts each macro group into lexical order.
16984 Previously the output order depended on the order the files
16985 appeared in the directory, resulting in needless rewriting
16990 * Patches to make OpenSSL compile under Win32 again. Mostly
16991 work arounds for the VC++ problem that it treats func() as
16992 func(void). Also stripped out the parts of mkdef.pl that
16993 added extra typesafe functions: these no longer exist.
16997 * Reorganisation of the stack code. The macros are now all
16998 collected in safestack.h . Each macro is defined in terms of
16999 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
17000 DEBUG_SAFESTACK is now handled in terms of function casts,
17001 this has the advantage of retaining type safety without the
17002 use of additional functions. If DEBUG_SAFESTACK is not defined
17003 then the non typesafe macros are used instead. Also modified the
17004 mkstack.pl script to handle the new form. Needs testing to see
17005 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
17006 the default if no major problems. Similar behaviour for ASN1_SET_OF
17007 and PKCS12_STACK_OF.
17011 * When some versions of IIS use the 'NET' form of private key the
17012 key derivation algorithm is different. Normally MD5(password) is
17013 used as a 128 bit RC4 key. In the modified case
17014 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
17015 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
17016 as the old Netscape_RSA functions except they have an additional
17017 'sgckey' parameter which uses the modified algorithm. Also added
17018 an -sgckey command line option to the rsa utility. Thanks to
17019 Adrian Peck <bertie@ncipher.com> for posting details of the modified
17020 algorithm to openssl-dev.
17024 * The evp_local.h macros were using 'c.##kname' which resulted in
17025 invalid expansion on some systems (SCO 5.0.5 for example).
17026 Corrected to 'c.kname'.
17028 *Phillip Porch <root@theporch.com>*
17030 * New X509_get1_email() and X509_REQ_get1_email() functions that return
17031 a STACK of email addresses from a certificate or request, these look
17032 in the subject name and the subject alternative name extensions and
17033 omit any duplicate addresses.
17037 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
17038 This makes DSA verification about 2 % faster.
17042 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
17043 (meaning that now 2^5 values will be precomputed, which is only 4 KB
17044 plus overhead for 1024 bit moduli).
17045 This makes exponentiations about 0.5 % faster for 1024 bit
17046 exponents (as measured by "openssl speed rsa2048").
17050 * Rename memory handling macros to avoid conflicts with other
17052 Malloc => OPENSSL_malloc
17053 Malloc_locked => OPENSSL_malloc_locked
17054 Realloc => OPENSSL_realloc
17055 Free => OPENSSL_free
17059 * New function BN_mod_exp_mont_word for small bases (roughly 15%
17060 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
17064 * CygWin32 support.
17066 *John Jarvie <jjarvie@newsguy.com>*
17068 * The type-safe stack code has been rejigged. It is now only compiled
17069 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
17070 by default all type-specific stack functions are "#define"d back to
17071 standard stack functions. This results in more streamlined output
17072 but retains the type-safety checking possibilities of the original
17077 * The STACK code has been cleaned up, and certain type declarations
17078 that didn't make a lot of sense have been brought in line. This has
17079 also involved a cleanup of sorts in safestack.h to more correctly
17080 map type-safe stack functions onto their plain stack counterparts.
17081 This work has also resulted in a variety of "const"ifications of
17082 lots of the code, especially `_cmp` operations which should normally
17083 be prototyped with "const" parameters anyway.
17087 * When generating bytes for the first time in md_rand.c, 'stir the pool'
17088 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
17089 (The PRNG state consists of two parts, the large pool 'state' and 'md',
17090 where all of 'md' is used each time the PRNG is used, but 'state'
17091 is used only indexed by a cyclic counter. As entropy may not be
17092 well distributed from the beginning, 'md' is important as a
17093 chaining variable. However, the output function chains only half
17094 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
17095 all of 'md', and seeding with STATE_SIZE dummy bytes will result
17096 in all of 'state' being rewritten, with the new values depending
17097 on virtually all of 'md'. This overcomes the 80 bit limitation.)
17101 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
17102 the handshake is continued after ssl_verify_cert_chain();
17103 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
17104 can lead to 'unexplainable' connection aborts later.
17106 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
17108 * Major EVP API cipher revision.
17109 Add hooks for extra EVP features. This allows various cipher
17110 parameters to be set in the EVP interface. Support added for variable
17111 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
17112 setting of RC2 and RC5 parameters.
17114 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
17117 Remove lots of duplicated code from the EVP library. For example *every*
17118 cipher init() function handles the 'iv' in the same way according to the
17119 cipher mode. They also all do nothing if the 'key' parameter is NULL and
17120 for CFB and OFB modes they zero ctx->num.
17122 New functionality allows removal of S/MIME code RC2 hack.
17124 Most of the routines have the same form and so can be declared in terms
17127 By shifting this to the top level EVP_CipherInit() it can be removed from
17128 all individual ciphers. If the cipher wants to handle IVs or keys
17129 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
17132 Change lots of functions like EVP_EncryptUpdate() to now return a
17133 value: although software versions of the algorithms cannot fail
17134 any installed hardware versions can.
17138 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
17139 this option is set, tolerate broken clients that send the negotiated
17140 protocol version number instead of the requested protocol version
17145 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
17146 i.e. non-zero for export ciphersuites, zero otherwise.
17147 Previous versions had this flag inverted, inconsistent with
17148 rsa_tmp_cb (..._TMP_RSA_CB).
17150 *Bodo Moeller; problem reported by Amit Chopra*
17152 * Add missing DSA library text string. Work around for some IIS
17153 key files with invalid SEQUENCE encoding.
17157 * Add a document (doc/standards.txt) that list all kinds of standards
17158 and so on that are implemented in OpenSSL.
17162 * Enhance c_rehash script. Old version would mishandle certificates
17163 with the same subject name hash and wouldn't handle CRLs at all.
17164 Added -fingerprint option to crl utility, to support new c_rehash
17169 * Eliminate non-ANSI declarations in crypto.h and stack.h.
17173 * Fix for SSL server purpose checking. Server checking was
17174 rejecting certificates which had extended key usage present
17175 but no ssl client purpose.
17177 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
17179 * Make PKCS#12 code work with no password. The PKCS#12 spec
17180 is a little unclear about how a blank password is handled.
17181 Since the password in encoded as a BMPString with terminating
17182 double NULL a zero length password would end up as just the
17183 double NULL. However no password at all is different and is
17184 handled differently in the PKCS#12 key generation code. NS
17185 treats a blank password as zero length. MSIE treats it as no
17186 password on export: but it will try both on import. We now do
17187 the same: PKCS12_parse() tries zero length and no password if
17188 the password is set to "" or NULL (NULL is now a valid password:
17189 it wasn't before) as does the pkcs12 application.
17193 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17194 perror when PEM_read_bio_X509_REQ fails, the error message must
17195 be obtained from the error queue.
17199 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17200 it in ERR_remove_state if appropriate, and change ERR_get_state
17201 accordingly to avoid race conditions (this is necessary because
17202 thread_hash is no longer constant once set).
17206 * Bugfix for linux-elf makefile.one.
17210 * RSA_get_default_method() will now cause a default
17211 RSA_METHOD to be chosen if one doesn't exist already.
17212 Previously this was only set during a call to RSA_new()
17213 or RSA_new_method(NULL) meaning it was possible for
17214 RSA_get_default_method() to return NULL.
17218 * Added native name translation to the existing DSO code
17219 that will convert (if the flag to do so is set) filenames
17220 that are sufficiently small and have no path information
17221 into a canonical native form. Eg. "blah" converted to
17222 "libblah.so" or "blah.dll" etc.
17226 * New function ERR_error_string_n(e, buf, len) which is like
17227 ERR_error_string(e, buf), but writes at most 'len' bytes
17228 including the 0 terminator. For ERR_error_string_n, 'buf'
17231 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17233 * CONF library reworked to become more general. A new CONF
17234 configuration file reader "class" is implemented as well as a
17235 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17236 old `CONF_*` functions are still there, but are reimplemented to
17237 work in terms of the new functions. Also, a set of functions
17238 to handle the internal storage of the configuration data is
17239 provided to make it easier to write new configuration file
17240 reader "classes" (I can definitely see something reading a
17241 configuration file in XML format, for example), called `_CONF_*`,
17242 or "the configuration storage API"...
17244 The new configuration file reading functions are:
17246 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17247 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17249 NCONF_default, NCONF_WIN32
17251 NCONF_dump_fp, NCONF_dump_bio
17253 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17254 NCONF_new creates a new CONF object. This works in the same way
17255 as other interfaces in OpenSSL, like the BIO interface.
17256 `NCONF_dump_*` dump the internal storage of the configuration file,
17257 which is useful for debugging. All other functions take the same
17258 arguments as the old `CONF_*` functions with the exception of the
17259 first that must be a `CONF *` instead of a `LHASH *`.
17261 To make it easier to use the new classes with the old `CONF_*` functions,
17262 the function CONF_set_default_method is provided.
17266 * Add '-tls1' option to 'openssl ciphers', which was already
17267 mentioned in the documentation but had not been implemented.
17268 (This option is not yet really useful because even the additional
17269 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17273 * Initial DSO code added into libcrypto for letting OpenSSL (and
17274 OpenSSL-based applications) load shared libraries and bind to
17275 them in a portable way.
17277 *Geoff Thorpe, with contributions from Richard Levitte*
17279 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17281 * Make sure _lrotl and _lrotr are only used with MSVC.
17283 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17284 (the default implementation of RAND_status).
17286 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17287 to '-clrext' (= clear extensions), as intended and documented.
17288 *Bodo Moeller; inconsistency pointed out by Michael Attili
17289 <attili@amaxo.com>*
17291 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17292 was larger than the MD block size.
17294 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17296 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17297 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17298 using the passed key: if the passed key was a private key the result
17299 of X509_print(), for example, would be to print out all the private key
17304 * des_quad_cksum() byte order bug fix.
17305 *Ulf Möller, using the problem description in krb4-0.9.7, where
17306 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17308 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17311 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17313 * For easily testing in shell scripts whether some command
17314 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17315 returns with exit code 0 iff no command of the given name is available.
17316 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17317 the output goes to stdout and nothing is printed to stderr.
17318 Additional arguments are always ignored.
17320 Since for each cipher there is a command of the same name,
17321 the 'no-cipher' compilation switches can be tested this way.
17323 ('openssl no-XXX' is not able to detect pseudo-commands such
17324 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17328 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17332 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17333 is set; it will be thrown away anyway because each handshake creates
17335 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17336 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17337 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17338 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17342 * New s_client option -ign_eof: EOF at stdin is ignored, and
17343 'Q' and 'R' lose their special meanings (quit/renegotiate).
17344 This is part of what -quiet does; unlike -quiet, -ign_eof
17345 does not suppress any output.
17349 * Add compatibility options to the purpose and trust code. The
17350 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17351 accepts a certificate or CA, this was the previous behaviour,
17352 with all the associated security issues.
17354 X509_TRUST_COMPAT is the old trust behaviour: only and
17355 automatically trust self signed roots in certificate store. A
17356 new trust setting X509_TRUST_DEFAULT is used to specify that
17357 a purpose has no associated trust setting and it should instead
17358 use the value in the default purpose.
17362 * Fix the PKCS#8 DSA private key code so it decodes keys again
17363 and fix a memory leak.
17367 * In util/mkerr.pl (which implements 'make errors'), preserve
17368 reason strings from the previous version of the .c file, as
17369 the default to have only downcase letters (and digits) in
17370 automatically generated reasons codes is not always appropriate.
17374 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17375 using strerror. Previously, ERR_reason_error_string() returned
17376 library names as reason strings for SYSerr; but SYSerr is a special
17377 case where small numbers are errno values, not library numbers.
17381 * Add '-dsaparam' option to 'openssl dhparam' application. This
17382 converts DSA parameters into DH parameters. (When creating parameters,
17383 DSA_generate_parameters is used.)
17387 * Include 'length' (recommended exponent length) in C code generated
17388 by 'openssl dhparam -C'.
17392 * The second argument to set_label in perlasm was already being used
17393 so couldn't be used as a "file scope" flag. Moved to third argument
17398 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17399 instead of RAND_bytes for encryption IVs and salts.
17403 * Include RAND_status() into RAND_METHOD instead of implementing
17404 it only for md_rand.c Otherwise replacing the PRNG by calling
17405 RAND_set_rand_method would be impossible.
17409 * Don't let DSA_generate_key() enter an infinite loop if the random
17410 number generation fails.
17414 * New 'rand' application for creating pseudo-random output.
17418 * Added configuration support for Linux/IA64
17420 *Rolf Haberrecker <rolf@suse.de>*
17422 * Assembler module support for Mingw32.
17426 * Shared library support for HPUX (in shlib/).
17428 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17430 * Shared library support for Solaris gcc.
17432 *Lutz Behnke <behnke@trustcenter.de>*
17434 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17436 * PKCS7_encrypt() was adding text MIME headers twice because they
17437 were added manually and by SMIME_crlf_copy().
17441 * In bntest.c don't call BN_rand with zero bits argument.
17443 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17445 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17446 case was implemented. This caused BN_div_recp() to fail occasionally.
17450 * Add an optional second argument to the set_label() in the perl
17451 assembly language builder. If this argument exists and is set
17452 to 1 it signals that the assembler should use a symbol whose
17453 scope is the entire file, not just the current function. This
17454 is needed with MASM which uses the format label:: for this scope.
17456 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17458 * Change the ASN1 types so they are typedefs by default. Before
17459 almost all types were #define'd to ASN1_STRING which was causing
17460 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17465 * Change names of new functions to the new get1/get0 naming
17466 convention: After 'get1', the caller owns a reference count
17467 and has to call `..._free`; 'get0' returns a pointer to some
17468 data structure without incrementing reference counters.
17469 (Some of the existing 'get' functions increment a reference
17470 counter, some don't.)
17471 Similarly, 'set1' and 'add1' functions increase reference
17472 counters or duplicate objects.
17476 * Allow for the possibility of temp RSA key generation failure:
17477 the code used to assume it always worked and crashed on failure.
17481 * Fix potential buffer overrun problem in BIO_printf().
17482 *Ulf Möller, using public domain code by Patrick Powell; problem
17483 pointed out by David Sacerdote <das33@cornell.edu>*
17485 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17486 RAND_egd() and RAND_status(). In the command line application,
17487 the EGD socket can be specified like a seed file using RANDFILE
17492 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17493 Some CAs (e.g. Verisign) distribute certificates in this form.
17497 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17498 list to exclude them. This means that no special compilation option
17499 is needed to use anonymous DH: it just needs to be included in the
17504 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17505 EVP_MD_type. The old functionality is available in a new macro called
17506 EVP_MD_md(). Change code that uses it and update docs.
17510 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17511 where the `void *` argument is replaced by a function pointer argument.
17512 Previously `void *` was abused to point to functions, which works on
17513 many platforms, but is not correct. As these functions are usually
17514 called by macros defined in OpenSSL header files, most source code
17515 should work without changes.
17519 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17520 sections with information on -D... compiler switches used for
17521 compiling the library so that applications can see them. To enable
17522 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17523 must be defined. E.g.,
17524 #define OPENSSL_ALGORITHM_DEFINES
17525 #include <openssl/opensslconf.h>
17526 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17528 *Richard Levitte, Ulf and Bodo Möller*
17530 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17535 * Change the 'other' type in certificate aux info to a STACK_OF
17536 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17537 the required ASN1 format: arbitrary types determined by an OID.
17541 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17542 argument to 'req'. This is not because the function is newer or
17543 better than others it just uses the work 'NEW' in the certificate
17544 request header lines. Some software needs this.
17548 * Reorganise password command line arguments: now passwords can be
17549 obtained from various sources. Delete the PEM_cb function and make
17550 it the default behaviour: i.e. if the callback is NULL and the
17551 usrdata argument is not NULL interpret it as a null terminated pass
17552 phrase. If usrdata and the callback are NULL then the pass phrase
17553 is prompted for as usual.
17557 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17558 the support is automatically enabled. The resulting binaries will
17559 autodetect the card and use it if present.
17561 *Ben Laurie and Compaq Inc.*
17563 * Work around for Netscape hang bug. This sends certificate request
17564 and server done in one record. Since this is perfectly legal in the
17565 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17566 the bugs/SSLv3 entry for more info.
17570 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17574 * Add -rand argument to smime and pkcs12 applications and read/write
17579 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17583 * Add command line password options to the remaining applications.
17587 * Bug fix for BN_div_recp() for numerators with an even number of
17592 * More tests in bntest.c, and changed test_bn output.
17596 * ./config recognizes MacOS X now.
17600 * Bug fix for BN_div() when the first words of num and divisor are
17601 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17605 * Add support for various broken PKCS#8 formats, and command line
17606 options to produce them.
17610 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17611 get temporary BIGNUMs from a BN_CTX.
17615 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17620 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17621 include a #define from the old name to the new. The original intent
17622 was that statically linked binaries could for example just call
17623 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17624 link with digests. This never worked because SSLeay_add_all_digests()
17625 and SSLeay_add_all_ciphers() were in the same source file so calling
17626 one would link with the other. They are now in separate source files.
17630 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17634 * Use a less unusual form of the Miller-Rabin primality test (it used
17635 a binary algorithm for exponentiation integrated into the Miller-Rabin
17636 loop, our standard modexp algorithms are faster).
17640 * Support for the EBCDIC character set completed.
17642 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17644 * Source code cleanups: use const where appropriate, eliminate casts,
17645 use `void *` instead of `char *` in lhash.
17649 * Bugfix: ssl3_send_server_key_exchange was not restartable
17650 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17651 this the server could overwrite ephemeral keys that the client
17656 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17657 using 50 iterations of the Rabin-Miller test.
17659 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17660 iterations of the Rabin-Miller test as required by the appendix
17661 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17662 As BN_is_prime_fasttest includes trial division, DSA parameter
17663 generation becomes much faster.
17665 This implies a change for the callback functions in DSA_is_prime
17666 and DSA_generate_parameters: The callback function is called once
17667 for each positive witness in the Rabin-Miller test, not just
17668 occasionally in the inner loop; and the parameters to the
17669 callback function now provide an iteration count for the outer
17670 loop rather than for the current invocation of the inner loop.
17671 DSA_generate_parameters additionally can call the callback
17672 function with an 'iteration count' of -1, meaning that a
17673 candidate has passed the trial division test (when q is generated
17674 from an application-provided seed, trial division is skipped).
17678 * New function BN_is_prime_fasttest that optionally does trial
17679 division before starting the Rabin-Miller test and has
17680 an additional BN_CTX * argument (whereas BN_is_prime always
17681 has to allocate at least one BN_CTX).
17682 'callback(1, -1, cb_arg)' is called when a number has passed the
17683 trial division stage.
17687 * Fix for bug in CRL encoding. The validity dates weren't being handled
17692 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17696 * New function BN_pseudo_rand().
17700 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17701 bignum version of BN_from_montgomery() with the working code from
17702 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17707 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17708 made it impossible to use the same SSL_SESSION data structure in
17709 SSL2 clients in multiple threads.
17713 * The return value of RAND_load_file() no longer counts bytes obtained
17714 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17715 to seed the PRNG (previously an explicit byte count was required).
17717 *Ulf Möller, Bodo Möller*
17719 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17720 used `char *` instead of `void *` and had casts all over the place.
17724 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17728 * Retain source code compatibility for BN_prime_checks macro:
17729 BN_is_prime(..., BN_prime_checks, ...) now uses
17730 BN_prime_checks_for_size to determine the appropriate number of
17731 Rabin-Miller iterations.
17735 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17736 DH_CHECK_P_NOT_SAFE_PRIME.
17737 (Check if this is true? OpenPGP calls them "strong".)
17741 * Merge the functionality of "dh" and "gendh" programs into a new program
17742 "dhparam". The old programs are retained for now but will handle DH keys
17743 (instead of parameters) in future.
17747 * Make the ciphers, s_server and s_client programs check the return values
17748 when a new cipher list is set.
17752 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17753 ciphers. Before when the 56bit ciphers were enabled the sorting was
17756 The syntax for the cipher sorting has been extended to support sorting by
17757 cipher-strength (using the strength_bits hard coded in the tables).
17758 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17760 Fix a bug in the cipher-command parser: when supplying a cipher command
17761 string with an "undefined" symbol (neither command nor alphanumeric
17762 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17763 an error is flagged.
17765 Due to the strength-sorting extension, the code of the
17766 ssl_create_cipher_list() function was completely rearranged. I hope that
17767 the readability was also increased :-)
17769 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17771 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17772 for the first serial number and places 2 in the serial number file. This
17773 avoids problems when the root CA is created with serial number zero and
17774 the first user certificate has the same issuer name and serial number
17779 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17780 the new code. Add documentation for this stuff.
17784 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17785 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17786 structures and behave in an analogous way to the X509v3 functions:
17787 they shouldn't be called directly but wrapper functions should be used
17790 So we also now have some wrapper functions that call the X509at functions
17791 when passed certificate requests. (TO DO: similar things can be done with
17792 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17793 things. Some of these need some d2i or i2d and print functionality
17794 because they handle more complex structures.)
17798 * Add missing #ifndefs that caused missing symbols when building libssl
17799 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17800 NO_RSA in `ssl/s2*.c`.
17802 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17804 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17805 has a return value which indicates the quality of the random data
17806 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17807 error queue. New function RAND_pseudo_bytes() generates output that is
17808 guaranteed to be unique but not unpredictable. RAND_add is like
17809 RAND_seed, but takes an extra argument for an entropy estimate
17810 (RAND_seed always assumes full entropy).
17814 * Do more iterations of Rabin-Miller probable prime test (specifically,
17815 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17816 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17817 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17818 false-positive rate of at most 2^-80 for random input.
17822 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17826 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17827 in the 0.9.5 release), this returns the chain
17828 from an X509_CTX structure with a dup of the stack and all
17829 the X509 reference counts upped: so the stack will exist
17830 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17833 Also make SSL_SESSION_print() print out the verify return
17838 * Add manpage for the pkcs12 command. Also change the default
17839 behaviour so MAC iteration counts are used unless the new
17840 -nomaciter option is used. This improves file security and
17841 only older versions of MSIE (4.0 for example) need it.
17845 * Honor the no-xxx Configure options when creating .DEF files.
17849 * Add PKCS#10 attributes to field table: challengePassword,
17850 unstructuredName and unstructuredAddress. These are taken from
17851 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17852 international characters are used.
17854 More changes to X509_ATTRIBUTE code: allow the setting of types
17855 based on strings. Remove the 'loc' parameter when adding
17856 attributes because these will be a SET OF encoding which is sorted
17861 * Initial changes to the 'req' utility to allow request generation
17862 automation. This will allow an application to just generate a template
17863 file containing all the field values and have req construct the
17866 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17867 used all over the place including certificate requests and PKCS#7
17868 structures. They are currently handled manually where necessary with
17869 some primitive wrappers for PKCS#7. The new functions behave in a
17870 manner analogous to the X509 extension functions: they allow
17871 attributes to be looked up by NID and added.
17873 Later something similar to the X509V3 code would be desirable to
17874 automatically handle the encoding, decoding and printing of the
17875 more complex types. The string types like challengePassword can
17876 be handled by the string table functions.
17878 Also modified the multi byte string table handling. Now there is
17879 a 'global mask' which masks out certain types. The table itself
17880 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17881 is useful when for example there is only one permissible type
17882 (as in countryName) and using the mask might result in no valid
17887 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17888 SSL_get_peer_finished to allow applications to obtain the latest
17889 Finished messages sent to the peer or expected from the peer,
17890 respectively. (SSL_get_peer_finished is usually the Finished message
17891 actually received from the peer, otherwise the protocol will be aborted.)
17893 As the Finished message are message digests of the complete handshake
17894 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17895 be used for external authentication procedures when the authentication
17896 provided by SSL/TLS is not desired or is not enough.
17900 * Enhanced support for Alpha Linux is added. Now ./config checks if
17901 the host supports BWX extension and if Compaq C is present on the
17902 $PATH. Just exploiting of the BWX extension results in 20-30%
17903 performance kick for some algorithms, e.g. DES and RC4 to mention
17904 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17909 * Add support for MS "fast SGC". This is arguably a violation of the
17910 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17911 weak crypto and after checking the certificate is SGC a second one
17912 with strong crypto. MS SGC stops the first handshake after receiving
17913 the server certificate message and sends a second client hello. Since
17914 a server will typically do all the time consuming operations before
17915 expecting any further messages from the client (server key exchange
17916 is the most expensive) there is little difference between the two.
17918 To get OpenSSL to support MS SGC we have to permit a second client
17919 hello message after we have sent server done. In addition we have to
17920 reset the MAC if we do get this second client hello.
17924 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17925 if a DER encoded private key is RSA or DSA traditional format. Changed
17926 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17927 format DER encoded private key. Newer code should use PKCS#8 format which
17928 has the key type encoded in the ASN1 structure. Added DER private key
17929 support to pkcs8 application.
17933 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17934 ciphersuites has been selected (as required by the SSL 3/TLS 1
17935 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17936 is set, we interpret this as a request to violate the specification
17937 (the worst that can happen is a handshake failure, and 'correct'
17938 behaviour would result in a handshake failure anyway).
17942 * In SSL_CTX_add_session, take into account that there might be multiple
17943 SSL_SESSION structures with the same session ID (e.g. when two threads
17944 concurrently obtain them from an external cache).
17945 The internal cache can handle only one SSL_SESSION with a given ID,
17946 so if there's a conflict, we now throw out the old one to achieve
17951 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17952 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17953 some routines that use cipher OIDs: some ciphers do not have OIDs
17954 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17959 * Simplify the trust setting structure and code. Now we just have
17960 two sequences of OIDs for trusted and rejected settings. These will
17961 typically have values the same as the extended key usage extension
17962 and any application specific purposes.
17964 The trust checking code now has a default behaviour: it will just
17965 check for an object with the same NID as the passed id. Functions can
17966 be provided to override either the default behaviour or the behaviour
17967 for a given id. SSL client, server and email already have functions
17968 in place for compatibility: they check the NID and also return "trusted"
17969 if the certificate is self signed.
17973 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17974 traditional format into an EVP_PKEY structure.
17978 * Add a password callback function PEM_cb() which either prompts for
17979 a password if usr_data is NULL or otherwise assumes it is a null
17980 terminated password. Allow passwords to be passed on command line
17981 environment or config files in a few more utilities.
17985 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17986 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17987 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17988 Update documentation.
17992 * Support for ASN1 "NULL" type. This could be handled before by using
17993 ASN1_TYPE but there wasn't any function that would try to read a NULL
17994 and produce an error if it couldn't. For compatibility we also have
17995 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17996 don't allocate anything because they don't need to.
18000 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
18003 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
18005 * Rebuild of the memory allocation routines used by OpenSSL code and
18006 possibly others as well. The purpose is to make an interface that
18007 provide hooks so anyone can build a separate set of allocation and
18008 deallocation routines to be used by OpenSSL, for example memory
18009 pool implementations, or something else, which was previously hard
18010 since Malloc(), Realloc() and Free() were defined as macros having
18011 the values malloc, realloc and free, respectively (except for Win32
18012 compilations). The same is provided for memory debugging code.
18013 OpenSSL already comes with functionality to find memory leaks, but
18014 this gives people a chance to debug other memory problems.
18016 With these changes, a new set of functions and macros have appeared:
18018 CRYPTO_set_mem_debug_functions() [F]
18019 CRYPTO_get_mem_debug_functions() [F]
18020 CRYPTO_dbg_set_options() [F]
18021 CRYPTO_dbg_get_options() [F]
18022 CRYPTO_malloc_debug_init() [M]
18024 The memory debug functions are NULL by default, unless the library
18025 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
18026 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
18027 gives the standard debugging functions that come with OpenSSL) or
18028 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
18029 provided by the library user) must be used. When the standard
18030 debugging functions are used, CRYPTO_dbg_set_options can be used to
18031 request additional information:
18032 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
18033 the CRYPTO_MDEBUG_xxx macro when compiling the library.
18035 Also, things like CRYPTO_set_mem_functions will always give the
18036 expected result (the new set of functions is used for allocation
18037 and deallocation) at all times, regardless of platform and compiler
18040 To finish it up, some functions that were never use in any other
18041 way than through macros have a new API and new semantic:
18043 CRYPTO_dbg_malloc()
18044 CRYPTO_dbg_realloc()
18047 All macros of value have retained their old syntax.
18049 *Richard Levitte and Bodo Moeller*
18051 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
18052 ordering of SMIMECapabilities wasn't in "strength order" and there
18053 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
18058 * Some ASN1 types with illegal zero length encoding (INTEGER,
18059 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
18061 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
18063 * Merge in my S/MIME library for OpenSSL. This provides a simple
18064 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
18065 functionality to handle multipart/signed properly) and a utility
18066 called 'smime' to call all this stuff. This is based on code I
18067 originally wrote for Celo who have kindly allowed it to be
18068 included in OpenSSL.
18072 * Add variants des_set_key_checked and des_set_key_unchecked of
18073 des_set_key (aka des_key_sched). Global variable des_check_key
18074 decides which of these is called by des_set_key; this way
18075 des_check_key behaves as it always did, but applications and
18076 the library itself, which was buggy for des_check_key == 1,
18077 have a cleaner way to pick the version they need.
18081 * New function PKCS12_newpass() which changes the password of a
18086 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
18087 dynamic mix. In both cases the ids can be used as an index into the
18088 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
18089 functions so they accept a list of the field values and the
18090 application doesn't need to directly manipulate the X509_TRUST
18095 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
18100 * Modify the way the V3 extension code looks up extensions. This now
18101 works in a similar way to the object code: we have some "standard"
18102 extensions in a static table which is searched with OBJ_bsearch()
18103 and the application can add dynamic ones if needed. The file
18104 crypto/x509v3/ext_dat.h now has the info: this file needs to be
18105 updated whenever a new extension is added to the core code and kept
18106 in ext_nid order. There is a simple program 'tabtest.c' which checks
18107 this. New extensions are not added too often so this file can readily
18108 be maintained manually.
18110 There are two big advantages in doing things this way. The extensions
18111 can be looked up immediately and no longer need to be "added" using
18112 X509V3_add_standard_extensions(): this function now does nothing.
18113 Side note: I get *lots* of email saying the extension code doesn't
18114 work because people forget to call this function.
18115 Also no dynamic allocation is done unless new extensions are added:
18116 so if we don't add custom extensions there is no need to call
18117 X509V3_EXT_cleanup().
18121 * Modify enc utility's salting as follows: make salting the default. Add a
18122 magic header, so unsalted files fail gracefully instead of just decrypting
18123 to garbage. This is because not salting is a big security hole, so people
18124 should be discouraged from doing it.
18128 * Fixes and enhancements to the 'x509' utility. It allowed a message
18129 digest to be passed on the command line but it only used this
18130 parameter when signing a certificate. Modified so all relevant
18131 operations are affected by the digest parameter including the
18132 -fingerprint and -x509toreq options. Also -x509toreq choked if a
18133 DSA key was used because it didn't fix the digest.
18137 * Initial certificate chain verify code. Currently tests the untrusted
18138 certificates for consistency with the verify purpose (which is set
18139 when the X509_STORE_CTX structure is set up) and checks the pathlength.
18141 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
18142 this is because it will reject chains with invalid extensions whereas
18143 every previous version of OpenSSL and SSLeay made no checks at all.
18145 Trust code: checks the root CA for the relevant trust settings. Trust
18146 settings have an initial value consistent with the verify purpose: e.g.
18147 if the verify purpose is for SSL client use it expects the CA to be
18148 trusted for SSL client use. However the default value can be changed to
18149 permit custom trust settings: one example of this would be to only trust
18150 certificates from a specific "secure" set of CAs.
18152 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
18153 which should be used for version portability: especially since the
18154 verify structure is likely to change more often now.
18156 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
18157 to set them. If not set then assume SSL clients will verify SSL servers
18160 Two new options to the verify program: -untrusted allows a set of
18161 untrusted certificates to be passed in and -purpose which sets the
18162 intended purpose of the certificate. If a purpose is set then the
18163 new chain verify code is used to check extension consistency.
18167 * Support for the authority information access extension.
18171 * Modify RSA and DSA PEM read routines to transparently handle
18172 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
18173 public keys in a format compatible with certificate
18174 SubjectPublicKeyInfo structures. Unfortunately there were already
18175 functions called *_PublicKey_* which used various odd formats so
18176 these are retained for compatibility: however the DSA variants were
18177 never in a public release so they have been deleted. Changed dsa/rsa
18178 utilities to handle the new format: note no releases ever handled public
18179 keys so we should be OK.
18181 The primary motivation for this change is to avoid the same fiasco
18182 that dogs private keys: there are several incompatible private key
18183 formats some of which are standard and some OpenSSL specific and
18184 require various evil hacks to allow partial transparent handling and
18185 even then it doesn't work with DER formats. Given the option anything
18186 other than PKCS#8 should be dumped: but the other formats have to
18187 stay in the name of compatibility.
18189 With public keys and the benefit of hindsight one standard format
18190 is used which works with EVP_PKEY, RSA or DSA structures: though
18191 it clearly returns an error if you try to read the wrong kind of key.
18193 Added a -pubkey option to the 'x509' utility to output the public key.
18194 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18195 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18196 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18197 that do the same as the `EVP_PKEY_assign_*()` except they up the
18198 reference count of the added key (they don't "swallow" the
18203 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18204 CRLs would fail if the file contained no certificates or no CRLs:
18205 added a new function to read in both types and return the number
18206 read: this means that if none are read it will be an error. The
18207 DER versions of the certificate and CRL reader would always fail
18208 because it isn't possible to mix certificates and CRLs in DER format
18209 without choking one or the other routine. Changed this to just read
18210 a certificate: this is the best we can do. Also modified the code
18211 in `apps/verify.c` to take notice of return codes: it was previously
18212 attempting to read in certificates from NULL pointers and ignoring
18213 any errors: this is one reason why the cert and CRL reader seemed
18214 to work. It doesn't check return codes from the default certificate
18215 routines: these may well fail if the certificates aren't installed.
18219 * Code to support otherName option in GeneralName.
18223 * First update to verify code. Change the verify utility
18224 so it warns if it is passed a self signed certificate:
18225 for consistency with the normal behaviour. X509_verify
18226 has been modified to it will now verify a self signed
18227 certificate if *exactly* the same certificate appears
18228 in the store: it was previously impossible to trust a
18229 single self signed certificate. This means that:
18230 openssl verify ss.pem
18231 now gives a warning about a self signed certificate but
18232 openssl verify -CAfile ss.pem ss.pem
18237 * For servers, store verify_result in SSL_SESSION data structure
18238 (and add it to external session representation).
18239 This is needed when client certificate verifications fails,
18240 but an application-provided verification callback (set by
18241 SSL_CTX_set_cert_verify_callback) allows accepting the session
18242 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18243 but returns 1): When the session is reused, we have to set
18244 ssl->verify_result to the appropriate error code to avoid
18247 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18249 * Fix a bug in the new PKCS#7 code: it didn't consider the
18250 case in PKCS7_dataInit() where the signed PKCS7 structure
18251 didn't contain any existing data because it was being created.
18253 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18255 * Add a salt to the key derivation routines in enc.c. This
18256 forms the first 8 bytes of the encrypted file. Also add a
18257 -S option to allow a salt to be input on the command line.
18261 * New function X509_cmp(). Oddly enough there wasn't a function
18262 to compare two certificates. We do this by working out the SHA1
18263 hash and comparing that. X509_cmp() will be needed by the trust
18268 * SSL_get1_session() is like SSL_get_session(), but increments
18269 the reference count in the SSL_SESSION returned.
18271 *Geoff Thorpe <geoff@eu.c2.net>*
18273 * Fix for 'req': it was adding a null to request attributes.
18274 Also change the X509_LOOKUP and X509_INFO code to handle
18275 certificate auxiliary information.
18279 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18284 * Add the possibility to add extra information to the memory leak
18285 detecting output, to form tracebacks, showing from where each
18286 allocation was originated: CRYPTO_push_info("constant string") adds
18287 the string plus current file name and line number to a per-thread
18288 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18289 is like calling CYRPTO_pop_info() until the stack is empty.
18290 Also updated memory leak detection code to be multi-thread-safe.
18294 * Add options -text and -noout to pkcs7 utility and delete the
18295 encryption options which never did anything. Update docs.
18299 * Add options to some of the utilities to allow the pass phrase
18300 to be included on either the command line (not recommended on
18301 OSes like Unix) or read from the environment. Update the
18302 manpages and fix a few bugs.
18306 * Add a few manpages for some of the openssl commands.
18310 * Fix the -revoke option in ca. It was freeing up memory twice,
18311 leaking and not finding already revoked certificates.
18315 * Extensive changes to support certificate auxiliary information.
18316 This involves the use of X509_CERT_AUX structure and X509_AUX
18317 functions. An X509_AUX function such as PEM_read_X509_AUX()
18318 can still read in a certificate file in the usual way but it
18319 will also read in any additional "auxiliary information". By
18320 doing things this way a fair degree of compatibility can be
18321 retained: existing certificates can have this information added
18322 using the new 'x509' options.
18324 Current auxiliary information includes an "alias" and some trust
18325 settings. The trust settings will ultimately be used in enhanced
18326 certificate chain verification routines: currently a certificate
18327 can only be trusted if it is self signed and then it is trusted
18332 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18333 The problem was that one of the replacement routines had not been working
18334 since SSLeay releases. For now the offending routine has been replaced
18335 with non-optimised assembler. Even so, this now gives around 95%
18336 performance improvement for 1024 bit RSA signs.
18340 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18341 handling. Most clients have the effective key size in bits equal to
18342 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18343 A few however don't do this and instead use the size of the decrypted key
18344 to determine the RC2 key length and the AlgorithmIdentifier to determine
18345 the effective key length. In this case the effective key length can still
18346 be 40 bits but the key length can be 168 bits for example. This is fixed
18347 by manually forcing an RC2 key into the EVP_PKEY structure because the
18348 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18349 the key length and effective key length are equal.
18353 * Add a bunch of functions that should simplify the creation of
18354 X509_NAME structures. Now you should be able to do:
18355 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18356 and have it automatically work out the correct field type and fill in
18357 the structures. The more adventurous can try:
18358 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18359 and it will (hopefully) work out the correct multibyte encoding.
18363 * Change the 'req' utility to use the new field handling and multibyte
18364 copy routines. Before the DN field creation was handled in an ad hoc
18365 way in req, ca, and x509 which was rather broken and didn't support
18366 BMPStrings or UTF8Strings. Since some software doesn't implement
18367 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18368 using the dirstring_type option. See the new comment in the default
18369 openssl.cnf for more info.
18373 * Make crypto/rand/md_rand.c more robust:
18374 - Assure unique random numbers after fork().
18375 - Make sure that concurrent threads access the global counter and
18376 md serializably so that we never lose entropy in them
18377 or use exactly the same state in multiple threads.
18378 Access to the large state is not always serializable because
18379 the additional locking could be a performance killer, and
18380 md should be large enough anyway.
18384 * New file `apps/app_rand.c` with commonly needed functionality
18385 for handling the random seed file.
18387 Use the random seed file in some applications that previously did not:
18389 dsaparam -genkey (which also ignored its '-rand' option),
18392 x509 (when signing).
18393 Except on systems with /dev/urandom, it is crucial to have a random
18394 seed file at least for key creation, DSA signing, and for DH exchanges;
18395 for RSA signatures we could do without one.
18397 gendh and gendsa (unlike genrsa) used to read only the first byte
18398 of each file listed in the '-rand' option. The function as previously
18399 found in genrsa is now in app_rand.c and is used by all programs
18400 that support '-rand'.
18404 * In RAND_write_file, use mode 0600 for creating files;
18405 don't just chmod when it may be too late.
18409 * Report an error from X509_STORE_load_locations
18410 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18414 * New function ASN1_mbstring_copy() this copies a string in either
18415 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18416 into an ASN1_STRING type. A mask of permissible types is passed
18417 and it chooses the "minimal" type to use or an error if not type
18422 * Add function equivalents to the various macros in asn1.h. The old
18423 macros are retained with an `M_` prefix. Code inside the library can
18424 use the `M_` macros. External code (including the openssl utility)
18425 should *NOT* in order to be "shared library friendly".
18429 * Add various functions that can check a certificate's extensions
18430 to see if it usable for various purposes such as SSL client,
18431 server or S/MIME and CAs of these types. This is currently
18432 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18433 verification. Also added a -purpose flag to x509 utility to
18434 print out all the purposes.
18438 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18443 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18444 for, obtain and decode and extension and obtain its critical flag.
18445 This allows all the necessary extension code to be handled in a
18446 single function call.
18450 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18451 platforms. See crypto/rc4/rc4_enc.c for further details.
18455 * New -noout option to asn1parse. This causes no output to be produced
18456 its main use is when combined with -strparse and -out to extract data
18457 from a file (which may not be in ASN.1 format).
18461 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18462 when producing the local key id.
18464 *Richard Levitte <levitte@stacken.kth.se>*
18466 * New option -dhparam in s_server. This allows a DH parameter file to be
18467 stated explicitly. If it is not stated then it tries the first server
18468 certificate file. The previous behaviour hard coded the filename
18473 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18474 a public key to be input or output. For example:
18475 openssl rsa -in key.pem -pubout -out pubkey.pem
18476 Also added necessary DSA public key functions to handle this.
18480 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18481 in the message. This was handled by allowing
18482 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18484 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18486 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18487 to the end of the strings whereas this didn't. This would cause problems
18488 if strings read with d2i_ASN1_bytes() were later modified.
18490 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18492 * Fix for base64 decode bug. When a base64 bio reads only one line of
18493 data and it contains EOF it will end up returning an error. This is
18494 caused by input 46 bytes long. The cause is due to the way base64
18495 BIOs find the start of base64 encoded data. They do this by trying a
18496 trial decode on each line until they find one that works. When they
18497 do a flag is set and it starts again knowing it can pass all the
18498 data directly through the decoder. Unfortunately it doesn't reset
18499 the context it uses. This means that if EOF is reached an attempt
18500 is made to pass two EOFs through the context and this causes the
18501 resulting error. This can also cause other problems as well. As is
18502 usual with these problems it takes *ages* to find and the fix is
18503 trivial: move one line.
18505 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18507 * Ugly workaround to get s_client and s_server working under Windows. The
18508 old code wouldn't work because it needed to select() on sockets and the
18509 tty (for keypresses and to see if data could be written). Win32 only
18510 supports select() on sockets so we select() with a 1s timeout on the
18511 sockets and then see if any characters are waiting to be read, if none
18512 are present then we retry, we also assume we can always write data to
18513 the tty. This isn't nice because the code then blocks until we've
18514 received a complete line of data and it is effectively polling the
18515 keyboard at 1s intervals: however it's quite a bit better than not
18516 working at all :-) A dedicated Windows application might handle this
18517 with an event loop for example.
18521 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18522 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18523 will be called when RSA_sign() and RSA_verify() are used. This is useful
18524 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18525 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18526 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18527 This necessitated the support of an extra signature type NID_md5_sha1
18528 for SSL signatures and modifications to the SSL library to use it instead
18529 of calling RSA_public_decrypt() and RSA_private_encrypt().
18533 * Add new -verify -CAfile and -CApath options to the crl program, these
18534 will lookup a CRL issuers certificate and verify the signature in a
18535 similar way to the verify program. Tidy up the crl program so it
18536 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18537 less strict. It will now permit CRL extensions even if it is not
18538 a V2 CRL: this will allow it to tolerate some broken CRLs.
18542 * Initialize all non-automatic variables each time one of the openssl
18543 sub-programs is started (this is necessary as they may be started
18544 multiple times from the "OpenSSL>" prompt).
18546 *Lennart Bang, Bodo Moeller*
18548 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18549 removing all other RSA functionality (this is what NO_RSA does). This
18550 is so (for example) those in the US can disable those operations covered
18551 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18556 * Non-copying interface to BIO pairs.
18557 (still largely untested)
18561 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18562 ASCII string. This was handled independently in various places before.
18566 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18567 UTF8 strings a character at a time.
18571 * Use client_version from client hello to select the protocol
18572 (s23_srvr.c) and for RSA client key exchange verification
18573 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18577 * Add various utility functions to handle SPKACs, these were previously
18578 handled by poking round in the structure internals. Added new function
18579 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18580 print, verify and generate SPKACs. Based on an original idea from
18581 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18585 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18589 * Allow the config file extension section to be overwritten on the
18590 command line. Based on an original idea from Massimiliano Pala
18591 <madwolf@comune.modena.it>. The new option is called -extensions
18592 and can be applied to ca, req and x509. Also -reqexts to override
18593 the request extensions in req and -crlexts to override the crl extensions
18598 * Add new feature to the SPKAC handling in ca. Now you can include
18599 the same field multiple times by preceding it by "XXXX." for example:
18602 this is the same syntax as used in the req config file.
18606 * Allow certificate extensions to be added to certificate requests. These
18607 are specified in a 'req_extensions' option of the req section of the
18608 config file. They can be printed out with the -text option to req but
18609 are otherwise ignored at present.
18613 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18614 data read consists of only the final block it would not decrypted because
18615 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18616 A misplaced 'break' also meant the decrypted final block might not be
18617 copied until the next read.
18621 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18622 a few extra parameters to the DH structure: these will be useful if
18623 for example we want the value of 'q' or implement X9.42 DH.
18627 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18628 provides hooks that allow the default DSA functions or functions on a
18629 "per key" basis to be replaced. This allows hardware acceleration and
18630 hardware key storage to be handled without major modification to the
18631 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18632 associated functions.
18636 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18637 as "read only": it can't be written to and the buffer it points to will
18638 not be freed. Reading from a read only BIO is much more efficient than
18639 a normal memory BIO. This was added because there are several times when
18640 an area of memory needs to be read from a BIO. The previous method was
18641 to create a memory BIO and write the data to it, this results in two
18642 copies of the data and an O(n^2) reading algorithm. There is a new
18643 function BIO_new_mem_buf() which creates a read only memory BIO from
18644 an area of memory. Also modified the PKCS#7 routines to use read only
18649 * Bugfix: ssl23_get_client_hello did not work properly when called in
18650 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18651 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18652 but a retry condition occurred while trying to read the rest.
18656 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18657 NID_pkcs7_encrypted by default: this was wrong since this should almost
18658 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18659 the encrypted data type: this is a more sensible place to put it and it
18660 allows the PKCS#12 code to be tidied up that duplicated this
18665 * Changed obj_dat.pl script so it takes its input and output files on
18666 the command line. This should avoid shell escape redirection problems
18671 * Initial support for certificate extension requests, these are included
18672 in things like Xenroll certificate requests. Included functions to allow
18673 extensions to be obtained and added.
18677 * -crlf option to s_client and s_server for sending newlines as
18678 CRLF (as required by many protocols).
18682 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18684 * Install libRSAglue.a when OpenSSL is built with RSAref.
18686 *Ralf S. Engelschall*
18688 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18690 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18692 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18697 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18698 DH parameters/keys (q is lost during that conversion, but the resulting
18699 DH parameters contain its length).
18701 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18702 much faster than DH_generate_parameters (which creates parameters
18703 where `p = 2*q + 1`), and also the smaller q makes DH computations
18704 much more efficient (160-bit exponentiation instead of 1024-bit
18705 exponentiation); so this provides a convenient way to support DHE
18706 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18707 utter importance to use
18708 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18710 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18711 when such DH parameters are used, because otherwise small subgroup
18712 attacks may become possible!
18716 * Avoid memory leak in i2d_DHparams.
18720 * Allow the -k option to be used more than once in the enc program:
18721 this allows the same encrypted message to be read by multiple recipients.
18725 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18726 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18727 it will always use the numerical form of the OID, even if it has a short
18732 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18733 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18734 otherwise bn_mod_exp was called. In the case of hardware keys for example
18735 no private key components need be present and it might store extra data
18736 in the RSA structure, which cannot be accessed from bn_mod_exp.
18737 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18738 private key operations.
18742 * Added support for SPARC Linux.
18746 * pem_password_cb function type incompatibly changed from
18747 typedef int pem_password_cb(char *buf, int size, int rwflag);
18749 ....(char *buf, int size, int rwflag, void *userdata);
18750 so that applications can pass data to their callbacks:
18751 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18752 additional void * argument, which is just handed through whenever
18753 the password callback is called.
18755 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18757 New function SSL_CTX_set_default_passwd_cb_userdata.
18759 Compatibility note: As many C implementations push function arguments
18760 onto the stack in reverse order, the new library version is likely to
18761 interoperate with programs that have been compiled with the old
18762 pem_password_cb definition (PEM_whatever takes some data that
18763 happens to be on the stack as its last argument, and the callback
18764 just ignores this garbage); but there is no guarantee whatsoever that
18767 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18768 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18769 problems not only on Windows, but also on some Unix platforms.
18770 To avoid problematic command lines, these definitions are now in an
18771 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18772 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18776 * MIPS III/IV assembler module is reimplemented.
18780 * More DES library cleanups: remove references to srand/rand and
18781 delete an unused file.
18785 * Add support for the free Netwide assembler (NASM) under Win32,
18786 since not many people have MASM (ml) and it can be hard to obtain.
18787 This is currently experimental but it seems to work OK and pass all
18788 the tests. Check out INSTALL.W32 for info.
18792 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18793 without temporary keys kept an extra copy of the server key,
18794 and connections with temporary keys did not free everything in case
18799 * New function RSA_check_key and new openssl rsa option -check
18800 for verifying the consistency of RSA keys.
18802 *Ulf Moeller, Bodo Moeller*
18804 * Various changes to make Win32 compile work:
18805 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18806 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18807 comparison" warnings.
18808 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18812 * Add a debugging option to PKCS#5 v2 key generation function: when
18813 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18814 derived keys are printed to stderr.
18818 * Copy the flags in ASN1_STRING_dup().
18820 *Roman E. Pavlov <pre@mo.msk.ru>*
18822 * The x509 application mishandled signing requests containing DSA
18823 keys when the signing key was also DSA and the parameters didn't match.
18825 It was supposed to omit the parameters when they matched the signing key:
18826 the verifying software was then supposed to automatically use the CA's
18827 parameters if they were absent from the end user certificate.
18829 Omitting parameters is no longer recommended. The test was also
18830 the wrong way round! This was probably due to unusual behaviour in
18831 EVP_cmp_parameters() which returns 1 if the parameters match.
18832 This meant that parameters were omitted when they *didn't* match and
18833 the certificate was useless. Certificates signed with 'ca' didn't have
18836 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18838 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18839 The interface is as follows:
18840 Applications can use
18841 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18842 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18843 "off" is now the default.
18844 The library internally uses
18845 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18846 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18847 to disable memory-checking temporarily.
18849 Some inconsistent states that previously were possible (and were
18850 even the default) are now avoided.
18852 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18853 with each memory chunk allocated; this is occasionally more helpful
18854 than just having a counter.
18856 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18858 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18863 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18864 which largely parallels "options", but is for changing API behaviour,
18865 whereas "options" are about protocol behaviour.
18866 Initial "mode" flags are:
18868 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18869 a single record has been written.
18870 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18871 retries use the same buffer location.
18872 (But all of the contents must be
18877 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18880 * Fix problems with no-hmac etc.
18882 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18884 * New functions RSA_get_default_method(), RSA_set_method() and
18885 RSA_get_method(). These allows replacement of RSA_METHODs without having
18886 to mess around with the internals of an RSA structure.
18890 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18891 Also really enable memory leak checks in openssl.c and in some
18894 *Chad C. Mulligan, Bodo Moeller*
18896 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18897 up the length of negative integers. This has now been simplified to just
18898 store the length when it is first determined and use it later, rather
18899 than trying to keep track of where data is copied and updating it to
18901 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18903 * Add a new function PKCS7_signatureVerify. This allows the verification
18904 of a PKCS#7 signature but with the signing certificate passed to the
18905 function itself. This contrasts with PKCS7_dataVerify which assumes the
18906 certificate is present in the PKCS#7 structure. This isn't always the
18907 case: certificates can be omitted from a PKCS#7 structure and be
18908 distributed by "out of band" means (such as a certificate database).
18912 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18913 function prototypes in pem.h, also change util/mkdef.pl to add the
18914 necessary function names.
18918 * mk1mf.pl (used by Windows builds) did not properly read the
18919 options set by Configure in the top level Makefile, and Configure
18920 was not even able to write more than one option correctly.
18921 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18925 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18926 file to be loaded from a BIO or FILE pointer. The BIO version will
18927 for example allow memory BIOs to contain config info.
18931 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18932 Whoever hopes to achieve shared-library compatibility across versions
18933 must use this, not the compile-time macro.
18934 (Exercise 0.9.4: Which is the minimum library version required by
18936 Note: All this applies only to multi-threaded programs, others don't
18941 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18942 through a BIO pair triggered the default case, i.e.
18943 SSLerr(...,SSL_R_UNKNOWN_STATE).
18947 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18948 can use the SSL library even if none of the specific BIOs is
18953 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18954 for the encoded length.
18956 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18958 * Add initial documentation of the X509V3 functions.
18962 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18963 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18964 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18965 secure PKCS#8 private key format with a high iteration count.
18969 * Fix determination of Perl interpreter: A perl or perl5
18970 *directory* in $PATH was also accepted as the interpreter.
18972 *Ralf S. Engelschall*
18974 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18975 wrong with it but it was very old and did things like calling
18976 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18977 unusual formatting.
18981 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18982 to use the new extension code.
18986 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18987 with macros. This should make it easier to change their form, add extra
18988 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18993 * Add to configuration table a new entry that can specify an alternative
18994 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18995 according to Mark Crispin <MRC@Panda.COM>.
18999 * DES CBC did not update the IV. Weird.
19003 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
19004 Changing the behaviour of the former might break existing programs --
19005 where IV updating is needed, des_ncbc_encrypt can be used.
19008 * When bntest is run from "make test" it drives bc to check its
19009 calculations, as well as internally checking them. If an internal check
19010 fails, it needs to cause bc to give a non-zero result or make test carries
19011 on without noticing the failure. Fixed.
19015 * DES library cleanups.
19019 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
19020 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
19021 ciphers. NOTE: although the key derivation function has been verified
19022 against some published test vectors it has not been extensively tested
19023 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
19028 * Instead of "mkdir -p", which is not fully portable, use new
19029 Perl script "util/mkdir-p.pl".
19033 * Rewrite the way password based encryption (PBE) is handled. It used to
19034 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
19035 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
19036 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
19037 the 'parameter' field of the AlgorithmIdentifier is passed to the
19038 underlying key generation function so it must do its own ASN1 parsing.
19039 This has also changed the EVP_PBE_CipherInit() function which now has a
19040 'parameter' argument instead of literal salt and iteration count values
19041 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
19045 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
19046 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
19047 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
19048 KEY" because this clashed with PKCS#8 unencrypted string. Since this
19049 value was just used as a "magic string" and not used directly its
19050 value doesn't matter.
19054 * Introduce some semblance of const correctness to BN. Shame C doesn't
19059 * "linux-sparc64" configuration (ultrapenguin).
19061 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
19062 "linux-sparc" configuration.
19064 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
19066 * config now generates no-xxx options for missing ciphers.
19070 * Support the EBCDIC character set (work in progress).
19071 File ebcdic.c not yet included because it has a different license.
19073 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19075 * Support BS2000/OSD-POSIX.
19077 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19079 * Make callbacks for key generation use `void *` instead of `char *`.
19083 * Make S/MIME samples compile (not yet tested).
19087 * Additional typesafe stacks.
19091 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
19095 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
19097 * New configuration variant "sco5-gcc".
19099 * Updated some demos.
19101 *Sean O Riordain, Wade Scholine*
19103 * Add missing BIO_free at exit of pkcs12 application.
19107 * Fix memory leak in conf.c.
19111 * Updates for Win32 to assembler version of MD5.
19115 * Set #! path to perl in `apps/der_chop` to where we found it
19116 instead of using a fixed path.
19120 * SHA library changes for irix64-mips4-cc.
19124 * Improvements for VMS support.
19128 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
19130 * Bignum library bug fix. IRIX 6 passes "make test" now!
19131 This also avoids the problems with SC4.2 and unpatched SC5.
19133 *Andy Polyakov <appro@fy.chalmers.se>*
19135 * New functions sk_num, sk_value and sk_set to replace the previous macros.
19136 These are required because of the typesafe stack would otherwise break
19137 existing code. If old code used a structure member which used to be STACK
19138 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
19139 sk_num or sk_value it would produce an error because the num, data members
19140 are not present in STACK_OF. Now it just produces a warning. sk_set
19141 replaces the old method of assigning a value to sk_value
19142 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
19143 that does this will no longer work (and should use sk_set instead) but
19144 this could be regarded as a "questionable" behaviour anyway.
19148 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
19149 correctly handle encrypted S/MIME data.
19153 * Change type of various DES function arguments from des_cblock
19154 (which means, in function argument declarations, pointer to char)
19155 to des_cblock * (meaning pointer to array with 8 char elements),
19156 which allows the compiler to do more typechecking; it was like
19157 that back in SSLeay, but with lots of ugly casts.
19159 Introduce new type const_des_cblock.
19163 * Reorganise the PKCS#7 library and get rid of some of the more obvious
19164 problems: find RecipientInfo structure that matches recipient certificate
19165 and initialise the ASN1 structures properly based on passed cipher.
19169 * Belatedly make the BN tests actually check the results.
19173 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
19174 to and from BNs: it was completely broken. New compilation option
19175 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
19176 key elements as negative integers.
19180 * Reorganize and speed up MD5.
19182 *Andy Polyakov <appro@fy.chalmers.se>*
19186 *Richard Levitte <richard@levitte.org>*
19188 * New option -out to asn1parse to allow the parsed structure to be
19189 output to a file. This is most useful when combined with the -strparse
19190 option to examine the output of things like OCTET STRINGS.
19194 * Make SSL library a little more fool-proof by not requiring any longer
19195 that `SSL_set_{accept,connect}_state` be called before
19196 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19197 in many applications because usually everything *appeared* to work as
19198 intended anyway -- now it really works as intended).
19202 * Move openssl.cnf out of lib/.
19206 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19207 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19208 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19210 *Ralf S. Engelschall*
19212 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19213 handle PKCS#7 enveloped data properly.
19215 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19217 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19218 copying pointers. The cert_st handling is changed by this in
19219 various ways (and thus what used to be known as ctx->default_cert
19220 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19221 any longer when s->cert does not give us what we need).
19222 ssl_cert_instantiate becomes obsolete by this change.
19223 As soon as we've got the new code right (possibly it already is?),
19224 we have solved a couple of bugs of the earlier code where s->cert
19225 was used as if it could not have been shared with other SSL structures.
19227 Note that using the SSL API in certain dirty ways now will result
19228 in different behaviour than observed with earlier library versions:
19229 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19230 does not influence s as it used to.
19232 In order to clean up things more thoroughly, inside SSL_SESSION
19233 we don't use CERT any longer, but a new structure SESS_CERT
19234 that holds per-session data (if available); currently, this is
19235 the peer's certificate chain and, for clients, the server's certificate
19236 and temporary key. CERT holds only those values that can have
19237 meaningful defaults in an SSL_CTX.
19241 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19242 from the internal representation. Various PKCS#7 fixes: remove some
19243 evil casts and set the enc_dig_alg field properly based on the signing
19248 * Allow PKCS#12 password to be set from the command line or the
19249 environment. Let 'ca' get its config file name from the environment
19250 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19255 * Allow certificate policies extension to use an IA5STRING for the
19256 organization field. This is contrary to the PKIX definition but
19257 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19262 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19263 without disallowing inline assembler and the like for non-pedantic builds.
19267 * Support Borland C++ builder.
19269 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19275 * SHA-1 cleanups and performance enhancements.
19277 *Andy Polyakov <appro@fy.chalmers.se>*
19279 * Sparc v8plus assembler for the bignum library.
19281 *Andy Polyakov <appro@fy.chalmers.se>*
19283 * Accept any -xxx and +xxx compiler options in Configure.
19287 * Update HPUX configuration.
19291 * Add missing `sk_<type>_unshift()` function to safestack.h
19293 *Ralf S. Engelschall*
19295 * New function SSL_CTX_use_certificate_chain_file that sets the
19296 "extra_cert"s in addition to the certificate. (This makes sense
19297 only for "PEM" format files, as chains as a whole are not
19302 * Support verify_depth from the SSL API.
19303 x509_vfy.c had what can be considered an off-by-one-error:
19304 Its depth (which was not part of the external interface)
19305 was actually counting the number of certificates in a chain;
19306 now it really counts the depth.
19310 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19311 instead of X509err, which often resulted in confusing error
19312 messages since the error codes are not globally unique
19313 (e.g. an alleged error in ssl3_accept when a certificate
19314 didn't match the private key).
19316 * New function SSL_CTX_set_session_id_context that allows to set a default
19317 value (so that you don't need SSL_set_session_id_context for each
19318 connection using the SSL_CTX).
19322 * OAEP decoding bug fix.
19326 * Support INSTALL_PREFIX for package builders, as proposed by
19331 * New Configure options "threads" and "no-threads". For systems
19332 where the proper compiler options are known (currently Solaris
19333 and Linux), "threads" is the default.
19337 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19341 * Install various scripts to $(OPENSSLDIR)/misc, not to
19342 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19343 such as /usr/local/bin.
19347 * "make linux-shared" to build shared libraries.
19349 *Niels Poppe <niels@netbox.org>*
19351 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19355 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19356 extension adding in x509 utility.
19360 * Remove NOPROTO sections and error code comments.
19364 * Partial rewrite of the DEF file generator to now parse the ANSI
19369 * New Configure options --prefix=DIR and --openssldir=DIR.
19373 * Complete rewrite of the error code script(s). It is all now handled
19374 by one script at the top level which handles error code gathering,
19375 header rewriting and C source file generation. It should be much better
19376 than the old method: it now uses a modified version of Ulf's parser to
19377 read the ANSI prototypes in all header files (thus the old K&R definitions
19378 aren't needed for error creation any more) and do a better job of
19379 translating function codes into names. The old 'ASN1 error code embedded
19380 in a comment' is no longer necessary and it doesn't use .err files which
19381 have now been deleted. Also the error code call doesn't have to appear all
19382 on one line (which resulted in some large lines...).
19386 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19390 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19391 0 (which usually indicates a closed connection), but continue reading.
19395 * Fix some race conditions.
19399 * Add support for CRL distribution points extension. Add Certificate
19400 Policies and CRL distribution points documentation.
19404 * Move the autogenerated header file parts to crypto/opensslconf.h.
19408 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19409 8 of keying material. Merlin has also confirmed interop with this fix
19410 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19412 *Merlin Hughes <merlin@baltimore.ie>*
19414 * Fix lots of warnings.
19416 *Richard Levitte <levitte@stacken.kth.se>*
19418 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19419 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19421 *Richard Levitte <levitte@stacken.kth.se>*
19423 * Fix problems with sizeof(long) == 8.
19425 *Andy Polyakov <appro@fy.chalmers.se>*
19427 * Change functions to ANSI C.
19431 * Fix typos in error codes.
19433 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19435 * Remove defunct assembler files from Configure.
19439 * SPARC v8 assembler BIGNUM implementation.
19441 *Andy Polyakov <appro@fy.chalmers.se>*
19443 * Support for Certificate Policies extension: both print and set.
19444 Various additions to support the r2i method this uses.
19448 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19449 return a const string when you are expecting an allocated buffer.
19453 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19454 types DirectoryString and DisplayText.
19458 * Add code to allow r2i extensions to access the configuration database,
19459 add an LHASH database driver and add several ctx helper functions.
19463 * Fix an evil bug in bn_expand2() which caused various BN functions to
19464 fail when they extended the size of a BIGNUM.
19468 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19469 support typesafe stack.
19473 * Fix typo in SSL_[gs]et_options().
19475 *Nils Frostberg <nils@medcom.se>*
19477 * Delete various functions and files that belonged to the (now obsolete)
19478 old X509V3 handling code.
19482 * New Configure option "rsaref".
19486 * Don't auto-generate pem.h.
19490 * Introduce type-safe ASN.1 SETs.
19494 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19496 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19498 * Introduce type-safe STACKs. This will almost certainly break lots of code
19499 that links with OpenSSL (well at least cause lots of warnings), but fear
19500 not: the conversion is trivial, and it eliminates loads of evil casts. A
19501 few STACKed things have been converted already. Feel free to convert more.
19502 In the fullness of time, I'll do away with the STACK type altogether.
19506 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19507 specified in `<certfile>` by updating the entry in the index.txt file.
19508 This way one no longer has to edit the index.txt file manually for
19509 revoking a certificate. The -revoke option does the gory details now.
19511 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19513 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19514 `-text` option at all and this way the `-noout -text` combination was
19515 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19517 *Ralf S. Engelschall*
19519 * Make sure a corresponding plain text error message exists for the
19520 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19521 verify callback function determined that a certificate was revoked.
19523 *Ralf S. Engelschall*
19525 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19526 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19527 all available ciphers including rc5, which was forgotten until now.
19528 In order to let the testing shell script know which algorithms
19529 are available, a new (up to now undocumented) command
19530 `openssl list-cipher-commands` is used.
19534 * Bugfix: s_client occasionally would sleep in select() when
19535 it should have checked SSL_pending() first.
19539 * New functions DSA_do_sign and DSA_do_verify to provide access to
19540 the raw DSA values prior to ASN.1 encoding.
19544 * Tweaks to Configure
19546 *Niels Poppe <niels@netbox.org>*
19548 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19553 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19557 * New config option to avoid instructions that are illegal on the 80386.
19558 The default code is faster, but requires at least a 486.
19562 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19563 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19564 same as SSL2_VERSION anyway.
19568 * New "-showcerts" option for s_client.
19572 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19573 application. Various cleanups and fixes.
19577 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19578 modify error routines to work internally. Add error codes and PBE init
19579 to library startup routines.
19583 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19584 packing functions to asn1 and evp. Changed function names and error
19585 codes along the way.
19589 * PKCS12 integration: and so it begins... First of several patches to
19590 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19591 objects to objects.h
19595 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19596 and display support for Thawte strong extranet extension.
19600 * Add LinuxPPC support.
19602 *Jeff Dubrule <igor@pobox.org>*
19604 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19605 bn_div_words in alpha.s.
19607 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19609 * Make sure the RSA OAEP test is skipped under -DRSAref because
19610 OAEP isn't supported when OpenSSL is built with RSAref.
19612 *Ulf Moeller <ulf@fitug.de>*
19614 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19615 so they no longer are missing under -DNOPROTO.
19617 *Soren S. Jorvang <soren@t.dk>*
19619 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19621 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19622 doesn't work when the session is reused. Coming soon!
19626 * Fix a security hole, that allows sessions to be reused in the wrong
19627 context thus bypassing client cert protection! All software that uses
19628 client certs and session caches in multiple contexts NEEDS PATCHING to
19629 allow session reuse! A fuller solution is in the works.
19631 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19633 * Some more source tree cleanups (removed obsolete files
19634 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19635 permission on "config" script to be executable) and a fix for the INSTALL
19638 *Ulf Moeller <ulf@fitug.de>*
19640 * Remove some legacy and erroneous uses of malloc, free instead of
19643 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19645 * Make rsa_oaep_test return non-zero on error.
19647 *Ulf Moeller <ulf@fitug.de>*
19649 * Add support for native Solaris shared libraries. Configure
19650 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19651 if someone would make that last step automatic.
19653 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19655 * ctx_size was not built with the right compiler during "make links". Fixed.
19659 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19660 except NULL ciphers". This means the default cipher list will no longer
19661 enable NULL ciphers. They need to be specifically enabled e.g. with
19662 the string "DEFAULT:eNULL".
19666 * Fix to RSA private encryption routines: if p < q then it would
19667 occasionally produce an invalid result. This will only happen with
19668 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19672 * Be less restrictive and allow also `perl util/perlpath.pl
19673 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19674 because this way one can also use an interpreter named `perl5` (which is
19675 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19676 installed as `perl`).
19678 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19680 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19682 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19684 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19685 advapi32.lib to Win32 build and change the pem test comparison
19686 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19687 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19688 and crypto/des/ede_cbcm_enc.c.
19692 * DES quad checksum was broken on big-endian architectures. Fixed.
19696 * Comment out two functions in bio.h that aren't implemented. Fix up the
19697 Win32 test batch file so it (might) work again. The Win32 test batch file
19698 is horrible: I feel ill....
19702 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19703 in e_os.h. Audit of header files to check ANSI and non ANSI
19704 sections: 10 functions were absent from non ANSI section and not exported
19705 from Windows DLLs. Fixed up libeay.num for new functions.
19709 * Make `openssl version` output lines consistent.
19711 *Ralf S. Engelschall*
19713 * Fix Win32 symbol export lists for BIO functions: Added
19714 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19715 to ms/libeay{16,32}.def.
19717 *Ralf S. Engelschall*
19719 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19720 fine under Unix and passes some trivial tests I've now added. But the
19721 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19722 added to make sure no one expects that this stuff really works in the
19723 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19724 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19727 *Ralf S. Engelschall*
19729 * Fix the generation of two part addresses in perl.
19731 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19733 * Add config entry for Linux on MIPS.
19735 *John Tobey <jtobey@channel1.com>*
19737 * Make links whenever Configure is run, unless we are on Windoze.
19741 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19742 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19747 * Add a useful kludge to allow package maintainers to specify compiler and
19748 other platforms details on the command line without having to patch the
19749 Configure script every time: One now can use
19750 `perl Configure <id>:<details>`,
19751 i.e. platform ids are allowed to have details appended
19752 to them (separated by colons). This is treated as there would be a static
19753 pre-configured entry in Configure's %table under key `<id>` with value
19754 `<details>` and `perl Configure <id>` is called. So, when you want to
19755 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19756 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19757 now, which overrides the FreeBSD-elf entry on-the-fly.
19759 *Ralf S. Engelschall*
19761 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19765 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19766 on the `perl Configure ...` command line. This way one can compile
19767 OpenSSL libraries with Position Independent Code (PIC) which is needed
19768 for linking it into DSOs.
19770 *Ralf S. Engelschall*
19772 * Remarkably, export ciphers were totally broken and no-one had noticed!
19777 * Cleaned up the LICENSE document: The official contact for any license
19778 questions now is the OpenSSL core team under openssl-core@openssl.org.
19779 And add a paragraph about the dual-license situation to make sure people
19780 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19781 to the OpenSSL toolkit.
19783 *Ralf S. Engelschall*
19785 * General source tree makefile cleanups: Made `making xxx in yyy...`
19786 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19787 Additionally cleaned up the `make links` target: Remove unnecessary
19788 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19789 to speed processing and no longer clutter the display with confusing
19790 stuff. Instead only the actually done links are displayed.
19792 *Ralf S. Engelschall*
19794 * Permit null encryption ciphersuites, used for authentication only. It used
19795 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19796 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19801 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19802 signed attributes when verifying signatures (this would break them),
19803 the detached data encoding was wrong and public keys obtained using
19804 X509_get_pubkey() weren't freed.
19808 * Add text documentation for the BUFFER functions. Also added a work around
19809 to a Win95 console bug. This was triggered by the password read stuff: the
19810 last character typed gets carried over to the next fread(). If you were
19811 generating a new cert request using 'req' for example then the last
19812 character of the passphrase would be CR which would then enter the first
19817 * Added the new 'Includes OpenSSL Cryptography Software' button as
19818 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19819 button and can be used by applications based on OpenSSL to show the
19820 relationship to the OpenSSL project.
19822 *Ralf S. Engelschall*
19824 * Remove confusing variables in function signatures in files
19825 ssl/ssl_lib.c and ssl/ssl.h.
19827 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19829 * Don't install bss_file.c under PREFIX/include/
19831 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19833 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19834 functions that return function pointers and has support for NT specific
19835 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19836 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19837 unsigned to signed types: this was killing the Win32 compile.
19841 * Add new certificate file to stack functions,
19842 SSL_add_dir_cert_subjects_to_stack() and
19843 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19844 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19845 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19846 This means that Apache-SSL and similar packages don't have to mess around
19847 to add as many CAs as they want to the preferred list.
19851 * Experiment with doxygen documentation. Currently only partially applied to
19853 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19854 openssl.doxy as the configuration file.
19858 * Get rid of remaining C++-style comments which strict C compilers hate.
19860 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19862 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19863 compiled in by default: it has problems with large keys.
19867 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19868 DH private keys and/or callback functions which directly correspond to
19869 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19870 is needed for applications which have to configure certificates on a
19871 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19873 For the RSA certificate situation is makes no difference, but
19874 for the DSA certificate situation this fixes the "no shared cipher"
19875 problem where the OpenSSL cipher selection procedure failed because the
19876 temporary keys were not overtaken from the context and the API provided
19877 no way to reconfigure them.
19878 The new functions now let applications reconfigure the stuff and they
19879 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19880 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19881 non-public-API function ssl_cert_instantiate() is used as a helper
19882 function and also to reduce code redundancy inside ssl_rsa.c.
19884 *Ralf S. Engelschall*
19886 * Move s_server -dcert and -dkey options out of the undocumented feature
19887 area because they are useful for the DSA situation and should be
19888 recognized by the users.
19890 *Ralf S. Engelschall*
19892 * Fix the cipher decision scheme for export ciphers: the export bits are
19893 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19894 SSL_EXP_MASK. So, the original variable has to be used instead of the
19895 already masked variable.
19897 *Richard Levitte <levitte@stacken.kth.se>*
19899 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19901 *Richard Levitte <levitte@stacken.kth.se>*
19903 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19904 from `int` to `unsigned int` because it is a length and initialized by
19905 EVP_DigestFinal() which expects an `unsigned int *`.
19907 *Richard Levitte <levitte@stacken.kth.se>*
19909 * Don't hard-code path to Perl interpreter on shebang line of Configure
19910 script. Instead use the usual Shell->Perl transition trick.
19912 *Ralf S. Engelschall*
19914 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19915 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19916 -noout -modulus` as it's already the case for `openssl rsa -noout
19917 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19918 currently the public key is printed (a decision which was already done by
19919 `openssl dsa -modulus` in the past) which serves a similar purpose.
19920 Additionally the NO_RSA no longer completely removes the whole -modulus
19921 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19924 *Ralf S. Engelschall*
19926 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19927 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19929 *Arne Ansper <arne@ats.cyber.ee>*
19931 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19932 to be added. Now both 'req' and 'ca' can use new objects defined in the
19937 * Add cool BIO that does syslog (or event log on NT).
19939 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19941 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19942 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19943 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19944 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19948 * Add preliminary config info for new extension code.
19952 * Make RSA_NO_PADDING really use no padding.
19954 *Ulf Moeller <ulf@fitug.de>*
19956 * Generate errors when private/public key check is done.
19960 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19961 for some CRL extensions and new objects added.
19965 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19966 key usage extension and fuller support for authority key id.
19970 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19971 padding method for RSA, which is recommended for new applications in PKCS
19972 #1 v2.0 (RFC 2437, October 1998).
19973 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19974 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19975 against Bleichbacher's attack on RSA.
19976 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19979 * Updates to the new SSL compression code
19981 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19983 * Fix so that the version number in the master secret, when passed
19984 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19985 (because the server will not accept higher), that the version number
19986 is 0x03,0x01, not 0x03,0x00
19988 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19990 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19991 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19992 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19996 * Support for RAW extensions where an arbitrary extension can be
19997 created by including its DER encoding. See `apps/openssl.cnf` for
20002 * Make sure latest Perl versions don't interpret some generated C array
20003 code as Perl array code in the crypto/err/err_genc.pl script.
20005 *Lars Weber <3weber@informatik.uni-hamburg.de>*
20007 * Modify ms/do_ms.bat to not generate assembly language makefiles since
20008 not many people have the assembler. Various Win32 compilation fixes and
20009 update to the INSTALL.W32 file with (hopefully) more accurate Win32
20010 build instructions.
20014 * Modify configure script 'Configure' to automatically create crypto/date.h
20015 file under Win32 and also build pem.h from pem.org. New script
20016 util/mkfiles.pl to create the MINFO file on environments that can't do a
20017 'make files': perl util/mkfiles.pl >MINFO should work.
20021 * Major rework of DES function declarations, in the pursuit of correctness
20022 and purity. As a result, many evil casts evaporated, and some weirdness,
20023 too. You may find this causes warnings in your code. Zapping your evil
20024 casts will probably fix them. Mostly.
20028 * Fix for a typo in asn1.h. Bug fix to object creation script
20029 obj_dat.pl. It considered a zero in an object definition to mean
20030 "end of object": none of the objects in objects.h have any zeros
20031 so it wasn't spotted.
20033 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
20035 * Add support for Triple DES Cipher Block Chaining with Output Feedback
20036 Masking (CBCM). In the absence of test vectors, the best I have been able
20037 to do is check that the decrypt undoes the encrypt, so far. Send me test
20038 vectors if you have them.
20042 * Correct calculation of key length for export ciphers (too much space was
20043 allocated for null ciphers). This has not been tested!
20047 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
20048 message is now correct (it understands "crypto" and "ssl" on its
20049 command line). There is also now an "update" option. This will update
20050 the util/ssleay.num and util/libeay.num files with any new functions.
20052 perl util/mkdef.pl crypto ssl update
20053 it will update them.
20057 * Overhauled the Perl interface:
20058 - ported BN stuff to OpenSSL's different BN library
20059 - made the perl/ source tree CVS-aware
20060 - renamed the package from SSLeay to OpenSSL (the files still contain
20061 their history because I've copied them in the repository)
20062 - removed obsolete files (the test scripts will be replaced
20063 by better Test::Harness variants in the future)
20065 *Ralf S. Engelschall*
20067 * First cut for a very conservative source tree cleanup:
20068 1. merge various obsolete readme texts into doc/ssleay.txt
20069 where we collect the old documents and readme texts.
20070 2. remove the first part of files where I'm already sure that we no
20071 longer need them because of three reasons: either they are just temporary
20072 files which were left by Eric or they are preserved original files where
20073 I've verified that the diff is also available in the CVS via "cvs diff
20074 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
20075 the crypto/md/ stuff).
20077 *Ralf S. Engelschall*
20079 * More extension code. Incomplete support for subject and issuer alt
20080 name, issuer and authority key id. Change the i2v function parameters
20081 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
20082 what that's for :-) Fix to ASN1 macro which messed up
20083 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
20087 * Preliminary support for ENUMERATED type. This is largely copied from the
20092 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
20094 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20096 * Make sure `make rehash` target really finds the `openssl` program.
20098 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
20100 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
20101 like to hear about it if this slows down other processors.
20105 * Add CygWin32 platform information to Configure script.
20107 *Alan Batie <batie@aahz.jf.intel.com>*
20109 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
20111 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
20113 * New program nseq to manipulate netscape certificate sequences
20117 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
20122 * Fixes to BN code. Previously the default was to define BN_RECURSION
20123 but the BN code had some problems that would cause failures when
20124 doing certificate verification and some other functions.
20126 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20128 * Add ASN1 and PEM code to support netscape certificate sequences.
20132 * Add ASN1 and PEM code to support netscape certificate sequences.
20136 * Add several PKIX and private extended key usage OIDs.
20140 * Modify the 'ca' program to handle the new extension code. Modify
20141 openssl.cnf for new extension format, add comments.
20145 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
20146 and add a sample to openssl.cnf so req -x509 now adds appropriate
20151 * Continued X509 V3 changes. Add to other makefiles, integrate with the
20152 error code, add initial support to X509_print() and x509 application.
20156 * Takes a deep breath and start adding X509 V3 extension support code. Add
20157 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
20158 stuff is currently isolated and isn't even compiled yet.
20162 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
20163 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
20164 Removed the versions check from X509 routines when loading extensions:
20165 this allows certain broken certificates that don't set the version
20166 properly to be processed.
20170 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
20171 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
20172 can still be regenerated with "make depend".
20176 * Spelling mistake in C version of CAST-128.
20178 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
20180 * Changes to the error generation code. The perl script err-code.pl
20181 now reads in the old error codes and retains the old numbers, only
20182 adding new ones if necessary. It also only changes the .err files if new
20183 codes are added. The makefiles have been modified to only insert errors
20184 when needed (to avoid needlessly modifying header files). This is done
20185 by only inserting errors if the .err file is newer than the auto generated
20186 C file. To rebuild all the error codes from scratch (the old behaviour)
20187 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
20188 or delete all the .err files.
20192 * CAST-128 was incorrectly implemented for short keys. The C version has
20193 been fixed, but is untested. The assembler versions are also fixed, but
20194 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20195 to regenerate it if needed.
20196 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20197 Hagino <itojun@kame.net>*
20199 * File was opened incorrectly in randfile.c.
20201 *Ulf Möller <ulf@fitug.de>*
20203 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20204 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20205 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20206 al: it's just almost always a UTCTime. Note this patch adds new error
20207 codes so do a "make errors" if there are problems.
20211 * Correct Linux 1 recognition in config.
20213 *Ulf Möller <ulf@fitug.de>*
20215 * Remove pointless MD5 hash when using DSA keys in ca.
20217 *Anonymous <nobody@replay.com>*
20219 * Generate an error if given an empty string as a cert directory. Also
20220 generate an error if handed NULL (previously returned 0 to indicate an
20221 error, but didn't set one).
20223 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20225 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20229 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20230 parameters. This was causing a warning which killed off the Win32 compile.
20234 * Remove C++ style comments from crypto/bn/bn_local.h.
20236 *Neil Costigan <neil.costigan@celocom.com>*
20238 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20239 based on a text string, looking up short and long names and finally
20240 "dot" format. The "dot" format stuff didn't work. Added new function
20241 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20242 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20243 OID is not part of the table.
20247 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20248 X509_LOOKUP_by_alias().
20252 * Sort openssl functions by name.
20256 * Get the `gendsa` command working and add it to the `list` command. Remove
20257 encryption from sample DSA keys (in case anyone is interested the password
20262 * Make *all* `*_free` functions accept a NULL pointer.
20264 *Frans Heymans <fheymans@isaserver.be>*
20266 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20269 *Anonymous <nobody@replay.com>*
20271 * s_server should send the CAfile as acceptable CAs, not its own cert.
20273 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20275 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20277 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20279 * Temp key "for export" tests were wrong in s3_srvr.c.
20281 *Anonymous <nobody@replay.com>*
20283 * Add prototype for temp key callback functions
20284 SSL_CTX_set_tmp_{rsa,dh}_callback().
20288 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20289 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20293 * X509_name_add_entry() freed the wrong thing after an error.
20295 *Arne Ansper <arne@ats.cyber.ee>*
20297 * rsa_eay.c would attempt to free a NULL context.
20299 *Arne Ansper <arne@ats.cyber.ee>*
20301 * BIO_s_socket() had a broken should_retry() on Windoze.
20303 *Arne Ansper <arne@ats.cyber.ee>*
20305 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20307 *Arne Ansper <arne@ats.cyber.ee>*
20309 * Make sure the already existing X509_STORE->depth variable is initialized
20310 in X509_STORE_new(), but document the fact that this variable is still
20311 unused in the certificate verification process.
20313 *Ralf S. Engelschall*
20315 * Fix the various library and `apps/` files to free up pkeys obtained from
20316 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20320 * Fix reference counting in X509_PUBKEY_get(). This makes
20321 demos/maurice/example2.c work, amongst others, probably.
20323 *Steve Henson and Ben Laurie*
20325 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20326 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20327 are no longer created. This way we have a single and consistent command
20328 line interface `openssl <command>`, similar to `cvs <command>`.
20330 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20332 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20333 BIT STRING wrapper always have zero unused bits.
20337 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20341 * Make the top-level INSTALL documentation easier to understand.
20345 * Makefiles updated to exit if an error occurs in a sub-directory
20346 make (including if user presses ^C) [Paul Sutton]
20348 * Make Montgomery context stuff explicit in RSA data structure.
20352 * Fix build order of pem and err to allow for generated pem.h.
20356 * Fix renumbering bug in X509_NAME_delete_entry().
20360 * Enhanced the err-ins.pl script so it makes the error library number
20361 global and can add a library name. This is needed for external ASN1 and
20362 other error libraries.
20366 * Fixed sk_insert which never worked properly.
20370 * Fix ASN1 macros so they can handle indefinite length constructed
20371 EXPLICIT tags. Some non standard certificates use these: they can now
20376 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20377 into a single doc/ssleay.txt bundle. This way the information is still
20378 preserved but no longer messes up this directory. Now it's new room for
20379 the new set of documentation files.
20381 *Ralf S. Engelschall*
20383 * SETs were incorrectly DER encoded. This was a major pain, because they
20384 shared code with SEQUENCEs, which aren't coded the same. This means that
20385 almost everything to do with SETs or SEQUENCEs has either changed name or
20386 number of arguments.
20388 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20390 * Fix test data to work with the above.
20394 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20395 was already fixed by Eric for 0.9.1 it seems.
20397 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20399 * Autodetect FreeBSD3.
20403 * Fix various bugs in Configure. This affects the following platforms:
20407 unixware-2.0-pentium
20412 * Eliminate generated files from CVS. Reorder tests to regenerate files
20413 before they are needed.
20417 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20421 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20423 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20424 changed SSLeay to OpenSSL in version strings.
20426 *Ralf S. Engelschall*
20428 * Some fixups to the top-level documents.
20432 * Fixed the nasty bug where rsaref.h was not found under compile-time
20433 because the symlink to include/ was missing.
20435 *Ralf S. Engelschall*
20437 * Incorporated the popular no-RSA/DSA-only patches
20438 which allow to compile an RSA-free SSLeay.
20440 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20442 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20443 when "ssleay" is still not found.
20445 *Ralf S. Engelschall*
20447 * Added more platforms to Configure: Cray T3E, HPUX 11,
20449 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20451 * Updated the README file.
20453 *Ralf S. Engelschall*
20455 * Added various .cvsignore files in the CVS repository subdirs
20456 to make a "cvs update" really silent.
20458 *Ralf S. Engelschall*
20460 * Recompiled the error-definition header files and added
20461 missing symbols to the Win32 linker tables.
20463 *Ralf S. Engelschall*
20465 * Cleaned up the top-level documents;
20466 o new files: CHANGES and LICENSE
20467 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20468 o merged COPYRIGHT into LICENSE
20469 o removed obsolete TODO file
20470 o renamed MICROSOFT to INSTALL.W32
20472 *Ralf S. Engelschall*
20474 * Removed dummy files from the 0.9.1b source tree:
20475 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20476 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20477 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20478 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20479 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20481 *Ralf S. Engelschall*
20483 * Added various platform portability fixes.
20487 * The Genesis of the OpenSSL rpject:
20488 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20489 Young and Tim J. Hudson created while they were working for C2Net until
20492 *The OpenSSL Project*
20494 ### Changes between 0.9.0b and 0.9.1b [not released]
20496 * Updated a few CA certificates under certs/
20500 * Changed some BIGNUM api stuff.
20504 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20505 DGUX x86, Linux Alpha, etc.
20509 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20510 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20515 * Add -strparse option to asn1pars program which parses nested
20518 *Dr Stephen Henson <shenson@bigfoot.com>*
20520 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20524 * DSA fix for "ca" program.
20528 * Added "-genkey" option to "dsaparam" program.
20532 * Added RIPE MD160 (rmd160) message digest.
20536 * Added -a (all) option to "ssleay version" command.
20540 * Added PLATFORM define which is the id given to Configure.
20544 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20548 * Extended the ASN.1 parser routines.
20552 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20556 * Added a BN_CTX to the BN library.
20560 * Fixed the weak key values in DES library
20564 * Changed API in EVP library for cipher aliases.
20568 * Added support for RC2/64bit cipher.
20572 * Converted the lhash library to the crypto/mem.c functions.
20576 * Added more recognized ASN.1 object ids.
20580 * Added more RSA padding checks for SSL/TLS.
20584 * Added BIO proxy/filter functionality.
20588 * Added extra_certs to SSL_CTX which can be used
20589 send extra CA certificates to the client in the CA cert chain sending
20590 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20594 * Now Fortezza is denied in the authentication phase because
20595 this is key exchange mechanism is not supported by SSLeay at all.
20599 * Additional PKCS1 checks.
20603 * Support the string "TLSv1" for all TLS v1 ciphers.
20607 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20608 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20612 * Fixed a few memory leaks.
20616 * Fixed various code and comment typos.
20620 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20621 bytes sent in the client random.
20623 *Edward Bishop <ebishop@spyglass.com>*
20627 [CVE-2024-2511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-2511
20628 [CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727
20629 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20630 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20631 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20632 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20633 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20634 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20635 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20636 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20637 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20638 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20639 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20640 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20641 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20642 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20643 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20644 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20645 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20646 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20647 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20648 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20649 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20650 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20651 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20652 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20653 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20654 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20655 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20656 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20657 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20658 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20659 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20660 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20661 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20662 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20663 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20664 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20665 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20666 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20667 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20668 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20669 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20670 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20671 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20672 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20673 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20674 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20675 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20676 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20677 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20678 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20679 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20680 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20681 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20682 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20683 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20684 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20685 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20686 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20687 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20688 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20689 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20690 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20691 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20692 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20693 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20694 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20695 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20696 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20697 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20698 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20699 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20700 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20701 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20702 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20703 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20704 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20705 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20706 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20707 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20708 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20709 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20710 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20711 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20712 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20713 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20714 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20715 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20716 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20717 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20718 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20719 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20720 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20721 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20722 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20723 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20724 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20725 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20726 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20727 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20728 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20729 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20730 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20731 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20732 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20733 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20734 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20735 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20736 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20737 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20738 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20739 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20740 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20741 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20742 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20743 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20744 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20745 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20746 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20747 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20748 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20749 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20750 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20751 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20752 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20753 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20754 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20755 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20756 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20757 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20758 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20759 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20760 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20761 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20762 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20763 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20764 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20765 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20766 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20767 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20768 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20769 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20770 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20771 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20772 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20773 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20774 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20775 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20776 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20777 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20778 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20779 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20780 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20781 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20782 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20783 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20784 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20785 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20786 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20787 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20788 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20789 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20790 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20791 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20792 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20793 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20794 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20795 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20796 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20797 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20798 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20799 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20800 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20801 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20802 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20803 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20804 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20805 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20806 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20807 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20808 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20809 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20810 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20811 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20812 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20813 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20814 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20815 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655