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.
39 * The X25519 and X448 key exchange implementation in the FIPS provider
40 is unapproved and has `fips=no` property.
44 * SHAKE-128 and SHAKE-256 implementations have no default digest length
45 anymore. That means these algorithms cannot be used with
46 EVP_DigestFinal/_ex() unless the `xoflen` param is set before.
48 This change was necessary because the preexisting default lengths were
49 half the size necessary for full collision resistance supported by these
54 * Setting `config_diagnostics=1` in the config file will cause errors to
55 be returned from SSL_CTX_new() and SSL_CTX_new_ex() if there is an error
56 in the ssl module configuration.
60 * Use an empty renegotiate extension in TLS client hellos instead of
61 the empty renegotiation SCSV, for all connections with a minimum TLS
66 * Added support for integrity-only cipher suites TLS_SHA256_SHA256 and
67 TLS_SHA384_SHA384 in TLS 1.3, as defined in RFC 9150.
69 This work was sponsored by Siemens AG.
73 * Added support for requesting CRL in CMP.
75 This work was sponsored by Siemens AG.
79 * Added Attribute Certificate (RFC 5755) support. Attribute
80 Certificates can be created, parsed, modified and printed via the
81 public API. There is no command-line tool support at this time.
83 *Damian Hobson-Garcia*
85 * Added support to build Position Independent Executables (PIE). Configuration
86 option `enable-pie` configures the cflag '-fPIE' and ldflag '-pie' to
87 support Address Space Layout Randomization (ASLR) in the openssl executable,
88 removes reliance on external toolchain configurations.
95 ### Changes between 3.3.0 and 3.3.1 [xx XXX xxxx]
97 * Fixed an issue where checking excessively long DSA keys or parameters may
100 Applications that use the functions EVP_PKEY_param_check() or
101 EVP_PKEY_public_check() to check a DSA public key or DSA parameters may
102 experience long delays. Where the key or parameters that are being checked
103 have been obtained from an untrusted source this may lead to a Denial of
106 To resolve this issue DSA keys larger than OPENSSL_DSA_MAX_MODULUS_BITS
107 will now fail the check immediately with a DSA_R_MODULUS_TOO_LARGE error
114 ### Changes between 3.2 and 3.3.0 [9 Apr 2024]
116 * The `-verify` option to the `openssl crl` and `openssl req` will make
117 the program exit with 1 on failure.
121 * The BIO_get_new_index() function can only be called 127 times before it
122 reaches its upper bound of BIO_TYPE_MASK. It will now correctly return an
123 error of -1 once it is exhausted. Users may need to reserve using this
124 function for cases where BIO_find_type() is required. Either BIO_TYPE_NONE
125 or BIO_get_new_index() can be used to supply a type to BIO_meth_new().
129 * Added API functions SSL_SESSION_get_time_ex(), SSL_SESSION_set_time_ex()
130 using time_t which is Y2038 safe on 32 bit systems when 64 bit time
131 is enabled (e.g via setting glibc macro _TIME_BITS=64).
135 * The d2i_ASN1_GENERALIZEDTIME(), d2i_ASN1_UTCTIME(), ASN1_TIME_check(), and
136 related functions have been augmented to check for a minimum length of
137 the input string, in accordance with ITU-T X.690 section 11.7 and 11.8.
141 * Unknown entries in TLS SignatureAlgorithms, ClientSignatureAlgorithms
142 config options and the respective calls to SSL[_CTX]_set1_sigalgs() and
143 SSL[_CTX]_set1_client_sigalgs() that start with `?` character are
144 ignored and the configuration will still be used.
146 Similarly unknown entries that start with `?` character in a TLS
147 Groups config option or set with SSL[_CTX]_set1_groups_list() are ignored
148 and the configuration will still be used.
150 In both cases if the resulting list is empty, an error is returned.
154 * The EVP_PKEY_fromdata function has been augmented to allow for the derivation
155 of CRT (Chinese Remainder Theorem) parameters when requested. See the
156 OSSL_PKEY_PARAM_RSA_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
160 * The activate and soft_load configuration settings for providers in
161 openssl.cnf have been updated to require a value of [1|yes|true|on]
162 (in lower or UPPER case) to enable the setting. Conversely a value
163 of [0|no|false|off] will disable the setting. All other values, or the
164 omission of a value for these settings will result in an error.
168 * Added `-set_issuer` and `-set_subject` options to `openssl x509` to
169 override the Issuer and Subject when creating a certificate. The `-subj`
170 option now is an alias for `-set_subject`.
172 *Job Snijders, George Michaelson*
174 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
175 if called with a NULL stack argument.
179 * In `openssl speed`, changed the default hash function used with `hmac` from
184 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
185 - `certProfile` request message header and respective `-profile` CLI option
186 - support for delayed delivery of all types of response messages
190 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
191 be less hard coded in the build file templates, and to allow easier
192 addition of more exporters. With that, an exporter for CMake is also
197 * The BLAKE2s hash algorithm matches BLAKE2b's support
198 for configurable output length.
200 *Ahelenia Ziemiańska*
202 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
203 server to prefer session resumption using PSK-only key exchange over PSK
204 with DHE, if both are available.
206 *Markus Minichmayr, Tapkey GmbH*
208 * New API `SSL_write_ex2`, which can be used to send an end-of-stream (FIN)
209 condition in an optimised way when using QUIC.
213 * New atexit configuration switch, which controls whether the OPENSSL_cleanup
214 is registered when libcrypto is unloaded. This is turned off on NonStop
215 configurations because of loader differences on that platform compared to
220 * Support for qlog for tracing QUIC connections has been added.
222 The qlog output from OpenSSL currently uses a pre-standard draft version of
223 qlog. The output from OpenSSL will change in incompatible ways in future
224 releases, and is not subject to any format stability or compatibility
225 guarantees at this time. This functionality can be
226 disabled with the build-time option `no-unstable-qlog`. See the
227 openssl-qlog(7) manpage for details.
231 * Added APIs to allow configuring the negotiated idle timeout for QUIC
232 connections, and to allow determining the number of additional streams
233 that can currently be created for a QUIC connection.
237 * Added APIs to allow disabling implicit QUIC event processing for
238 QUIC SSL objects, allowing applications to control when event handling
239 occurs. Refer to the SSL_get_value_uint(3) manpage for details.
243 * Limited support for polling of QUIC connection and stream objects in a
244 non-blocking manner. Refer to the SSL_poll(3) manpage for details.
248 * Added APIs to allow querying the size and utilisation of a QUIC stream's
249 write buffer. Refer to the SSL_get_value_uint(3) manpage for details.
253 * New limit on HTTP response headers is introduced to HTTP client. The
254 default limit is set to 256 header lines. If limit is exceeded the
255 response processing stops with error HTTP_R_RESPONSE_TOO_MANY_HDRLINES.
256 Application may call OSSL_HTTP_REQ_CTX_set_max_response_hdr_lines(3)
257 to change the default. Setting the value to 0 disables the limit.
259 *Alexandr Nedvedicky*
261 * Applied AES-GCM unroll8 optimisation to Microsoft Azure Cobalt 100
265 * Added X509_STORE_get1_objects to avoid issues with the existing
266 X509_STORE_get0_objects API in multi-threaded applications. Refer to the
267 documentation for details.
271 * Added assembly implementation for md5 on loongarch64
275 * Optimized AES-CTR for ARM Neoverse V1 and V2
279 * Enable AES and SHA3 optimisations on Applie Silicon M3-based MacOS systems
284 * Added a new EVP_DigestSqueeze() API. This allows SHAKE to squeeze multiple
285 times with different output sizes.
287 *Shane Lontis, Holger Dengler*
289 * Various optimizations for cryptographic routines using RISC-V vector crypto
292 *Christoph Müllner, Charalampos Mitrodimas, Ard Biesheuvel, Phoebe Chen,
295 * Accept longer context for TLS 1.2 exporters
297 While RFC 5705 implies that the maximum length of a context for exporters is
298 65535 bytes as the length is embedded in uint16, the previous implementation
299 enforced a much smaller limit, which is less than 1024 bytes. This
300 restriction has been removed.
307 ### Changes between 3.2.1 and 3.2.2 [xx XXX xxxx]
309 * Fixed an issue where some non-default TLS server configurations can cause
310 unbounded memory growth when processing TLSv1.3 sessions. An attacker may
311 exploit certain server configurations to trigger unbounded memory growth that
312 would lead to a Denial of Service
314 This problem can occur in TLSv1.3 if the non-default SSL_OP_NO_TICKET option
315 is being used (but not if early_data is also configured and the default
316 anti-replay protection is in use). In this case, under certain conditions,
317 the session cache can get into an incorrect state and it will fail to flush
318 properly as it fills. The session cache will continue to grow in an unbounded
319 manner. A malicious client could deliberately create the scenario for this
320 failure to force a Denial of Service. It may also happen by accident in
327 * Fixed bug where SSL_export_keying_material() could not be used with QUIC
328 connections. (#23560)
332 ### Changes between 3.2.0 and 3.2.1 [30 Jan 2024]
334 * A file in PKCS12 format can contain certificates and keys and may come from
335 an untrusted source. The PKCS12 specification allows certain fields to be
336 NULL, but OpenSSL did not correctly check for this case. A fix has been
337 applied to prevent a NULL pointer dereference that results in OpenSSL
338 crashing. If an application processes PKCS12 files from an untrusted source
339 using the OpenSSL APIs then that application will be vulnerable to this
340 issue prior to this fix.
342 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
343 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
344 and PKCS12_newpass().
346 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
347 function is related to writing data we do not consider it security
354 * When function EVP_PKEY_public_check() is called on RSA public keys,
355 a computation is done to confirm that the RSA modulus, n, is composite.
356 For valid RSA keys, n is a product of two or more large primes and this
357 computation completes quickly. However, if n is an overly large prime,
358 then this computation would take a long time.
360 An application that calls EVP_PKEY_public_check() and supplies an RSA key
361 obtained from an untrusted source could be vulnerable to a Denial of Service
364 The function EVP_PKEY_public_check() is not called from other OpenSSL
365 functions however it is called from the OpenSSL pkey command line
366 application. For that reason that application is also vulnerable if used
367 with the "-pubin" and "-check" options on untrusted data.
369 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
370 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
376 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
377 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
382 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
383 for PowerPC CPUs saves the contents of vector registers in different
384 order than they are restored. Thus the contents of some of these vector
385 registers is corrupted when returning to the caller. The vulnerable code is
386 used only on newer PowerPC processors supporting the PowerISA 2.07
389 The consequences of this kind of internal application state corruption can
390 be various - from no consequences, if the calling application does not
391 depend on the contents of non-volatile XMM registers at all, to the worst
392 consequences, where the attacker could get complete control of the
393 application process. However unless the compiler uses the vector registers
394 for storing pointers, the most likely consequence, if any, would be an
395 incorrect result of some application dependent calculations or a crash
396 leading to a denial of service.
402 * Disable building QUIC server utility when OpenSSL is configured with
407 ### Changes between 3.1 and 3.2.0 [23 Nov 2023]
409 * Fix excessive time spent in DH check / generation with large Q parameter
412 Applications that use the functions DH_generate_key() to generate an
413 X9.42 DH key may experience long delays. Likewise, applications that use
414 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
415 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
416 Where the key or parameters that are being checked have been obtained from
417 an untrusted source this may lead to a Denial of Service.
423 * The BLAKE2b hash algorithm supports a configurable output length
424 by setting the "size" parameter.
426 *Čestmír Kalina and Tomáš Mráz*
428 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
432 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
433 and the corresponding provider-storemgmt API function
434 OSSL_FUNC_store_delete().
438 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
439 a passphrase callback when opening a store.
443 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
444 from 8 bytes to 16 bytes.
445 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
446 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
447 requires a salt length of 128 bits. This affects OpenSSL command line
448 applications such as "genrsa" and "pkcs8" and API's such as
449 PEM_write_bio_PrivateKey() that are reliant on the default value.
450 The additional commandline option 'saltlen' has been added to the
451 OpenSSL command line applications for "pkcs8" and "enc" to allow the
452 salt length to be set to a non default value.
456 * Changed the default value of the `ess_cert_id_alg` configuration
457 option which is used to calculate the TSA's public key certificate
458 identifier. The default algorithm is updated to be sha256 instead
461 *Małgorzata Olszówka*
463 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
464 table for point multiplication of the base point, which increases the size of
465 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
466 been added to disable the precomputed table.
470 * Added client side support for QUIC
472 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
474 * Added multiple tutorials on the OpenSSL library and in particular
475 on writing various clients (using TLS and QUIC protocols) with libssl.
479 * Added secp384r1 implementation using Solinas' reduction to improve
480 speed of the NIST P-384 elliptic curve. To enable the implementation
481 the build option `enable-ec_nistp_64_gcc_128` must be used.
485 * Improved RFC7468 compliance of the asn1parse command.
487 *Matthias St. Pierre*
489 * Added SHA256/192 algorithm support.
493 * Added support for securely getting root CA certificate update in
498 * Improved contention on global write locks by using more read locks where
503 * Improved performance of OSSL_PARAM lookups in performance critical
508 * Added the SSL_get0_group_name() function to provide access to the
509 name of the group used for the TLS key exchange.
513 * Provide a new configure option `no-http` that can be used to disable the
514 HTTP support. Provide new configure options `no-apps` and `no-docs` to
515 disable building the openssl command line application and the documentation.
519 * Provide a new configure option `no-ecx` that can be used to disable the
520 X25519, X448, and EdDSA support.
524 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
525 the EVP_KDF_CTX_set_params() function they are now concatenated not just
526 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
530 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
531 the provider context as a parameter.
535 * TLS round-trip time calculation was added by a Brigham Young University
536 Capstone team partnering with Sandia National Laboratories. A new function
537 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
542 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
543 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
544 option. Use of the "advanced" s_client command command via the "-adv" option
549 * Added an "advanced" command mode to s_client. Use this with the "-adv"
550 option. The old "basic" command mode recognises certain letters that must
551 always appear at the start of a line and cannot be escaped. The advanced
552 command mode enables commands to be entered anywhere and there is an
553 escaping mechanism. After starting s_client with "-adv" type "{help}"
554 to show a list of available commands.
558 * Add Raw Public Key (RFC7250) support. Authentication is supported
559 by matching keys against either local policy (TLSA records synthesised
560 from the expected keys) or DANE (TLSA records obtained by the
561 application from DNS). TLSA records will also match the same key in
562 the server certificate, should RPK use not happen to be negotiated.
566 * Added support for modular exponentiation and CRT offloading for the
571 * Added further assembler code for the RISC-V architecture.
575 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
576 from a given EC_GROUP.
580 * Improved support for non-default library contexts and property queries
581 when parsing PKCS#12 files.
585 * Implemented support for all five instances of EdDSA from RFC8032:
586 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
587 The streaming is not yet supported for the HashEdDSA variants
588 (Ed25519ph and Ed448ph).
592 * Added SM4 optimization for ARM processors using ASIMD and AES HW
597 * Implemented SM4-XTS support.
601 * Added platform-agnostic OSSL_sleep() function.
605 * Implemented deterministic ECDSA signatures (RFC6979) support.
609 * Implemented AES-GCM-SIV (RFC8452) support.
613 * Added support for pluggable (provider-based) TLS signature algorithms.
614 This enables TLS 1.3 authentication operations with algorithms embedded
615 in providers not included by default in OpenSSL. In combination with
616 the already available pluggable KEM and X.509 support, this enables
617 for example suitable providers to deliver post-quantum or quantum-safe
618 cryptography to OpenSSL users.
622 * Added support for pluggable (provider-based) CMS signature algorithms.
623 This enables CMS sign and verify operations with algorithms embedded
624 in providers not included by default in OpenSSL.
628 * Added support for Hybrid Public Key Encryption (HPKE) as defined
629 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
630 Message Layer Security (MLS) and other IETF specifications.
631 HPKE can also be used by other applications that require
632 encrypting "to" an ECDH public key. External APIs are defined in
633 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
637 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
642 * Add support for certificate compression (RFC8879), including
643 library support for Brotli and Zstandard compression.
647 * Add the ability to add custom attributes to PKCS12 files. Add a new API
648 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
649 for a user specified callback and optional argument.
650 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
651 added to the existing STACK_OF attrs.
655 * Major refactor of the libssl record layer.
659 * Add a mac salt length option for the pkcs12 command.
663 * Add more SRTP protection profiles from RFC8723 and RFC8269.
667 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
669 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
671 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
672 supported and enabled.
676 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
677 to the list of ciphersuites providing Perfect Forward Secrecy as
678 required by SECLEVEL >= 3.
680 *Dmitry Belyavskiy, Nicola Tuveri*
682 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
683 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
684 SSL_get0_iana_groups() function-like macro, retrieves the list of
685 supported groups sent by the peer.
686 The function SSL_client_hello_get_extension_order() populates
687 a caller-supplied array with the list of extension types present in the
688 ClientHello, in order of appearance.
692 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
693 to make it possible to use empty passphrase strings.
697 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
701 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
702 to change functions used for allocating the memory of asynchronous call stack.
706 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
710 * A failure exit code is returned when using the openssl x509 command to check
711 certificate attributes and the checks fail.
715 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
716 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
717 of 160 bits and above and less than 224 bits were previously accepted by
718 default but are now no longer allowed. By default TLS compression was
719 already disabled in previous OpenSSL versions. At security level 2 it cannot
724 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
729 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
730 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
731 will need to load the legacy crypto provider.
735 * CCM8 cipher suites in TLS have been downgraded to security level zero
736 because they use a short authentication tag which lowers their strength.
740 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
741 by default. Also spaces surrounding `=` in DN output are removed.
745 * Add X.509 certificate codeSigning purpose and related checks on key usage and
746 extended key usage of the leaf certificate according to the CA/Browser Forum.
750 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
751 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
752 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
753 X.509 version 3 if the certificate information includes X.509 extensions.
757 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
758 such as adding a trace facility for debugging certificate chain building.
762 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
763 in particular supporting requests for central key generation, generalized
764 polling, and various types of genm/genp exchanges defined in CMP Updates.
768 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
769 like correcting the TLS and proxy support and adding tracing for debugging.
773 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
777 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
778 a certificate to be added is already present. `CMS_sign_ex()` and
779 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
780 and no longer throw an error for them.
784 * Fixed and extended `util/check-format.pl` for checking adherence to the
785 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
786 The checks are meanwhile more complete and yield fewer false positives.
790 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
791 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
792 calls. They can be used as the transport BIOs for QUIC.
794 *Hugo Landau, Matt Caswell and Tomáš Mráz*
796 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
797 sending and receiving multiple messages in a single call. An implementation
798 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
802 * Support for loading root certificates from the Windows certificate store
803 has been added. The support is in the form of a store which recognises the
804 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
805 arguments. This store is built by default and can be disabled using the new
806 compile-time option `no-winstore`. This store is not currently used by
807 default and must be loaded explicitly using the above store URI. It is
808 expected to be loaded by default in the future.
812 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
813 kernel versions that support KTLS have a known bug in CCM processing. That
814 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
815 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
820 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
825 * Zerocopy KTLS sendfile() support on Linux.
829 * The OBJ_ calls are now thread safe using a global lock.
833 * New parameter `-digest` for openssl cms command allowing signing
834 pre-computed digests and new CMS API functions supporting that
839 * OPENSSL_malloc() and other allocation functions now raise errors on
840 allocation failures. The callers do not need to explicitly raise errors
841 unless they want to for tracing purposes.
845 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
846 decryption as a protection against Bleichenbacher-like attacks.
847 The RSA decryption API will now return a randomly generated deterministic
848 message instead of an error in case it detects an error when checking
849 padding during PKCS#1 v1.5 decryption. This is a general protection against
850 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
852 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
853 on the RSA decryption context.
857 * Added support for Brainpool curves in TLS-1.3.
859 *Bernd Edlinger and Matt Caswell*
861 * Added OpenBSD specific build targets.
865 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
866 a basic thread pool implementation for select platforms.
873 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
875 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
876 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
877 that alter the key or IV length ([CVE-2023-5363]).
881 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
883 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
885 The POLY1305 MAC (message authentication code) implementation in OpenSSL
886 does not save the contents of non-volatile XMM registers on Windows 64
887 platform when calculating the MAC of data larger than 64 bytes. Before
888 returning to the caller all the XMM registers are set to zero rather than
889 restoring their previous content. The vulnerable code is used only on newer
890 x86_64 processors supporting the AVX512-IFMA instructions.
892 The consequences of this kind of internal application state corruption can
893 be various - from no consequences, if the calling application does not
894 depend on the contents of non-volatile XMM registers at all, to the worst
895 consequences, where the attacker could get complete control of the
896 application process. However given the contents of the registers are just
897 zeroized so the attacker cannot put arbitrary values inside, the most likely
898 consequence, if any, would be an incorrect result of some application
899 dependent calculations or a crash leading to a denial of service.
905 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
907 * Fix excessive time spent checking DH q parameter value.
909 The function DH_check() performs various checks on DH parameters. After
910 fixing CVE-2023-3446 it was discovered that a large q parameter value can
911 also trigger an overly long computation during some of these checks.
912 A correct q value, if present, cannot be larger than the modulus p
913 parameter, thus it is unnecessary to perform these checks if q is larger
916 If DH_check() is called with such q parameter value,
917 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
918 intensive checks are skipped.
924 * Fix DH_check() excessive time with over sized modulus.
926 The function DH_check() performs various checks on DH parameters. One of
927 those checks confirms that the modulus ("p" parameter) is not too large.
928 Trying to use a very large modulus is slow and OpenSSL will not normally use
929 a modulus which is over 10,000 bits in length.
931 However the DH_check() function checks numerous aspects of the key or
932 parameters that have been supplied. Some of those checks use the supplied
933 modulus value even if it has already been found to be too large.
935 A new limit has been added to DH_check of 32,768 bits. Supplying a
936 key/parameters with a modulus over this size will simply cause DH_check() to
943 * Do not ignore empty associated data entries with AES-SIV.
945 The AES-SIV algorithm allows for authentication of multiple associated
946 data entries along with the encryption. To authenticate empty data the
947 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
948 with NULL pointer as the output buffer and 0 as the input buffer length.
949 The AES-SIV implementation in OpenSSL just returns success for such call
950 instead of performing the associated data authentication operation.
951 The empty data thus will not be authenticated. ([CVE-2023-2975])
953 Thanks to Juerg Wullschleger (Google) for discovering the issue.
955 The fix changes the authentication tag value and the ciphertext for
956 applications that use empty associated data entries with AES-SIV.
957 To decrypt data encrypted with previous versions of OpenSSL the application
958 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
963 * When building with the `enable-fips` option and using the resulting
964 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
965 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
966 not operate with truncated digests (FIPS 140-3 IG G.R).
970 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
972 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
973 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
975 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
976 numeric text form. For gigantic sub-identifiers, this would take a very
977 long time, the time complexity being O(n^2) where n is the size of that
978 sub-identifier. ([CVE-2023-2650])
980 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
981 IDENTIFIER to canonical numeric text form if the size of that OBJECT
982 IDENTIFIER is 586 bytes or less, and fail otherwise.
984 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
985 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
986 most 128 sub-identifiers, and that the maximum value that each sub-
987 identifier may have is 2^32-1 (4294967295 decimal).
989 For each byte of every sub-identifier, only the 7 lower bits are part of
990 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
991 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
996 * Multiple algorithm implementation fixes for ARM BE platforms.
1000 * Added a -pedantic option to fipsinstall that adjusts the various
1001 settings to ensure strict FIPS compliance rather than backwards
1006 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
1007 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
1008 trigger a crash of an application using AES-XTS decryption if the memory
1009 just after the buffer being decrypted is not mapped.
1010 Thanks to Anton Romanov (Amazon) for discovering the issue.
1015 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
1016 The previous fix for this timing side channel turned out to cause
1017 a severe 2-3x performance regression in the typical use case
1018 compared to 3.0.7. The new fix uses existing constant time
1019 code paths, and restores the previous performance level while
1020 fully eliminating all existing timing side channels.
1021 The fix was developed by Bernd Edlinger with testing support
1026 * Add FIPS provider configuration option to disallow the use of
1027 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
1028 The option '-no_drbg_truncated_digests' can optionally be
1029 supplied to 'openssl fipsinstall'.
1033 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
1034 that it does not enable policy checking. Thanks to David Benjamin for
1035 discovering this issue.
1040 * Fixed an issue where invalid certificate policies in leaf certificates are
1041 silently ignored by OpenSSL and other certificate policy checks are skipped
1042 for that certificate. A malicious CA could use this to deliberately assert
1043 invalid certificate policies in order to circumvent policy checking on the
1044 certificate altogether.
1049 * Limited the number of nodes created in a policy tree to mitigate
1050 against CVE-2023-0464. The default limit is set to 1000 nodes, which
1051 should be sufficient for most installations. If required, the limit
1052 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
1053 time define to a desired maximum number of nodes or zero to allow
1059 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
1061 * Add FIPS provider configuration option to enforce the
1062 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
1063 The option '-ems-check' can optionally be supplied to
1064 'openssl fipsinstall'.
1068 * The FIPS provider includes a few non-approved algorithms for
1069 backward compatibility purposes and the "fips=yes" property query
1070 must be used for all algorithm fetches to ensure FIPS compliance.
1072 The algorithms that are included but not approved are Triple DES ECB,
1073 Triple DES CBC and EdDSA.
1077 * Added support for KMAC in KBKDF.
1081 * RNDR and RNDRRS support in provider functions to provide
1082 random number generation for Arm CPUs (aarch64).
1086 * s_client and s_server apps now explicitly say when the TLS version
1087 does not include the renegotiation mechanism. This avoids confusion
1088 between that scenario versus when the TLS version includes secure
1089 renegotiation but the peer lacks support for it.
1093 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
1095 *Tomasz Kantecki, Andrey Matyukov*
1097 * The various OBJ_* functions have been made thread safe.
1101 * Parallel dual-prime 1536/2048-bit modular exponentiation for
1102 AVX512_IFMA capable processors.
1104 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1106 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
1107 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
1108 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
1109 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
1110 `OPENSSL_NO_DEPRECATED_3_1`.
1112 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
1113 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
1114 definitions for these functions regardless of whether
1115 `OPENSSL_NO_DEPRECATED_3_1` is defined.
1117 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
1118 functions regardless of whether they are using them. It is recommended that
1119 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
1123 * When generating safe-prime DH parameters set the recommended private key
1124 length equivalent to minimum key lengths as in RFC 7919.
1128 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
1129 maximum size that is smaller or equal to the digest length to comply with
1130 FIPS 186-4 section 5. This is implemented by a new option
1131 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
1132 `rsa_pss_saltlen` parameter, which is now the default. Signature
1133 verification is not affected by this change and continues to work as before.
1140 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
1141 listed here are only a brief description.
1142 The migration guide contains more detailed information related to new features,
1143 breaking changes, and mappings for the large list of deprecated functions.
1145 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
1147 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
1149 * Fixed NULL dereference during PKCS7 data verification.
1151 A NULL pointer can be dereferenced when signatures are being
1152 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
1153 algorithm used for the signature is known to the OpenSSL library but
1154 the implementation of the hash algorithm is not available the digest
1155 initialization will fail. There is a missing check for the return
1156 value from the initialization function which later leads to invalid
1157 usage of the digest API most likely leading to a crash.
1160 PKCS7 data is processed by the SMIME library calls and also by the
1161 time stamp (TS) library calls. The TLS implementation in OpenSSL does
1162 not call these functions however third party applications would be
1163 affected if they call these functions to verify signatures on untrusted
1168 * Fixed X.400 address type confusion in X.509 GeneralName.
1170 There is a type confusion vulnerability relating to X.400 address processing
1171 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
1172 but the public structure definition for GENERAL_NAME incorrectly specified
1173 the type of the x400Address field as ASN1_TYPE. This field is subsequently
1174 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
1175 than an ASN1_STRING.
1177 When CRL checking is enabled (i.e. the application sets the
1178 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
1179 pass arbitrary pointers to a memcmp call, enabling them to read memory
1180 contents or enact a denial of service.
1185 * Fixed NULL dereference validating DSA public key.
1187 An invalid pointer dereference on read can be triggered when an
1188 application tries to check a malformed DSA public key by the
1189 EVP_PKEY_public_check() function. This will most likely lead
1190 to an application crash. This function can be called on public
1191 keys supplied from untrusted sources which could allow an attacker
1192 to cause a denial of service attack.
1194 The TLS implementation in OpenSSL does not call this function
1195 but applications might call the function if there are additional
1196 security requirements imposed by standards such as FIPS 140-3.
1199 *Shane Lontis, Tomáš Mráz*
1201 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
1203 An invalid pointer dereference on read can be triggered when an
1204 application tries to load malformed PKCS7 data with the
1205 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
1207 The result of the dereference is an application crash which could
1208 lead to a denial of service attack. The TLS implementation in OpenSSL
1209 does not call this function however third party applications might
1210 call these functions on untrusted data.
1215 * Fixed Use-after-free following BIO_new_NDEF.
1217 The public API function BIO_new_NDEF is a helper function used for
1218 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
1219 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
1220 be called directly by end user applications.
1222 The function receives a BIO from the caller, prepends a new BIO_f_asn1
1223 filter BIO onto the front of it to form a BIO chain, and then returns
1224 the new head of the BIO chain to the caller. Under certain conditions,
1225 for example if a CMS recipient public key is invalid, the new filter BIO
1226 is freed and the function returns a NULL result indicating a failure.
1227 However, in this case, the BIO chain is not properly cleaned up and the
1228 BIO passed by the caller still retains internal pointers to the previously
1229 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
1230 then a use-after-free will occur. This will most likely result in a crash.
1233 *Viktor Dukhovni, Matt Caswell*
1235 * Fixed Double free after calling PEM_read_bio_ex.
1237 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
1238 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
1239 data. If the function succeeds then the "name_out", "header" and "data"
1240 arguments are populated with pointers to buffers containing the relevant
1241 decoded data. The caller is responsible for freeing those buffers. It is
1242 possible to construct a PEM file that results in 0 bytes of payload data.
1243 In this case PEM_read_bio_ex() will return a failure code but will populate
1244 the header argument with a pointer to a buffer that has already been freed.
1245 If the caller also frees this buffer then a double free will occur. This
1246 will most likely lead to a crash.
1248 The functions PEM_read_bio() and PEM_read() are simple wrappers around
1249 PEM_read_bio_ex() and therefore these functions are also directly affected.
1251 These functions are also called indirectly by a number of other OpenSSL
1252 functions including PEM_X509_INFO_read_bio_ex() and
1253 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
1254 internal uses of these functions are not vulnerable because the caller does
1255 not free the header argument if PEM_read_bio_ex() returns a failure code.
1258 *Kurt Roeckx, Matt Caswell*
1260 * Fixed Timing Oracle in RSA Decryption.
1262 A timing based side channel exists in the OpenSSL RSA Decryption
1263 implementation which could be sufficient to recover a plaintext across
1264 a network in a Bleichenbacher style attack. To achieve a successful
1265 decryption an attacker would have to be able to send a very large number
1266 of trial messages for decryption. The vulnerability affects all RSA padding
1267 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1270 *Dmitry Belyavsky, Hubert Kario*
1272 * Fixed X.509 Name Constraints Read Buffer Overflow.
1274 A read buffer overrun can be triggered in X.509 certificate verification,
1275 specifically in name constraint checking. The read buffer overrun might
1276 result in a crash which could lead to a denial of service attack.
1277 In a TLS client, this can be triggered by connecting to a malicious
1278 server. In a TLS server, this can be triggered if the server requests
1279 client authentication and a malicious client connects.
1284 * Fixed X.509 Policy Constraints Double Locking security issue.
1286 If an X.509 certificate contains a malformed policy constraint and
1287 policy processing is enabled, then a write lock will be taken twice
1288 recursively. On some operating systems (most widely: Windows) this
1289 results in a denial of service when the affected process hangs. Policy
1290 processing being enabled on a publicly facing server is not considered
1291 to be a common setup.
1296 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1297 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1298 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1299 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1300 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1301 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1302 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1303 for legacy EC and SM2 keys is also changed similarly to honor the
1304 equivalent conversion format flag as specified in the underlying
1305 `EC_KEY` object being exported to a provider, when this function is
1306 called through `EVP_PKEY_export()`.
1310 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1312 * Fixed two buffer overflows in punycode decoding functions.
1314 A buffer overrun can be triggered in X.509 certificate verification,
1315 specifically in name constraint checking. Note that this occurs after
1316 certificate chain signature verification and requires either a CA to
1317 have signed the malicious certificate or for the application to continue
1318 certificate verification despite failure to construct a path to a trusted
1321 In a TLS client, this can be triggered by connecting to a malicious
1322 server. In a TLS server, this can be triggered if the server requests
1323 client authentication and a malicious client connects.
1325 An attacker can craft a malicious email address to overflow
1326 an arbitrary number of bytes containing the `.` character (decimal 46)
1327 on the stack. This buffer overflow could result in a crash (causing a
1331 An attacker can craft a malicious email address to overflow four
1332 attacker-controlled bytes on the stack. This buffer overflow could
1333 result in a crash (causing a denial of service) or potentially remote code
1334 execution depending on stack layout for any given platform/compiler.
1339 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1340 parameters in OpenSSL code.
1341 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1342 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1343 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1344 Using these invalid names may cause algorithms to use slower methods
1345 that ignore the CRT parameters.
1349 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1354 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1355 data to be signed before signing the certificate.
1359 * Added RIPEMD160 to the default provider.
1363 * Ensured that the key share group sent or accepted for the key exchange
1364 is allowed for the protocol version.
1368 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1370 * OpenSSL supports creating a custom cipher via the legacy
1371 EVP_CIPHER_meth_new() function and associated function calls. This function
1372 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1373 to use the new provider mechanism in order to implement custom ciphers.
1375 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1376 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1377 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1378 and decryption initialisation functions). Instead of using the custom cipher
1379 directly it incorrectly tries to fetch an equivalent cipher from the
1380 available providers. An equivalent cipher is found based on the NID passed to
1381 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1382 given cipher. However it is possible for an application to incorrectly pass
1383 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1384 is used in this way the OpenSSL encryption/decryption initialisation function
1385 will match the NULL cipher as being equivalent and will fetch this from the
1386 available providers. This will succeed if the default provider has been
1387 loaded (or if a third party provider has been loaded that offers this
1388 cipher). Using the NULL cipher means that the plaintext is emitted as the
1391 Applications are only affected by this issue if they call
1392 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1393 encryption/decryption initialisation function. Applications that only use
1394 SSL/TLS are not impacted by this issue.
1399 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1404 * Fixed the linux-mips64 Configure target which was missing the
1405 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1410 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1415 * Correctly handle a retransmitted ClientHello in DTLS
1419 * Fixed detection of ktls support in cross-compile environment on Linux
1423 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1428 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1429 report correct results in some cases
1433 * Fix UWP builds by defining VirtualLock
1437 * For known safe primes use the minimum key length according to RFC 7919.
1438 Longer private key sizes unnecessarily raise the cycles needed to compute the
1439 shared secret without any increase of the real security. This fixes a
1440 regression from 1.1.1 where these shorter keys were generated for the known
1445 * Added the loongarch64 target
1449 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1450 only passed to the FIPS provider and not to the default or legacy provider.
1454 * Fixed reported performance degradation on aarch64. Restored the
1455 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1456 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1457 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1458 The new algorithm is still used for 32 bit targets.
1462 * Added a missing header for memcmp that caused compilation failure on some
1467 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1469 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1470 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1471 This issue makes the RSA implementation with 2048 bit private keys
1472 incorrect on such machines and memory corruption will happen during
1473 the computation. As a consequence of the memory corruption an attacker
1474 may be able to trigger a remote code execution on the machine performing
1477 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1478 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1479 are affected by this issue.
1484 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1485 implementation would not encrypt the entirety of the data under some
1486 circumstances. This could reveal sixteen bytes of data that was
1487 preexisting in the memory that wasn't written. In the special case of
1488 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1490 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1491 they are both unaffected.
1494 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1496 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1498 * In addition to the c_rehash shell command injection identified in
1499 CVE-2022-1292, further bugs where the c_rehash script does not
1500 properly sanitise shell metacharacters to prevent command injection have been
1503 When the CVE-2022-1292 was fixed it was not discovered that there
1504 are other places in the script where the file names of certificates
1505 being hashed were possibly passed to a command executed through the shell.
1507 This script is distributed by some operating systems in a manner where
1508 it is automatically executed. On such operating systems, an attacker
1509 could execute arbitrary commands with the privileges of the script.
1511 Use of the c_rehash script is considered obsolete and should be replaced
1512 by the OpenSSL rehash command line tool.
1515 *Daniel Fiala, Tomáš Mráz*
1517 * Case insensitive string comparison no longer uses locales. It has instead
1518 been directly implemented.
1522 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1524 * Case insensitive string comparison is reimplemented via new locale-agnostic
1525 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1526 comparison. The previous implementation had problems when the Turkish locale
1531 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1532 metacharacters to prevent command injection. This script is distributed by
1533 some operating systems in a manner where it is automatically executed. On
1534 such operating systems, an attacker could execute arbitrary commands with the
1535 privileges of the script.
1537 Use of the c_rehash script is considered obsolete and should be replaced
1538 by the OpenSSL rehash command line tool.
1543 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1544 certificate on an OCSP response. The bug caused the function in the case
1545 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1546 response (meaning a successful verification) even in the case where the
1547 response signing certificate fails to verify.
1549 It is anticipated that most users of `OCSP_basic_verify` will not use the
1550 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1551 a negative value (indicating a fatal error) in the case of a certificate
1552 verification failure. The normal expected return value in this case would be
1555 This issue also impacts the command line OpenSSL "ocsp" application. When
1556 verifying an ocsp response with the "-no_cert_checks" option the command line
1557 application will report that the verification is successful even though it
1558 has in fact failed. In this case the incorrect successful response will also
1559 be accompanied by error messages showing the failure and contradicting the
1560 apparently successful result.
1565 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1566 AAD data as the MAC key. This made the MAC key trivially predictable.
1568 An attacker could exploit this issue by performing a man-in-the-middle attack
1569 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1570 that the modified data would still pass the MAC integrity check.
1572 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1573 endpoint will always be rejected by the recipient and the connection will
1574 fail at that point. Many application protocols require data to be sent from
1575 the client to the server first. Therefore, in such a case, only an OpenSSL
1576 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1578 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1579 sent in both directions. In this case both clients and servers could be
1580 affected, regardless of the application protocol.
1582 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1583 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1584 the handshake when using this ciphersuite.
1586 The confidentiality of data is not impacted by this issue, i.e. an attacker
1587 cannot decrypt data that has been encrypted using this ciphersuite - they can
1590 In order for this attack to work both endpoints must legitimately negotiate
1591 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1592 OpenSSL 3.0, and is not available within the default provider or the default
1593 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1594 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1595 following must have occurred:
1597 1) OpenSSL must have been compiled with the (non-default) compile time option
1598 enable-weak-ssl-ciphers
1600 2) OpenSSL must have had the legacy provider explicitly loaded (either
1601 through application code or via configuration)
1603 3) The ciphersuite must have been explicitly added to the ciphersuite list
1605 4) The libssl security level must have been set to 0 (default is 1)
1607 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1609 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1610 others that both endpoints have in common
1615 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1616 occupied by the removed hash table entries.
1618 This function is used when decoding certificates or keys. If a long lived
1619 process periodically decodes certificates or keys its memory usage will
1620 expand without bounds and the process might be terminated by the operating
1621 system causing a denial of service. Also traversing the empty hash table
1622 entries will take increasingly more time.
1624 Typically such long lived processes might be TLS clients or TLS servers
1625 configured to accept client certificate authentication.
1628 *Hugo Landau, Aliaksei Levin*
1630 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1631 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1632 statistics are no longer supported. For compatibility, these statistics are
1633 still listed in the output but are now always reported as zero.
1637 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1639 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1640 for non-prime moduli.
1642 Internally this function is used when parsing certificates that contain
1643 elliptic curve public keys in compressed form or explicit elliptic curve
1644 parameters with a base point encoded in compressed form.
1646 It is possible to trigger the infinite loop by crafting a certificate that
1647 has invalid explicit curve parameters.
1649 Since certificate parsing happens prior to verification of the certificate
1650 signature, any process that parses an externally supplied certificate may thus
1651 be subject to a denial of service attack. The infinite loop can also be
1652 reached when parsing crafted private keys as they can contain explicit
1653 elliptic curve parameters.
1655 Thus vulnerable situations include:
1657 - TLS clients consuming server certificates
1658 - TLS servers consuming client certificates
1659 - Hosting providers taking certificates or private keys from customers
1660 - Certificate authorities parsing certification requests from subscribers
1661 - Anything else which parses ASN.1 elliptic curve parameters
1663 Also any other applications that use the BN_mod_sqrt() where the attacker
1664 can control the parameter values are vulnerable to this DoS issue.
1669 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1670 to the list of ciphersuites providing Perfect Forward Secrecy as
1671 required by SECLEVEL >= 3.
1673 *Dmitry Belyavskiy, Nicola Tuveri*
1675 * Made the AES constant time code for no-asm configurations
1676 optional due to the resulting 95% performance degradation.
1677 The AES constant time code can be enabled, for no assembly
1678 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1682 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1687 * The negative return value handling of the certificate verification callback
1688 was reverted. The replacement is to set the verification retry state with
1689 the SSL_set_retry_verify() function.
1693 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1695 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1696 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1697 verify a certificate supplied by a server. That function may return a
1698 negative return value to indicate an internal error (for example out of
1699 memory). Such a negative return value is mishandled by OpenSSL and will cause
1700 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1701 success and a subsequent call to SSL_get_error() to return the value
1702 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1703 returned by OpenSSL if the application has previously called
1704 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1705 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1706 totally unexpected and applications may not behave correctly as a result. The
1707 exact behaviour will depend on the application but it could result in
1708 crashes, infinite loops or other similar incorrect responses.
1710 This issue is made more serious in combination with a separate bug in OpenSSL
1711 3.0 that will cause X509_verify_cert() to indicate an internal error when
1712 processing a certificate chain. This will occur where a certificate does not
1713 include the Subject Alternative Name extension but where a Certificate
1714 Authority has enforced name constraints. This issue can occur even with valid
1720 * Corrected a few file name and file reference bugs in the build,
1721 installation and setup scripts, which lead to installation verification
1722 failures. Slightly enhanced the installation verification script.
1726 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1731 * Fixed PVK encoder to properly query for the passphrase.
1735 * Multiple fixes in the OSSL_HTTP API functions.
1739 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1740 OSSL_PARAM_INTEGER data type and return error on negative numbers
1741 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1742 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1746 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1750 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1754 * Multiple threading fixes.
1758 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1762 * Allow fetching an operation from the provider that owns an unexportable key
1763 as a fallback if that is still allowed by the property query.
1767 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1769 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1774 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1775 S390X capability vector to zero. This simplifies testing of different code
1776 paths on S390X architecture.
1780 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1781 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1782 SP 800-38D". The communication will fail at this point.
1786 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1787 confidential in EC_GROUP data.
1791 * The byte order mark (BOM) character is ignored if encountered at the
1792 beginning of a PEM-formatted file.
1796 * Added CMS support for the Russian GOST algorithms.
1800 * Due to move of the implementation of cryptographic operations
1801 to the providers, validation of various operation parameters can
1802 be postponed until the actual operation is executed where previously
1803 it happened immediately when an operation parameter was set.
1805 For example when setting an unsupported curve with
1806 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1807 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1809 *OpenSSL team members and many third party contributors*
1811 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1812 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1813 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1814 instead to retrieve these algorithms from a provider.
1818 * On build targets where the multilib postfix is set in the build
1819 configuration the libdir directory was changing based on whether
1820 the lib directory with the multilib postfix exists on the system
1821 or not. This unpredictable behavior was removed and eventual
1822 multilib postfix is now always added to the default libdir. Use
1823 `--libdir=lib` to override the libdir if adding the postfix is
1828 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1829 no longer interoperable with OpenSSL 1.1.1.
1833 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1834 function codes, this function can only cause problems for calling
1839 * Add a configurable flag to output date formats as ISO 8601. Does not
1840 change the default date format.
1844 * Version of MSVC earlier than 1300 could get link warnings, which could
1845 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1846 Support for this flag has been removed.
1850 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1851 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1852 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1853 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1854 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1858 * The signatures of the functions to get and set options on SSL and
1859 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1860 Some source code changes may be required.
1864 * The public definitions of conf_method_st and conf_st have been
1865 deprecated. They will be made opaque in a future release.
1867 *Rich Salz and Tomáš Mráz*
1869 * Client-initiated renegotiation is disabled by default. To allow it, use
1870 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1871 flag, or the "ClientRenegotiation" config parameter as appropriate.
1875 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1876 or modify relative pathname inclusion.
1880 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1881 validated. Please consult the README-FIPS and
1882 README-PROVIDERS files, as well as the migration guide.
1884 *OpenSSL team members and many third party contributors*
1886 * For the key types DH and DHX the allowed settable parameters are now different.
1890 * The openssl commands that read keys, certificates, and CRLs now
1891 automatically detect the PEM or DER format of the input files.
1893 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1895 * Added enhanced PKCS#12 APIs which accept a library context.
1899 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1903 * Added support for Kernel TLS (KTLS).
1905 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1907 * Support for RFC 5746 secure renegotiation is now required by default for
1908 SSL or TLS connections to succeed.
1912 * The signature of the `copy` functional parameter of the
1913 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1914 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1915 the signature of the `pub_decode` functional parameter of the
1916 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1917 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1921 * The error return values from some control calls (ctrl) have changed.
1925 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1929 * Many functions in the EVP_ namespace that are getters of values from
1930 implementations or contexts were renamed to include get or get0 in their
1931 names. Old names are provided as macro aliases for compatibility and
1936 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1937 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1938 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1943 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1946 * The output from the command line applications may have minor
1951 * The output from numerous "printing" may have minor changes.
1955 * Windows thread synchronization uses read/write primitives (SRWLock) when
1956 supported by the OS, otherwise CriticalSection continues to be used.
1960 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1961 work on read only BIO source/sinks that do not support these functions.
1962 This allows piping or redirection of a file BIO using stdin to be buffered
1963 into memory. This is used internally in OSSL_DECODER_from_bio().
1967 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1968 this function would return one of the values OSSL_STORE_INFO_NAME,
1969 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1970 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1971 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1972 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1973 using this function should be amended to handle the changed return value.
1977 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1978 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1979 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1980 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1981 contains more than one certificate identifier: This means that all
1982 certificates referenced there MUST be part of the validation chain.
1986 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1987 RC5, DESX and DES have been moved to the legacy provider.
1991 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1992 RIPEMD-160 have been moved to the legacy provider.
1996 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
2001 * The deprecated functions EVP_PKEY_get0_RSA(),
2002 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
2003 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
2004 well as the similarly named "get1" functions behave differently in
2009 * A number of functions handling low-level keys or engines were deprecated
2010 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
2011 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
2012 EVP_PKEY_get0_siphash().
2016 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
2017 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
2018 will need to load the legacy crypto provider. This includes these PBE
2019 algorithms which use this KDF:
2020 - NID_pbeWithMD2AndDES_CBC
2021 - NID_pbeWithMD5AndDES_CBC
2022 - NID_pbeWithSHA1AndRC2_CBC
2023 - NID_pbeWithMD2AndRC2_CBC
2024 - NID_pbeWithMD5AndRC2_CBC
2025 - NID_pbeWithSHA1AndDES_CBC
2029 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
2030 BIO_debug_callback() functions.
2034 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
2035 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
2039 * The RAND_METHOD APIs have been deprecated.
2043 * The SRP APIs have been deprecated.
2047 * Add a compile time option to prevent the caching of provider fetched
2048 algorithms. This is enabled by including the no-cached-fetch option
2049 at configuration time.
2053 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
2054 count of PKCS12_DEFAULT_ITER.
2056 *Tomáš Mráz and Sahana Prasad*
2058 * The openssl speed command does not use low-level API calls anymore.
2062 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
2065 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
2067 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
2071 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
2072 providers may supply their own group implementations (using either the "key
2073 exchange" or the "key encapsulation" methods) which will automatically be
2074 detected and used by libssl.
2076 *Matt Caswell, Nicola Tuveri*
2078 * The undocumented function X509_certificate_type() has been deprecated;
2082 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
2086 * Removed RSA padding mode for SSLv23 (which was only used for
2087 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
2088 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
2093 * Deprecated the obsolete X9.31 RSA key generation related functions.
2095 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
2096 is not allowed to return a value > 1, this is no more taken as failure.
2098 *Viktor Dukhovni and David von Oheimb*
2100 * Deprecated the obsolete X9.31 RSA key generation related functions
2101 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
2102 BN_X931_generate_prime_ex().
2106 * The default key generation method for the regular 2-prime RSA keys was
2107 changed to the FIPS 186-4 B.3.6 method.
2111 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
2115 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
2119 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
2120 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
2122 *Rich Salz, Richard Levitte, and David von Oheimb*
2124 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
2128 * Deprecated `OCSP_parse_url()`.
2132 * Validation of SM2 keys has been separated from the validation of regular EC
2137 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
2138 switches: a validation failure triggers an early exit, returning a failure
2139 exit status to the parent process.
2143 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
2144 to ignore unknown ciphers.
2148 * The `-cipher-commands` and `-digest-commands` options
2149 of the command line utility `list` have been deprecated.
2150 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
2154 * Added convenience functions for generating asymmetric key pairs:
2155 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
2156 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
2160 * All of the low-level EC_KEY functions have been deprecated.
2162 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
2164 * Deprecated all the libcrypto and libssl error string loading
2169 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
2170 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
2175 * The `-crypt` option to the `passwd` command line tool has been removed.
2179 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
2184 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
2188 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
2189 EVP_PKEY_get1_tls_encodedpoint().
2193 * The security callback, which can be customised by application code, supports
2194 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
2195 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
2199 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
2200 interface. Their functionality remains unchanged.
2204 * Added new option for 'openssl list', '-providers', which will display the
2205 list of loaded providers, their names, version and status. It optionally
2206 displays their gettable parameters.
2210 * Removed EVP_PKEY_set_alias_type().
2214 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
2215 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
2219 * Changed all "STACK" functions to be macros instead of inline functions. Macro
2220 parameters are still checked for type safety at compile time via helper
2225 * Remove the RAND_DRBG API
2227 *Paul Dale and Matthias St. Pierre*
2229 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
2230 as well as actual hostnames.
2234 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2235 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2236 conversely, silently ignore DTLS protocol version bounds when configuring
2237 TLS-based contexts. The commands can be repeated to set bounds of both
2238 types. The same applies with the corresponding "min_protocol" and
2239 "max_protocol" command-line switches, in case some application uses both TLS
2242 SSL_CTX instances that are created for a fixed protocol version (e.g.
2243 `TLSv1_server_method()`) also silently ignore version bounds. Previously
2244 attempts to apply bounds to these protocol versions would result in an
2245 error. Now only the "version-flexible" SSL_CTX instances are subject to
2246 limits in configuration files in command-line options.
2250 * Deprecated the `ENGINE` API. Engines should be replaced with providers
2255 * Reworked the recorded ERR codes to make better space for system errors.
2256 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
2257 given code is a system error (true) or an OpenSSL error (false).
2261 * Reworked the test perl framework to better allow parallel testing.
2263 *Nicola Tuveri and David von Oheimb*
2265 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2266 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2270 * 'Configure' has been changed to figure out the configuration target if
2271 none is given on the command line. Consequently, the 'config' script is
2272 now only a mere wrapper. All documentation is changed to only mention
2275 *Rich Salz and Richard Levitte*
2277 * Added a library context `OSSL_LIB_CTX` that applications as well as
2278 other libraries can use to form a separate context within which
2279 libcrypto operations are performed.
2283 * Added various `_ex` functions to the OpenSSL API that support using
2284 a non-default `OSSL_LIB_CTX`.
2288 * Handshake now fails if Extended Master Secret extension is dropped
2293 * Dropped interactive mode from the `openssl` program.
2297 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2299 *David von Oheimb and Shane Lontis*
2301 * Deprecated `EC_METHOD_get_field_type()`.
2305 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2306 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2307 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2311 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2315 * Add CAdES-BES signature verification support, mostly derived
2316 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2318 *Filipe Raimundo da Silva*
2320 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2324 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2325 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2329 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2333 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2334 EC_KEY_precompute_mult().
2338 * Deprecated EC_POINTs_mul().
2342 * Removed FIPS_mode() and FIPS_mode_set().
2346 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2350 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2351 EC_POINT_get_Jprojective_coordinates_GFp().
2355 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2356 arrays to be more easily constructed via a series of utility functions.
2357 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2358 the various push functions and finally convert to a passable OSSL_PARAM
2359 array using OSSL_PARAM_BLD_to_param().
2363 * The security strength of SHA1 and MD5 based signatures in TLS has been
2368 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2369 contain a provider side internal key.
2373 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2377 * Project text documents not yet having a proper file name extension
2378 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2379 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2383 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2384 have been converted to Markdown with the goal to produce documents
2385 which not only look pretty when viewed online in the browser, but
2386 remain well readable inside a plain text editor.
2388 To achieve this goal, a 'minimalistic' Markdown style has been applied
2389 which avoids formatting elements that interfere too much with the
2390 reading flow in the text file. For example, it
2392 * avoids [ATX headings][] and uses [setext headings][] instead
2393 (which works for `<h1>` and `<h2>` headings only).
2394 * avoids [inline links][] and uses [reference links][] instead.
2395 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2397 [ATX headings]: https://github.github.com/gfm/#atx-headings
2398 [setext headings]: https://github.github.com/gfm/#setext-headings
2399 [inline links]: https://github.github.com/gfm/#inline-link
2400 [reference links]: https://github.github.com/gfm/#reference-link
2401 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2402 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2404 *Matthias St. Pierre*
2406 * The test suite is changed to preserve results of each test recipe.
2407 A new directory test-runs/ with subdirectories named like the
2408 test recipes are created in the build tree for this purpose.
2412 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2413 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2414 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2416 *David von Oheimb, Martin Peylo*
2418 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2419 It supports arbitrary request and response content types, GET redirection,
2420 TLS, connections via HTTP(S) proxies, connections and exchange via
2421 user-defined BIOs (allowing implicit connections), persistent connections,
2422 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2423 The legacy OCSP-focused (and only partly documented) API
2424 is retained for backward compatibility, while most of it is deprecated.
2428 * Added `util/check-format.pl`, a tool for checking adherence to the
2429 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2430 The checks performed are incomplete and yield some false positives.
2431 Still the tool should be useful for detecting most typical glitches.
2435 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2436 If domain name resolution yields multiple IP addresses all of them are tried
2437 after `connect()` failures.
2441 * All of the low-level RSA functions have been deprecated.
2445 * X509 certificates signed using SHA1 are no longer allowed at security
2450 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2451 modified to use PKEY APIs. These commands are now in maintenance mode
2452 and no new features will be added to them.
2456 * The command line utility rsautl has been deprecated.
2460 * The command line utilities genrsa and rsa have been modified to use PKEY
2461 APIs. They now write PKCS#8 keys by default. These commands are now in
2462 maintenance mode and no new features will be added to them.
2466 * All of the low-level DH functions have been deprecated.
2468 *Paul Dale and Matt Caswell*
2470 * All of the low-level DSA functions have been deprecated.
2474 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2475 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2479 * Deprecated low-level ECDH and ECDSA functions.
2483 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2487 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2488 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2489 a new formulation to include all the things it can be used for,
2490 as well as words of caution.
2494 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2498 * All of the low-level HMAC functions have been deprecated.
2500 *Paul Dale and David von Oheimb*
2502 * Over two thousand fixes were made to the documentation, including:
2503 - Common options (such as -rand/-writerand, TLS version control, etc)
2504 were refactored and point to newly-enhanced descriptions in openssl.pod.
2505 - Added style conformance for all options (with help from Richard Levitte),
2506 documented all reported missing options, added a CI build to check
2507 that all options are documented and that no unimplemented options
2509 - Documented some internals, such as all use of environment variables.
2510 - Addressed all internal broken L<> references.
2514 * All of the low-level CMAC functions have been deprecated.
2518 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2519 functions have been deprecated.
2521 *Paul Dale and David von Oheimb*
2523 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2524 set of functions. The documentation mentioned negative values for some
2525 errors, but this was never the case, so the mention of negative values
2528 Code that followed the documentation and thereby check with something
2529 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2533 * All of the low-level cipher functions have been deprecated.
2535 *Matt Caswell and Paul Dale*
2537 * Removed include/openssl/opensslconf.h.in and replaced it with
2538 include/openssl/configuration.h.in, which differs in not including
2539 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2540 was added to include both.
2542 This allows internal hacks where one might need to modify the set
2543 of configured macros, for example this if deprecated symbols are
2544 still supposed to be available internally:
2546 #include <openssl/configuration.h>
2548 #undef OPENSSL_NO_DEPRECATED
2549 #define OPENSSL_SUPPRESS_DEPRECATED
2551 #include <openssl/macros.h>
2553 This should not be used by applications that use the exported
2554 symbols, as that will lead to linking errors.
2558 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2559 used in exponentiation with 512-bit moduli. No EC algorithms are
2560 affected. Analysis suggests that attacks against 2-prime RSA1024,
2561 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2562 difficult to perform and are not believed likely. Attacks against DH512
2563 are considered just feasible. However, for an attack the target would
2564 have to reuse the DH512 private key, which is not recommended anyway.
2565 Also applications directly using the low-level API BN_mod_exp may be
2566 affected if they use BN_FLG_CONSTTIME.
2571 * Most memory-debug features have been deprecated, and the functionality
2572 replaced with no-ops.
2576 * Added documentation for the STACK API.
2580 * Introduced a new method type and API, OSSL_ENCODER, to represent
2581 generic encoders. These do the same sort of job that PEM writers
2582 and d2i functions do, but with support for methods supplied by
2583 providers, and the possibility for providers to support other
2588 * Introduced a new method type and API, OSSL_DECODER, to represent
2589 generic decoders. These do the same sort of job that PEM readers
2590 and i2d functions do, but with support for methods supplied by
2591 providers, and the possibility for providers to support other
2596 * Added a .pragma directive to the syntax of configuration files, to
2597 allow varying behavior in a supported and predictable manner.
2598 Currently added pragma:
2602 This allows dollar signs to be a keyword character unless it's
2603 followed by a opening brace or parenthesis. This is useful for
2604 platforms where dollar signs are commonly used in names, such as
2605 volume names and system directory names on VMS.
2609 * Added functionality to create an EVP_PKEY from user data.
2613 * Change the interpretation of the '--api' configuration option to
2614 mean that this is a desired API compatibility level with no
2615 further meaning. The previous interpretation, that this would
2616 also mean to remove all deprecated symbols up to and including
2617 the given version, no requires that 'no-deprecated' is also used
2618 in the configuration.
2620 When building applications, the desired API compatibility level
2621 can be set with the OPENSSL_API_COMPAT macro like before. For
2622 API compatibility version below 3.0, the old style numerical
2623 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2624 For version 3.0 and on, the value is expected to be the decimal
2625 value calculated from the major and minor version like this:
2627 MAJOR * 10000 + MINOR * 100
2631 -DOPENSSL_API_COMPAT=30000 For 3.0
2632 -DOPENSSL_API_COMPAT=30200 For 3.2
2634 To hide declarations that are deprecated up to and including the
2635 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2636 given when building the application as well.
2640 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2641 access to certificate and CRL stores via URIs and OSSL_STORE
2644 This adds the following functions:
2646 - X509_LOOKUP_store()
2647 - X509_STORE_load_file()
2648 - X509_STORE_load_path()
2649 - X509_STORE_load_store()
2650 - SSL_add_store_cert_subjects_to_stack()
2651 - SSL_CTX_set_default_verify_store()
2652 - SSL_CTX_load_verify_file()
2653 - SSL_CTX_load_verify_dir()
2654 - SSL_CTX_load_verify_store()
2658 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2659 The presence of this system service is determined at run-time.
2663 * Added functionality to create an EVP_PKEY context based on data
2664 for methods from providers. This takes an algorithm name and a
2665 property query string and simply stores them, with the intent
2666 that any operation that uses this context will use those strings
2667 to fetch the needed methods implicitly, thereby making the port
2668 of application written for pre-3.0 OpenSSL easier.
2672 * The undocumented function NCONF_WIN32() has been deprecated; for
2673 conversion details see the HISTORY section of doc/man5/config.pod
2677 * Introduced the new functions EVP_DigestSignInit_ex() and
2678 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2679 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2680 pages for further details.
2684 * Over two thousand fixes were made to the documentation, including:
2685 adding missing command flags, better style conformance, documentation
2688 *Rich Salz, Richard Levitte*
2690 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2691 X25519, X448, Ed25519 and Ed448.
2695 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2700 * Deprecated the public definition of `ERR_STATE` as well as the function
2701 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2706 * Added ERR functionality to give callers access to the stored function
2707 names that have replaced the older function code based functions.
2709 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2710 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2711 ERR_peek_error_all() and ERR_peek_last_error_all().
2713 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2714 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2715 ERR_func_error_string().
2719 * Extended testing to be verbose for failing tests only. The make variables
2720 VERBOSE_FAILURE or VF can be used to enable this:
2722 $ make VF=1 test # Unix
2723 $ mms /macro=(VF=1) test ! OpenVMS
2724 $ nmake VF=1 test # Windows
2728 * Added the `-copy_extensions` option to the `x509` command for use with
2729 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2730 all extensions in the request are copied to the certificate or vice versa.
2732 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2734 * Added the `-copy_extensions` option to the `req` command for use with
2735 `-x509`. When given with the `copy` or `copyall` argument,
2736 all extensions in the certification request are copied to the certificate.
2740 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2741 they generate are by default RFC 5280 compliant in the following sense:
2742 There is a subjectKeyIdentifier extension with a hash value of the public key
2743 and for not self-signed certs there is an authorityKeyIdentifier extension
2744 with a keyIdentifier field or issuer information identifying the signing key.
2745 This is done unless some configuration overrides the new default behavior,
2746 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2750 * Added several checks to `X509_verify_cert()` according to requirements in
2751 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2752 (which may be done by using the CLI option `-x509_strict`):
2753 * The basicConstraints of CA certificates must be marked critical.
2754 * CA certificates must explicitly include the keyUsage extension.
2755 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2756 * The issuer name of any certificate must not be empty.
2757 * The subject name of CA certs, certs with keyUsage crlSign,
2758 and certs without subjectAlternativeName must not be empty.
2759 * If a subjectAlternativeName extension is given it must not be empty.
2760 * The signatureAlgorithm field and the cert signature must be consistent.
2761 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2762 must not be marked critical.
2763 * The authorityKeyIdentifier must be given for X.509v3 certs
2764 unless they are self-signed.
2765 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2769 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2770 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2774 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2775 used even when parsing explicit parameters, when loading a encoded key
2776 or calling `EC_GROUP_new_from_ecpkparameters()`/
2777 `EC_GROUP_new_from_ecparameters()`.
2778 This prevents bypass of security hardening and performance gains,
2779 especially for curves with specialized EC_METHODs.
2780 By default, if a key encoded with explicit parameters is loaded and later
2781 encoded, the output is still encoded with explicit parameters, even if
2782 internally a "named" EC_GROUP is used for computation.
2786 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2787 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2788 NULL. After this change, only the cofactor parameter can be NULL. It also
2789 does some minimal sanity checks on the passed order.
2794 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2795 An attack is simple, if the first CMS_recipientInfo is valid but the
2796 second CMS_recipientInfo is chosen ciphertext. If the second
2797 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2798 encryption key will be replaced by garbage, and the message cannot be
2799 decoded, but if the RSA decryption fails, the correct encryption key is
2800 used and the recipient will not notice the attack.
2801 As a work around for this potential attack the length of the decrypted
2802 key must be equal to the cipher default key length, in case the
2803 certificate is not given and all recipientInfo are tried out.
2804 The old behaviour can be re-enabled in the CMS code by setting the
2805 CMS_DEBUG_DECRYPT flag.
2809 * Early start up entropy quality from the DEVRANDOM seed source has been
2810 improved for older Linux systems. The RAND subsystem will wait for
2811 /dev/random to be producing output before seeding from /dev/urandom.
2812 The seeded state is stored for future library initialisations using
2813 a system global shared memory segment. The shared memory identifier
2814 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2815 the desired value. The default identifier is 114.
2819 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2820 when primes for RSA keys are computed.
2821 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2822 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2823 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2824 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2825 This avoids possible fingerprinting of newly generated RSA modules.
2829 * Correct the extended master secret constant on EBCDIC systems. Without this
2830 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2831 negotiate EMS will fail. Unfortunately this also means that TLS connections
2832 between EBCDIC systems with this fix, and EBCDIC systems without this
2833 fix will fail if they negotiate EMS.
2837 * Changed the library initialisation so that the config file is now loaded
2838 by default. This was already the case for libssl. It now occurs for both
2839 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2840 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2844 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2845 where the former acts as a replacement for `ERR_put_error()`, and the
2846 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2847 `ERR_raise_data()` adds more flexibility by taking a format string and
2848 an arbitrary number of arguments following it, to be processed with
2853 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2854 to check if a named provider is loaded and available. When called, it
2855 will also activate all fallback providers if such are still present.
2859 * Enforce a minimum DH modulus size of 512 bits.
2863 * Changed DH parameters to generate the order q subgroup instead of 2q.
2864 Previously generated DH parameters are still accepted by DH_check
2865 but DH_generate_key works around that by clearing bit 0 of the
2866 private key for those. This avoids leaking bit 0 of the private key.
2870 * Significantly reduce secure memory usage by the randomness pools.
2874 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2879 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2880 algorithms. An implementation of a key exchange algorithm can be obtained
2881 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2882 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2883 the older EVP_PKEY_derive_init() function. See the man pages for the new
2884 functions for further details.
2888 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2892 * Removed the function names from error messages and deprecated the
2897 * Removed NextStep support and the macro OPENSSL_UNISTD
2901 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2902 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2903 Also removed "export var as function" capability; we do not export
2904 variables, only functions.
2908 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2909 an error and 1 indicating success. In previous versions of OpenSSL this
2910 was a void type. If a key was set longer than the maximum possible this
2915 * Support SM2 signing and verification schemes with X509 certificate.
2919 * Use SHA256 as the default digest for TS query in the `ts` app.
2923 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2927 * Default cipher lists/suites are now available via a function, the
2928 #defines are deprecated.
2932 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2933 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2934 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2938 * Join the directories crypto/x509 and crypto/x509v3
2942 * Added command 'openssl kdf' that uses the EVP_KDF API.
2946 * Added command 'openssl mac' that uses the EVP_MAC API.
2950 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2951 as default directories. Also added the command 'openssl info'
2952 for scripting purposes.
2956 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2961 * Add prediction resistance to the DRBG reseeding process.
2965 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2966 mandated by IEEE Std 1619-2018.
2970 * Added newline escaping functionality to a filename when using openssl dgst.
2971 This output format is to replicate the output format found in the `*sum`
2972 checksum programs. This aims to preserve backward compatibility.
2974 *Matt Eaton, Richard Levitte, and Paul Dale*
2976 * Removed the heartbeat message in DTLS feature, as it has very
2977 little usage and doesn't seem to fulfill a valuable purpose.
2978 The configuration option is now deprecated.
2982 * Changed the output of 'openssl {digestname} < file' to display the
2983 digest name in its output.
2987 * Added a new generic trace API which provides support for enabling
2988 instrumentation through trace output.
2990 *Richard Levitte & Matthias St. Pierre*
2992 * Added build tests for C++. These are generated files that only do one
2993 thing, to include one public OpenSSL head file each. This tests that
2994 the public header files can be usefully included in a C++ application.
2996 This test isn't enabled by default. It can be enabled with the option
2997 'enable-buildtest-c++'.
3001 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
3005 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
3009 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
3013 * Added KMAC to EVP_MAC.
3017 * Added property based algorithm implementation selection framework to
3022 * Added SCA hardening for modular field inversion in EC_GROUP through
3023 a new dedicated field_inv() pointer in EC_METHOD.
3024 This also addresses a leakage affecting conversions from projective
3025 to affine coordinates.
3027 *Billy Bob Brumley, Nicola Tuveri*
3029 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
3030 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
3031 those algorithms that were already supported through the EVP_PKEY API
3032 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
3033 and scrypt are now wrappers that call EVP_KDF.
3037 * Build devcrypto engine as a dynamic engine.
3039 *Eneas U de Queiroz*
3041 * Add keyed BLAKE2 to EVP_MAC.
3045 * Fix a bug in the computation of the endpoint-pair shared secret used
3046 by DTLS over SCTP. This breaks interoperability with older versions
3047 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
3048 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
3049 interoperability with such broken implementations. However, enabling
3050 this switch breaks interoperability with correct implementations.
3052 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3053 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3057 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3061 * Changed the license to the Apache License v2.0.
3065 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
3067 - Major releases (indicated by incrementing the MAJOR release number)
3068 may introduce incompatible API/ABI changes.
3069 - Minor releases (indicated by incrementing the MINOR release number)
3070 may introduce new features but retain API/ABI compatibility.
3071 - Patch releases (indicated by incrementing the PATCH number)
3072 are intended for bug fixes and other improvements of existing
3073 features only (like improving performance or adding documentation)
3074 and retain API/ABI compatibility.
3078 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
3082 * Remove the 'dist' target and add a tarball building script. The
3083 'dist' target has fallen out of use, and it shouldn't be
3084 necessary to configure just to create a source distribution.
3088 * Recreate the OS390-Unix config target. It no longer relies on a
3089 special script like it did for OpenSSL pre-1.1.0.
3093 * Instead of having the source directories listed in Configure, add
3094 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
3099 * Add GMAC to EVP_MAC.
3103 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
3107 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
3108 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
3109 to facilitate the continued use of MACs through raw private keys in
3110 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
3114 * Deprecate ECDH_KDF_X9_62().
3118 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
3119 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
3120 are retained for backwards compatibility.
3124 * AES-XTS mode now enforces that its two keys are different to mitigate
3125 the attacked described in "Efficient Instantiations of Tweakable
3126 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
3127 Details of this attack can be obtained from:
3128 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
3132 * Rename the object files, i.e. give them other names than in previous
3133 versions. Their names now include the name of the final product, as
3134 well as its type mnemonic (bin, lib, shlib).
3138 * Added new option for 'openssl list', '-objects', which will display the
3139 list of built in objects, i.e. OIDs with names.
3143 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
3144 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
3149 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
3150 improves application performance by removing data copies and providing
3151 applications with zero-copy system calls such as sendfile and splice.
3155 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
3159 * `PKCS12_parse` now maintains the order of the parsed certificates
3160 when outputting them via `*ca` (rather than reversing it).
3164 * Deprecated pthread fork support methods.
3168 * Added support for FFDHE key exchange in TLS 1.3.
3172 * Added a new concept for OpenSSL plugability: providers. This
3173 functionality is designed to replace the ENGINE API and ENGINE
3174 implementations, and to be much more dynamic, allowing provider
3175 authors to introduce new algorithms among other things, as long as
3176 there's an API that supports the algorithm type.
3178 With this concept comes a new core API for interaction between
3179 libcrypto and provider implementations. Public libcrypto functions
3180 that want to use providers do so through this core API.
3182 The main documentation for this core API is found in
3183 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
3184 refer to other manuals describing the API specific for supported
3185 algorithm types (also called operations).
3192 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
3194 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
3196 * Avoid loading of a dynamic engine twice.
3200 * Prioritise DANE TLSA issuer certs over peer certs
3204 * Fixed random API for MacOS prior to 10.12
3206 These MacOS versions don't support the CommonCrypto APIs
3210 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
3212 * Fixed an SM2 Decryption Buffer Overflow.
3214 In order to decrypt SM2 encrypted data an application is expected to
3215 call the API function EVP_PKEY_decrypt(). Typically an application will
3216 call this function twice. The first time, on entry, the "out" parameter
3217 can be NULL and, on exit, the "outlen" parameter is populated with the
3218 buffer size required to hold the decrypted plaintext. The application
3219 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
3220 again, but this time passing a non-NULL value for the "out" parameter.
3222 A bug in the implementation of the SM2 decryption code means that the
3223 calculation of the buffer size required to hold the plaintext returned
3224 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
3225 size required by the second call. This can lead to a buffer overflow
3226 when EVP_PKEY_decrypt() is called by the application a second time with
3227 a buffer that is too small.
3229 A malicious attacker who is able present SM2 content for decryption to
3230 an application could cause attacker chosen data to overflow the buffer
3231 by up to a maximum of 62 bytes altering the contents of other data held
3232 after the buffer, possibly changing application behaviour or causing
3233 the application to crash. The location of the buffer is application
3234 dependent but is typically heap allocated.
3239 * Fixed various read buffer overruns processing ASN.1 strings
3241 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
3242 structure which contains a buffer holding the string data and a field
3243 holding the buffer length. This contrasts with normal C strings which
3244 are represented as a buffer for the string data which is terminated
3245 with a NUL (0) byte.
3247 Although not a strict requirement, ASN.1 strings that are parsed using
3248 OpenSSL's own "d2i" functions (and other similar parsing functions) as
3249 well as any string whose value has been set with the ASN1_STRING_set()
3250 function will additionally NUL terminate the byte array in the
3251 ASN1_STRING structure.
3253 However, it is possible for applications to directly construct valid
3254 ASN1_STRING structures which do not NUL terminate the byte array by
3255 directly setting the "data" and "length" fields in the ASN1_STRING
3256 array. This can also happen by using the ASN1_STRING_set0() function.
3258 Numerous OpenSSL functions that print ASN.1 data have been found to
3259 assume that the ASN1_STRING byte array will be NUL terminated, even
3260 though this is not guaranteed for strings that have been directly
3261 constructed. Where an application requests an ASN.1 structure to be
3262 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3263 been directly constructed by the application without NUL terminating
3264 the "data" field, then a read buffer overrun can occur.
3266 The same thing can also occur during name constraints processing
3267 of certificates (for example if a certificate has been directly
3268 constructed by the application instead of loading it via the OpenSSL
3269 parsing functions, and the certificate contains non NUL terminated
3270 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3271 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3273 If a malicious actor can cause an application to directly construct an
3274 ASN1_STRING and then process it through one of the affected OpenSSL
3275 functions then this issue could be hit. This might result in a crash
3276 (causing a Denial of Service attack). It could also result in the
3277 disclosure of private memory contents (such as private keys, or
3278 sensitive plaintext).
3283 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3285 * Fixed a problem with verifying a certificate chain when using the
3286 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3287 the certificates present in a certificate chain. It is not set by default.
3289 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3290 the chain that have explicitly encoded elliptic curve parameters was added
3291 as an additional strict check.
3293 An error in the implementation of this check meant that the result of a
3294 previous check to confirm that certificates in the chain are valid CA
3295 certificates was overwritten. This effectively bypasses the check
3296 that non-CA certificates must not be able to issue other certificates.
3298 If a "purpose" has been configured then there is a subsequent opportunity
3299 for checks that the certificate is a valid CA. All of the named "purpose"
3300 values implemented in libcrypto perform this check. Therefore, where
3301 a purpose is set the certificate chain will still be rejected even when the
3302 strict flag has been used. A purpose is set by default in libssl client and
3303 server certificate verification routines, but it can be overridden or
3304 removed by an application.
3306 In order to be affected, an application must explicitly set the
3307 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3308 for the certificate verification or, in the case of TLS client or server
3309 applications, override the default purpose.
3314 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3315 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3316 renegotiation ClientHello omits the signature_algorithms extension (where it
3317 was present in the initial ClientHello), but includes a
3318 signature_algorithms_cert extension then a NULL pointer dereference will
3319 result, leading to a crash and a denial of service attack.
3321 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3322 (which is the default configuration). OpenSSL TLS clients are not impacted by
3326 *Peter Kästle and Samuel Sapalski*
3328 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3330 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3331 create a unique hash value based on the issuer and serial number data
3332 contained within an X509 certificate. However, it was failing to correctly
3333 handle any errors that may occur while parsing the issuer field (which might
3334 occur if the issuer field is maliciously constructed). This may subsequently
3335 result in a NULL pointer deref and a crash leading to a potential denial of
3341 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3342 padding mode to correctly check for rollback attacks. This is considered a
3343 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3348 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3349 functions. Previously they could overflow the output length argument in some
3350 cases where the input length is close to the maximum permissible length for
3351 an integer on the platform. In such cases the return value from the function
3352 call would be 1 (indicating success), but the output length value would be
3353 negative. This could cause applications to behave incorrectly or crash.
3358 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3359 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3360 could be exploited in a side channel attack to recover the password. Since
3361 the attack is local host only this is outside of the current OpenSSL
3362 threat model and therefore no CVE is assigned.
3364 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3369 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3371 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3372 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3373 If an attacker can control both items being compared then this could lead
3374 to a possible denial of service attack. OpenSSL itself uses the
3375 GENERAL_NAME_cmp function for two purposes:
3376 1) Comparing CRL distribution point names between an available CRL and a
3377 CRL distribution point embedded in an X509 certificate
3378 2) When verifying that a timestamp response token signer matches the
3379 timestamp authority name (exposed via the API functions
3380 TS_RESP_verify_response and TS_RESP_verify_token)
3385 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3387 * Certificates with explicit curve parameters are now disallowed in
3388 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3392 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3393 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3394 conversely, silently ignore DTLS protocol version bounds when configuring
3395 TLS-based contexts. The commands can be repeated to set bounds of both
3396 types. The same applies with the corresponding "min_protocol" and
3397 "max_protocol" command-line switches, in case some application uses both TLS
3400 SSL_CTX instances that are created for a fixed protocol version (e.g.
3401 TLSv1_server_method()) also silently ignore version bounds. Previously
3402 attempts to apply bounds to these protocol versions would result in an
3403 error. Now only the "version-flexible" SSL_CTX instances are subject to
3404 limits in configuration files in command-line options.
3408 * Handshake now fails if Extended Master Secret extension is dropped
3413 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3415 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3417 * Fixed segmentation fault in SSL_check_chain()
3418 Server or client applications that call the SSL_check_chain() function
3419 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3420 dereference as a result of incorrect handling of the
3421 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3422 or unrecognised signature algorithm is received from the peer. This could
3423 be exploited by a malicious peer in a Denial of Service attack.
3428 * Added AES consttime code for no-asm configurations
3429 an optional constant time support for AES was added
3430 when building openssl for no-asm.
3431 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3432 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3433 At this time this feature is by default disabled.
3434 It will be enabled by default in 3.0.
3438 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3440 * Revert the change of EOF detection while reading in libssl to avoid
3441 regressions in applications depending on the current way of reporting
3442 the EOF. As the existing method is not fully accurate the change to
3443 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3444 branch and will be present in the 3.0 release.
3448 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3449 when primes for RSA keys are computed.
3450 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3451 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3452 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3453 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3454 This avoids possible fingerprinting of newly generated RSA modules.
3458 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3460 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3461 while reading in libssl then we would report an error back to the
3462 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3463 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3464 therefore give a hint as to what went wrong.
3468 * Check that ed25519 and ed448 are allowed by the security level. Previously
3469 signature algorithms not using an MD were not being checked that they were
3470 allowed by the security level.
3474 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3475 was not quite right. The behaviour was not consistent between resumption
3476 and normal handshakes, and also not quite consistent with historical
3477 behaviour. The behaviour in various scenarios has been clarified and
3478 it has been updated to make it match historical behaviour as closely as
3483 * *[VMS only]* The header files that the VMS compilers include automatically,
3484 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3485 that the C++ compiler doesn't understand. This is a shortcoming in the
3486 compiler, but can be worked around with `__cplusplus` guards.
3488 C++ applications that use OpenSSL libraries must be compiled using the
3489 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3490 functions. Otherwise, only functions with symbols of less than 31
3491 characters can be used, as the linker will not be able to successfully
3492 resolve symbols with longer names.
3496 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3497 The presence of this system service is determined at run-time.
3501 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3506 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3508 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3509 number generator (RNG). This was intended to include protection in the
3510 event of a fork() system call in order to ensure that the parent and child
3511 processes did not share the same RNG state. However, this protection was not
3512 being used in the default case.
3514 A partial mitigation for this issue is that the output from a high
3515 precision timer is mixed into the RNG state so the likelihood of a parent
3516 and child process sharing state is significantly reduced.
3518 If an application already calls OPENSSL_init_crypto() explicitly using
3519 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3522 *Matthias St. Pierre*
3524 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3525 used even when parsing explicit parameters, when loading a encoded key
3526 or calling `EC_GROUP_new_from_ecpkparameters()`/
3527 `EC_GROUP_new_from_ecparameters()`.
3528 This prevents bypass of security hardening and performance gains,
3529 especially for curves with specialized EC_METHODs.
3530 By default, if a key encoded with explicit parameters is loaded and later
3531 encoded, the output is still encoded with explicit parameters, even if
3532 internally a "named" EC_GROUP is used for computation.
3536 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3537 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3538 NULL. After this change, only the cofactor parameter can be NULL. It also
3539 does some minimal sanity checks on the passed order.
3544 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3545 An attack is simple, if the first CMS_recipientInfo is valid but the
3546 second CMS_recipientInfo is chosen ciphertext. If the second
3547 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3548 encryption key will be replaced by garbage, and the message cannot be
3549 decoded, but if the RSA decryption fails, the correct encryption key is
3550 used and the recipient will not notice the attack.
3551 As a work around for this potential attack the length of the decrypted
3552 key must be equal to the cipher default key length, in case the
3553 certificate is not given and all recipientInfo are tried out.
3554 The old behaviour can be re-enabled in the CMS code by setting the
3555 CMS_DEBUG_DECRYPT flag.
3560 * Early start up entropy quality from the DEVRANDOM seed source has been
3561 improved for older Linux systems. The RAND subsystem will wait for
3562 /dev/random to be producing output before seeding from /dev/urandom.
3563 The seeded state is stored for future library initialisations using
3564 a system global shared memory segment. The shared memory identifier
3565 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3566 the desired value. The default identifier is 114.
3570 * Correct the extended master secret constant on EBCDIC systems. Without this
3571 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3572 negotiate EMS will fail. Unfortunately this also means that TLS connections
3573 between EBCDIC systems with this fix, and EBCDIC systems without this
3574 fix will fail if they negotiate EMS.
3578 * Use Windows installation paths in the mingw builds
3580 Mingw isn't a POSIX environment per se, which means that Windows
3581 paths should be used for installation.
3586 * Changed DH_check to accept parameters with order q and 2q subgroups.
3587 With order 2q subgroups the bit 0 of the private key is not secret
3588 but DH_generate_key works around that by clearing bit 0 of the
3589 private key for those. This avoids leaking bit 0 of the private key.
3593 * Significantly reduce secure memory usage by the randomness pools.
3597 * Revert the DEVRANDOM_WAIT feature for Linux systems
3599 The DEVRANDOM_WAIT feature added a select() call to wait for the
3600 /dev/random device to become readable before reading from the
3601 /dev/urandom device.
3603 It turned out that this change had negative side effects on
3604 performance which were not acceptable. After some discussion it
3605 was decided to revert this feature and leave it up to the OS
3606 resp. the platform maintainer to ensure a proper initialization
3607 during early boot time.
3609 *Matthias St. Pierre*
3611 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3613 * Add build tests for C++. These are generated files that only do one
3614 thing, to include one public OpenSSL head file each. This tests that
3615 the public header files can be usefully included in a C++ application.
3617 This test isn't enabled by default. It can be enabled with the option
3618 'enable-buildtest-c++'.
3622 * Enable SHA3 pre-hashing for ECDSA and DSA.
3626 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3627 This changes the size when using the `genpkey` command when no size is given.
3628 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3629 generation commands to use 2048 bits by default.
3633 * Reorganize the manual pages to consistently have RETURN VALUES,
3634 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3635 util/fix-doc-nits accordingly.
3637 *Paul Yang, Joshua Lock*
3639 * Add the missing accessor EVP_PKEY_get0_engine()
3643 * Have commands like `s_client` and `s_server` output the signature scheme
3644 along with other cipher suite parameters when debugging.
3648 * Make OPENSSL_config() error agnostic again.
3652 * Do the error handling in RSA decryption constant time.
3656 * Prevent over long nonces in ChaCha20-Poly1305.
3658 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3659 for every encryption operation. RFC 7539 specifies that the nonce value
3660 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3661 and front pads the nonce with 0 bytes if it is less than 12
3662 bytes. However it also incorrectly allows a nonce to be set of up to 16
3663 bytes. In this case only the last 12 bytes are significant and any
3664 additional leading bytes are ignored.
3666 It is a requirement of using this cipher that nonce values are
3667 unique. Messages encrypted using a reused nonce value are susceptible to
3668 serious confidentiality and integrity attacks. If an application changes
3669 the default nonce length to be longer than 12 bytes and then makes a
3670 change to the leading bytes of the nonce expecting the new value to be a
3671 new unique nonce then such an application could inadvertently encrypt
3672 messages with a reused nonce.
3674 Additionally the ignored bytes in a long nonce are not covered by the
3675 integrity guarantee of this cipher. Any application that relies on the
3676 integrity of these ignored leading bytes of a long nonce may be further
3677 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3678 is safe because no such use sets such a long nonce value. However user
3679 applications that use this cipher directly and set a non-default nonce
3680 length to be longer than 12 bytes may be vulnerable.
3682 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3688 * Add DEVRANDOM_WAIT feature for Linux systems
3690 On older Linux systems where the getrandom() system call is not available,
3691 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3692 Contrary to getrandom(), the /dev/urandom device will not block during
3693 early boot when the kernel CSPRNG has not been seeded yet.
3695 To mitigate this known weakness, use select() to wait for /dev/random to
3696 become readable before reading from /dev/urandom.
3698 * Ensure that SM2 only uses SM3 as digest algorithm
3702 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3704 * Change the info callback signals for the start and end of a post-handshake
3705 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3706 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3707 confused by this and assume that a TLSv1.2 renegotiation has started. This
3708 can break KeyUpdate handling. Instead we no longer signal the start and end
3709 of a post handshake message exchange (although the messages themselves are
3710 still signalled). This could break some applications that were expecting
3711 the old signals. However without this KeyUpdate is not usable for many
3716 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3718 * Timing vulnerability in DSA signature generation
3720 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3721 timing side channel attack. An attacker could use variations in the signing
3722 algorithm to recover the private key.
3724 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3729 * Timing vulnerability in ECDSA signature generation
3731 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3732 timing side channel attack. An attacker could use variations in the signing
3733 algorithm to recover the private key.
3735 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3740 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3741 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3742 of two gigabytes and the error handling improved.
3744 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3745 categorized as a normal bug, not a security issue, because the DRBG reseeds
3746 automatically and is fully functional even without additional randomness
3747 provided by the application.
3749 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3751 * Add a new ClientHello callback. Provides a callback interface that gives
3752 the application the ability to adjust the nascent SSL object at the
3753 earliest stage of ClientHello processing, immediately after extensions have
3754 been collected but before they have been processed. In particular, this
3755 callback can adjust the supported TLS versions in response to the contents
3760 * Add SM2 base algorithm support.
3764 * s390x assembly pack: add (improved) hardware-support for the following
3765 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3766 aes-cfb/cfb8, aes-ecb.
3770 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3771 parameter is no longer accepted, as it leads to a corrupt table. NULL
3772 pem_str is reserved for alias entries only.
3776 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3777 step for prime curves. The new implementation is based on formulae from
3778 differential addition-and-doubling in homogeneous projective coordinates
3779 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3780 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3781 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3782 to work in projective coordinates.
3784 *Billy Bob Brumley, Nicola Tuveri*
3786 * Change generating and checking of primes so that the error rate of not
3787 being prime depends on the intended use based on the size of the input.
3788 For larger primes this will result in more rounds of Miller-Rabin.
3789 The maximal error rate for primes with more than 1080 bits is lowered
3792 *Kurt Roeckx, Annie Yousar*
3794 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3798 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3799 moving between systems, and to avoid confusion when a Windows build is
3800 done with mingw vs with MSVC. For POSIX installs, there's still a
3801 symlink or copy named 'tsget' to avoid that confusion as well.
3805 * Revert blinding in ECDSA sign and instead make problematic addition
3806 length-invariant. Switch even to fixed-length Montgomery multiplication.
3810 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3811 step for binary curves. The new implementation is based on formulae from
3812 differential addition-and-doubling in mixed Lopez-Dahab projective
3813 coordinates, modified to independently blind the operands.
3815 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3817 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3818 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3819 EC_METHODs to implement their own specialized "ladder step", to take
3820 advantage of more favorable coordinate systems or more efficient
3821 differential addition-and-doubling algorithms.
3823 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3825 * Modified the random device based seed sources to keep the relevant
3826 file descriptors open rather than reopening them on each access.
3827 This allows such sources to operate in a chroot() jail without
3828 the associated device nodes being available. This behaviour can be
3829 controlled using RAND_keep_random_devices_open().
3833 * Numerous side-channel attack mitigations have been applied. This may have
3834 performance impacts for some algorithms for the benefit of improved
3835 security. Specific changes are noted in this change log by their respective
3840 * AIX shared library support overhaul. Switch to AIX "natural" way of
3841 handling shared libraries, which means collecting shared objects of
3842 different versions and bitnesses in one common archive. This allows to
3843 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3844 doesn't affect the way 3rd party applications are linked, only how
3845 multi-version installation is managed.
3849 * Make ec_group_do_inverse_ord() more robust and available to other
3850 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3851 mitigations are applied to the fallback BN_mod_inverse().
3852 When using this function rather than BN_mod_inverse() directly, new
3853 EC cryptosystem implementations are then safer-by-default.
3857 * Add coordinate blinding for EC_POINT and implement projective
3858 coordinate blinding for generic prime curves as a countermeasure to
3859 chosen point SCA attacks.
3861 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3863 * Add blinding to ECDSA and DSA signatures to protect against side channel
3864 attacks discovered by Keegan Ryan (NCC Group).
3868 * Enforce checking in the `pkeyutl` command to ensure that the input
3869 length does not exceed the maximum supported digest length when performing
3870 a sign, verify or verifyrecover operation.
3874 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3875 I/O in combination with something like select() or poll() will hang. This
3876 can be turned off again using SSL_CTX_clear_mode().
3877 Many applications do not properly handle non-application data records, and
3878 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3879 around the problems in those applications, but can also break some.
3880 It's recommended to read the manpages about SSL_read(), SSL_write(),
3881 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3882 SSL_CTX_set_read_ahead() again.
3886 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3887 now allow empty (zero character) pass phrases.
3891 * Apply blinding to binary field modular inversion and remove patent
3892 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3896 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3897 binary and prime elliptic curves.
3901 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3902 constant time fixed point multiplication.
3906 * Revise elliptic curve scalar multiplication with timing attack
3907 defenses: ec_wNAF_mul redirects to a constant time implementation
3908 when computing fixed point and variable point multiplication (which
3909 in OpenSSL are mostly used with secret scalars in keygen, sign,
3910 ECDH derive operations).
3911 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3914 * Updated CONTRIBUTING
3918 * Updated DRBG / RAND to request nonce and additional low entropy
3919 randomness from the system.
3921 *Matthias St. Pierre*
3923 * Updated 'openssl rehash' to use OpenSSL consistent default.
3927 * Moved the load of the ssl_conf module to libcrypto, which helps
3928 loading engines that libssl uses before libssl is initialised.
3932 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3936 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3938 *Ingo Schwarze, Rich Salz*
3940 * Added output of accepting IP address and port for 'openssl s_server'
3944 * Added a new API for TLSv1.3 ciphersuites:
3945 SSL_CTX_set_ciphersuites()
3946 SSL_set_ciphersuites()
3950 * Memory allocation failures consistently add an error to the error
3955 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3956 in libcrypto when run as setuid/setgid.
3960 * Load any config file by default when libssl is used.
3964 * Added new public header file <openssl/rand_drbg.h> and documentation
3965 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3967 *Matthias St. Pierre*
3969 * QNX support removed (cannot find contributors to get their approval
3970 for the license change).
3974 * TLSv1.3 replay protection for early data has been implemented. See the
3975 SSL_read_early_data() man page for further details.
3979 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3980 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3981 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3982 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3983 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3984 configuration has been separated out. See the ciphers man page or the
3985 SSL_CTX_set_ciphersuites() man page for more information.
3989 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3990 in responder mode now supports the new "-multi" option, which
3991 spawns the specified number of child processes to handle OCSP
3992 requests. The "-timeout" option now also limits the OCSP
3993 responder's patience to wait to receive the full client request
3994 on a newly accepted connection. Child processes are respawned
3995 as needed, and the CA index file is automatically reloaded
3996 when changed. This makes it possible to run the "ocsp" responder
3997 as a long-running service, making the OpenSSL CA somewhat more
3998 feature-complete. In this mode, most diagnostic messages logged
3999 after entering the event loop are logged via syslog(3) rather than
4004 * Added support for X448 and Ed448. Heavily based on original work by
4009 * Extend OSSL_STORE with capabilities to search and to narrow the set of
4010 objects loaded. This adds the functions OSSL_STORE_expect() and
4011 OSSL_STORE_find() as well as needed tools to construct searches and
4012 get the search data out of them.
4016 * Support for TLSv1.3 added. Note that users upgrading from an earlier
4017 version of OpenSSL should review their configuration settings to ensure
4018 that they are still appropriate for TLSv1.3. For further information see:
4019 <https://wiki.openssl.org/index.php/TLS1.3>
4023 * Grand redesign of the OpenSSL random generator
4025 The default RAND method now utilizes an AES-CTR DRBG according to
4026 NIST standard SP 800-90Ar1. The new random generator is essentially
4027 a port of the default random generator from the OpenSSL FIPS 2.0
4028 object module. It is a hybrid deterministic random bit generator
4029 using an AES-CTR bit stream and which seeds and reseeds itself
4030 automatically using trusted system entropy sources.
4032 Some of its new features are:
4033 - Support for multiple DRBG instances with seed chaining.
4034 - The default RAND method makes use of a DRBG.
4035 - There is a public and private DRBG instance.
4036 - The DRBG instances are fork-safe.
4037 - Keep all global DRBG instances on the secure heap if it is enabled.
4038 - The public and private DRBG instance are per thread for lock free
4041 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
4043 * Changed Configure so it only says what it does and doesn't dump
4044 so much data. Instead, ./configdata.pm should be used as a script
4045 to display all sorts of configuration data.
4049 * Added processing of "make variables" to Configure.
4053 * Added SHA512/224 and SHA512/256 algorithm support.
4057 * The last traces of Netware support, first removed in 1.1.0, have
4062 * Get rid of Makefile.shared, and in the process, make the processing
4063 of certain files (rc.obj, or the .def/.map/.opt files produced from
4064 the ordinal files) more visible and hopefully easier to trace and
4065 debug (or make silent).
4069 * Make it possible to have environment variable assignments as
4070 arguments to config / Configure.
4074 * Add multi-prime RSA (RFC 8017) support.
4078 * Add SM3 implemented according to GB/T 32905-2016
4079 *Jack Lloyd <jack.lloyd@ribose.com>,*
4080 *Ronald Tse <ronald.tse@ribose.com>,*
4081 *Erick Borsboom <erick.borsboom@ribose.com>*
4083 * Add 'Maximum Fragment Length' TLS extension negotiation and support
4084 as documented in RFC6066.
4085 Based on a patch from Tomasz Moń
4087 *Filipe Raimundo da Silva*
4089 * Add SM4 implemented according to GB/T 32907-2016.
4090 *Jack Lloyd <jack.lloyd@ribose.com>,*
4091 *Ronald Tse <ronald.tse@ribose.com>,*
4092 *Erick Borsboom <erick.borsboom@ribose.com>*
4094 * Reimplement -newreq-nodes and ERR_error_string_n; the
4095 original author does not agree with the license change.
4099 * Add ARIA AEAD TLS support.
4103 * Some macro definitions to support VS6 have been removed. Visual
4104 Studio 6 has not worked since 1.1.0
4108 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
4109 without clearing the errors.
4113 * Add "atfork" functions. If building on a system that without
4114 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
4115 requirements. The RAND facility now uses/requires this.
4123 * The UI API becomes a permanent and integral part of libcrypto, i.e.
4124 not possible to disable entirely. However, it's still possible to
4125 disable the console reading UI method, UI_OpenSSL() (use UI_null()
4128 To disable, configure with 'no-ui-console'. 'no-ui' is still
4129 possible to use as an alias. Check at compile time with the
4130 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
4131 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
4135 * Add a STORE module, which implements a uniform and URI based reader of
4136 stores that can contain keys, certificates, CRLs and numerous other
4137 objects. The main API is loosely based on a few stdio functions,
4138 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
4139 OSSL_STORE_error and OSSL_STORE_close.
4140 The implementation uses backends called "loaders" to implement arbitrary
4141 URI schemes. There is one built in "loader" for the 'file' scheme.
4145 * Add devcrypto engine. This has been implemented against cryptodev-linux,
4146 then adjusted to work on FreeBSD 8.4 as well.
4147 Enable by configuring with 'enable-devcryptoeng'. This is done by default
4148 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
4152 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
4153 util/mkerr.pl, which is adapted to allow those prefixes, leading to
4154 error code calls like this:
4156 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
4158 With this change, we claim the namespaces OSSL and OPENSSL in a manner
4159 that can be encoded in C. For the foreseeable future, this will only
4162 *Richard Levitte and Tim Hudson*
4164 * Removed BSD cryptodev engine.
4168 * Add a build target 'build_all_generated', to build all generated files
4169 and only that. This can be used to prepare everything that requires
4170 things like perl for a system that lacks perl and then move everything
4171 to that system and do the rest of the build there.
4175 * In the UI interface, make it possible to duplicate the user data. This
4176 can be used by engines that need to retain the data for a longer time
4177 than just the call where this user data is passed.
4181 * Ignore the '-named_curve auto' value for compatibility of applications
4184 *Tomáš Mráz <tmraz@fedoraproject.org>*
4186 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
4187 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
4188 alerts across multiple records (some of which could be empty). In practice
4189 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
4190 prohibits this altogether and other libraries (BoringSSL, NSS) do not
4191 support this at all. Supporting it adds significant complexity to the
4192 record layer, and its removal is unlikely to cause interoperability
4197 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
4198 with Z. These are meant to replace LONG and ZLONG and to be size safe.
4199 The use of LONG and ZLONG is discouraged and scheduled for deprecation
4204 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
4205 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
4207 *Richard Levitte, Andy Polyakov*
4209 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
4214 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4215 platform rather than 'mingw'.
4219 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
4220 success if they are asked to add an object which already exists
4221 in the store. This change cascades to other functions which load
4222 certificates and CRLs.
4226 * x86_64 assembly pack: annotate code with DWARF CFI directives to
4227 facilitate stack unwinding even from assembly subroutines.
4231 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
4232 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
4236 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4237 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4238 which is the minimum version we support.
4242 * Certificate time validation (X509_cmp_time) enforces stricter
4243 compliance with RFC 5280. Fractional seconds and timezone offsets
4244 are no longer allowed.
4248 * Add support for ARIA
4252 * s_client will now send the Server Name Indication (SNI) extension by
4253 default unless the new "-noservername" option is used. The server name is
4254 based on the host provided to the "-connect" option unless overridden by
4255 using "-servername".
4259 * Add support for SipHash
4263 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4264 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4265 prevent issues where no progress is being made and the peer continually
4266 sends unrecognised record types, using up resources processing them.
4270 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4271 using the algorithm defined in
4272 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4276 * Heartbeat support has been removed; the ABI is changed for now.
4278 *Richard Levitte, Rich Salz*
4280 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4284 * The RSA "null" method, which was partially supported to avoid patent
4285 issues, has been replaced to always returns NULL.
4292 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4294 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4295 used even when parsing explicit parameters, when loading a encoded key
4296 or calling `EC_GROUP_new_from_ecpkparameters()`/
4297 `EC_GROUP_new_from_ecparameters()`.
4298 This prevents bypass of security hardening and performance gains,
4299 especially for curves with specialized EC_METHODs.
4300 By default, if a key encoded with explicit parameters is loaded and later
4301 encoded, the output is still encoded with explicit parameters, even if
4302 internally a "named" EC_GROUP is used for computation.
4306 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4307 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4308 NULL. After this change, only the cofactor parameter can be NULL. It also
4309 does some minimal sanity checks on the passed order.
4314 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4315 An attack is simple, if the first CMS_recipientInfo is valid but the
4316 second CMS_recipientInfo is chosen ciphertext. If the second
4317 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4318 encryption key will be replaced by garbage, and the message cannot be
4319 decoded, but if the RSA decryption fails, the correct encryption key is
4320 used and the recipient will not notice the attack.
4321 As a work around for this potential attack the length of the decrypted
4322 key must be equal to the cipher default key length, in case the
4323 certificate is not given and all recipientInfo are tried out.
4324 The old behaviour can be re-enabled in the CMS code by setting the
4325 CMS_DEBUG_DECRYPT flag.
4330 * Use Windows installation paths in the mingw builds
4332 Mingw isn't a POSIX environment per se, which means that Windows
4333 paths should be used for installation.
4338 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4340 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4341 This changes the size when using the `genpkey` command when no size is given.
4342 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4343 generation commands to use 2048 bits by default.
4347 * Prevent over long nonces in ChaCha20-Poly1305.
4349 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4350 for every encryption operation. RFC 7539 specifies that the nonce value
4351 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4352 and front pads the nonce with 0 bytes if it is less than 12
4353 bytes. However it also incorrectly allows a nonce to be set of up to 16
4354 bytes. In this case only the last 12 bytes are significant and any
4355 additional leading bytes are ignored.
4357 It is a requirement of using this cipher that nonce values are
4358 unique. Messages encrypted using a reused nonce value are susceptible to
4359 serious confidentiality and integrity attacks. If an application changes
4360 the default nonce length to be longer than 12 bytes and then makes a
4361 change to the leading bytes of the nonce expecting the new value to be a
4362 new unique nonce then such an application could inadvertently encrypt
4363 messages with a reused nonce.
4365 Additionally the ignored bytes in a long nonce are not covered by the
4366 integrity guarantee of this cipher. Any application that relies on the
4367 integrity of these ignored leading bytes of a long nonce may be further
4368 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4369 is safe because no such use sets such a long nonce value. However user
4370 applications that use this cipher directly and set a non-default nonce
4371 length to be longer than 12 bytes may be vulnerable.
4373 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4379 * Added SCA hardening for modular field inversion in EC_GROUP through
4380 a new dedicated field_inv() pointer in EC_METHOD.
4381 This also addresses a leakage affecting conversions from projective
4382 to affine coordinates.
4384 *Billy Bob Brumley, Nicola Tuveri*
4386 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4387 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4391 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4395 * Remove the 'dist' target and add a tarball building script. The
4396 'dist' target has fallen out of use, and it shouldn't be
4397 necessary to configure just to create a source distribution.
4401 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4403 * Timing vulnerability in DSA signature generation
4405 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4406 timing side channel attack. An attacker could use variations in the signing
4407 algorithm to recover the private key.
4409 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4414 * Timing vulnerability in ECDSA signature generation
4416 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4417 timing side channel attack. An attacker could use variations in the signing
4418 algorithm to recover the private key.
4420 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4425 * Add coordinate blinding for EC_POINT and implement projective
4426 coordinate blinding for generic prime curves as a countermeasure to
4427 chosen point SCA attacks.
4429 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4431 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4433 * Client DoS due to large DH parameter
4435 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4436 malicious server can send a very large prime value to the client. This will
4437 cause the client to spend an unreasonably long period of time generating a
4438 key for this prime resulting in a hang until the client has finished. This
4439 could be exploited in a Denial Of Service attack.
4441 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4446 * Cache timing vulnerability in RSA Key Generation
4448 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4449 a cache timing side channel attack. An attacker with sufficient access to
4450 mount cache timing attacks during the RSA key generation process could
4451 recover the private key.
4453 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4454 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4459 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4460 parameter is no longer accepted, as it leads to a corrupt table. NULL
4461 pem_str is reserved for alias entries only.
4465 * Revert blinding in ECDSA sign and instead make problematic addition
4466 length-invariant. Switch even to fixed-length Montgomery multiplication.
4470 * Change generating and checking of primes so that the error rate of not
4471 being prime depends on the intended use based on the size of the input.
4472 For larger primes this will result in more rounds of Miller-Rabin.
4473 The maximal error rate for primes with more than 1080 bits is lowered
4476 *Kurt Roeckx, Annie Yousar*
4478 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4482 * Add blinding to ECDSA and DSA signatures to protect against side channel
4483 attacks discovered by Keegan Ryan (NCC Group).
4487 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4488 now allow empty (zero character) pass phrases.
4492 * Certificate time validation (X509_cmp_time) enforces stricter
4493 compliance with RFC 5280. Fractional seconds and timezone offsets
4494 are no longer allowed.
4498 * Fixed a text canonicalisation bug in CMS
4500 Where a CMS detached signature is used with text content the text goes
4501 through a canonicalisation process first prior to signing or verifying a
4502 signature. This process strips trailing space at the end of lines, converts
4503 line terminators to CRLF and removes additional trailing line terminators
4504 at the end of a file. A bug in the canonicalisation process meant that
4505 some characters, such as form-feed, were incorrectly treated as whitespace
4506 and removed. This is contrary to the specification (RFC5485). This fix
4507 could mean that detached text data signed with an earlier version of
4508 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4509 signed with a fixed OpenSSL may fail to verify with an earlier version of
4510 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4511 and use the "-binary" flag (for the "cms" command line application) or set
4512 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4516 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4518 * Constructed ASN.1 types with a recursive definition could exceed the stack
4520 Constructed ASN.1 types with a recursive definition (such as can be found
4521 in PKCS7) could eventually exceed the stack given malicious input with
4522 excessive recursion. This could result in a Denial Of Service attack. There
4523 are no such structures used within SSL/TLS that come from untrusted sources
4524 so this is considered safe.
4526 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4532 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4534 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4535 effectively reduced to only comparing the least significant bit of each
4536 byte. This allows an attacker to forge messages that would be considered as
4537 authenticated in an amount of tries lower than that guaranteed by the
4538 security claims of the scheme. The module can only be compiled by the
4539 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4541 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4547 * Add a build target 'build_all_generated', to build all generated files
4548 and only that. This can be used to prepare everything that requires
4549 things like perl for a system that lacks perl and then move everything
4550 to that system and do the rest of the build there.
4554 * Backport SSL_OP_NO_RENGOTIATION
4556 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4557 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4558 changes this is no longer possible in 1.1.0. Therefore, the new
4559 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4560 1.1.0 to provide equivalent functionality.
4562 Note that if an application built against 1.1.0h headers (or above) is run
4563 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4564 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4568 * Removed the OS390-Unix config target. It relied on a script that doesn't
4573 * rsaz_1024_mul_avx2 overflow bug on x86_64
4575 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4576 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4577 Analysis suggests that attacks against RSA and DSA as a result of this
4578 defect would be very difficult to perform and are not believed likely.
4579 Attacks against DH1024 are considered just feasible, because most of the
4580 work necessary to deduce information about a private key may be performed
4581 offline. The amount of resources required for such an attack would be
4582 significant. However, for an attack on TLS to be meaningful, the server
4583 would have to share the DH1024 private key among multiple clients, which is
4584 no longer an option since CVE-2016-0701.
4586 This only affects processors that support the AVX2 but not ADX extensions
4587 like Intel Haswell (4th generation).
4589 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4590 was originally found via the OSS-Fuzz project.
4595 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4597 * bn_sqrx8x_internal carry bug on x86_64
4599 There is a carry propagating bug in the x86_64 Montgomery squaring
4600 procedure. No EC algorithms are affected. Analysis suggests that attacks
4601 against RSA and DSA as a result of this defect would be very difficult to
4602 perform and are not believed likely. Attacks against DH are considered just
4603 feasible (although very difficult) because most of the work necessary to
4604 deduce information about a private key may be performed offline. The amount
4605 of resources required for such an attack would be very significant and
4606 likely only accessible to a limited number of attackers. An attacker would
4607 additionally need online access to an unpatched system using the target
4608 private key in a scenario with persistent DH parameters and a private
4609 key that is shared between multiple clients.
4611 This only affects processors that support the BMI1, BMI2 and ADX extensions
4612 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4614 This issue was reported to OpenSSL by the OSS-Fuzz project.
4619 * Malformed X.509 IPAddressFamily could cause OOB read
4621 If an X.509 certificate has a malformed IPAddressFamily extension,
4622 OpenSSL could do a one-byte buffer overread. The most likely result
4623 would be an erroneous display of the certificate in text format.
4625 This issue was reported to OpenSSL by the OSS-Fuzz project.
4630 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4632 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4633 platform rather than 'mingw'.
4637 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4638 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4639 which is the minimum version we support.
4643 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4645 * Encrypt-Then-Mac renegotiation crash
4647 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4648 negotiated where it was not in the original handshake (or vice-versa) then
4649 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4650 and servers are affected.
4652 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4657 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4659 * Truncated packet could crash via OOB read
4661 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4662 cipher is being used, then a truncated packet can cause that host to
4663 perform an out-of-bounds read, usually resulting in a crash.
4665 This issue was reported to OpenSSL by Robert Święcki of Google.
4670 * Bad (EC)DHE parameters cause a client crash
4672 If a malicious server supplies bad parameters for a DHE or ECDHE key
4673 exchange then this can result in the client attempting to dereference a
4674 NULL pointer leading to a client crash. This could be exploited in a Denial
4677 This issue was reported to OpenSSL by Guido Vranken.
4682 * BN_mod_exp may produce incorrect results on x86_64
4684 There is a carry propagating bug in the x86_64 Montgomery squaring
4685 procedure. No EC algorithms are affected. Analysis suggests that attacks
4686 against RSA and DSA as a result of this defect would be very difficult to
4687 perform and are not believed likely. Attacks against DH are considered just
4688 feasible (although very difficult) because most of the work necessary to
4689 deduce information about a private key may be performed offline. The amount
4690 of resources required for such an attack would be very significant and
4691 likely only accessible to a limited number of attackers. An attacker would
4692 additionally need online access to an unpatched system using the target
4693 private key in a scenario with persistent DH parameters and a private
4694 key that is shared between multiple clients. For example this can occur by
4695 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4696 similar to CVE-2015-3193 but must be treated as a separate problem.
4698 This issue was reported to OpenSSL by the OSS-Fuzz project.
4703 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4705 * ChaCha20/Poly1305 heap-buffer-overflow
4707 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4708 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4709 crash. This issue is not considered to be exploitable beyond a DoS.
4711 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4716 * CMS Null dereference
4718 Applications parsing invalid CMS structures can crash with a NULL pointer
4719 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4720 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4721 structure callback if an attempt is made to free certain invalid encodings.
4722 Only CHOICE structures using a callback which do not handle NULL value are
4725 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4730 * Montgomery multiplication may produce incorrect results
4732 There is a carry propagating bug in the Broadwell-specific Montgomery
4733 multiplication procedure that handles input lengths divisible by, but
4734 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4735 and DH private keys are impossible. This is because the subroutine in
4736 question is not used in operations with the private key itself and an input
4737 of the attacker's direct choice. Otherwise the bug can manifest itself as
4738 transient authentication and key negotiation failures or reproducible
4739 erroneous outcome of public-key operations with specially crafted input.
4740 Among EC algorithms only Brainpool P-512 curves are affected and one
4741 presumably can attack ECDH key negotiation. Impact was not analyzed in
4742 detail, because pre-requisites for attack are considered unlikely. Namely
4743 multiple clients have to choose the curve in question and the server has to
4744 share the private key among them, neither of which is default behaviour.
4745 Even then only clients that chose the curve will be affected.
4747 This issue was publicly reported as transient failures and was not
4748 initially recognized as a security issue. Thanks to Richard Morgan for
4749 providing reproducible case.
4754 * Removed automatic addition of RPATH in shared libraries and executables,
4755 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4759 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4761 * Fix Use After Free for large message sizes
4763 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4764 message larger than approx 16k is received then the underlying buffer to
4765 store the incoming message is reallocated and moved. Unfortunately a
4766 dangling pointer to the old location is left which results in an attempt to
4767 write to the previously freed location. This is likely to result in a
4768 crash, however it could potentially lead to execution of arbitrary code.
4770 This issue only affects OpenSSL 1.1.0a.
4772 This issue was reported to OpenSSL by Robert Święcki.
4777 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4779 * OCSP Status Request extension unbounded memory growth
4781 A malicious client can send an excessively large OCSP Status Request
4782 extension. If that client continually requests renegotiation, sending a
4783 large OCSP Status Request extension each time, then there will be unbounded
4784 memory growth on the server. This will eventually lead to a Denial Of
4785 Service attack through memory exhaustion. Servers with a default
4786 configuration are vulnerable even if they do not support OCSP. Builds using
4787 the "no-ocsp" build time option are not affected.
4789 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4794 * SSL_peek() hang on empty record
4796 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4797 sends an empty record. This could be exploited by a malicious peer in a
4798 Denial Of Service attack.
4800 This issue was reported to OpenSSL by Alex Gaynor.
4805 * Excessive allocation of memory in tls_get_message_header() and
4806 dtls1_preprocess_fragment()
4808 A (D)TLS message includes 3 bytes for its length in the header for the
4809 message. This would allow for messages up to 16Mb in length. Messages of
4810 this length are excessive and OpenSSL includes a check to ensure that a
4811 peer is sending reasonably sized messages in order to avoid too much memory
4812 being consumed to service a connection. A flaw in the logic of version
4813 1.1.0 means that memory for the message is allocated too early, prior to
4814 the excessive message length check. Due to way memory is allocated in
4815 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4816 to service a connection. This could lead to a Denial of Service through
4817 memory exhaustion. However, the excessive message length check still takes
4818 place, and this would cause the connection to immediately fail. Assuming
4819 that the application calls SSL_free() on the failed connection in a timely
4820 manner then the 21Mb of allocated memory will then be immediately freed
4821 again. Therefore, the excessive memory allocation will be transitory in
4822 nature. This then means that there is only a security impact if:
4824 1) The application does not call SSL_free() in a timely manner in the event
4825 that the connection fails
4827 2) The application is working in a constrained environment where there is
4828 very little free memory
4830 3) The attacker initiates multiple connection attempts such that there are
4831 multiple connections in a state where memory has been allocated for the
4832 connection; SSL_free() has not yet been called; and there is insufficient
4833 memory to service the multiple requests.
4835 Except in the instance of (1) above any Denial Of Service is likely to be
4836 transitory because as soon as the connection fails the memory is
4837 subsequently freed again in the SSL_free() call. However there is an
4838 increased risk during this period of application crashes due to the lack of
4839 memory - which would then mean a more serious Denial of Service.
4841 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4842 (CVE-2016-6307 and CVE-2016-6308)
4846 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4847 had to be removed. Primary reason is that vendor assembler can't
4848 assemble our modules with -KPIC flag. As result it, assembly
4849 support, was not even available as option. But its lack means
4850 lack of side-channel resistant code, which is incompatible with
4851 security by todays standards. Fortunately gcc is readily available
4852 prepackaged option, which we firmly point at...
4856 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4858 * Windows command-line tool supports UTF-8 opt-in option for arguments
4859 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4860 (to any value) allows Windows user to access PKCS#12 file generated
4861 with Windows CryptoAPI and protected with non-ASCII password, as well
4862 as files generated under UTF-8 locale on Linux also protected with
4867 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4868 have been disabled by default and removed from DEFAULT, just like RC4.
4869 See the RC4 item below to re-enable both.
4873 * The method for finding the storage location for the Windows RAND seed file
4874 has changed. First we check %RANDFILE%. If that is not set then we check
4875 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4876 all else fails we fall back to C:\.
4880 * The EVP_EncryptUpdate() function has had its return type changed from void
4881 to int. A return of 0 indicates and error while a return of 1 indicates
4886 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4887 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4888 off the constant time implementation for RSA, DSA and DH have been made
4889 no-ops and deprecated.
4893 * Windows RAND implementation was simplified to only get entropy by
4894 calling CryptGenRandom(). Various other RAND-related tickets
4897 *Joseph Wylie Yandle, Rich Salz*
4899 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4900 and `OPENSSL_LH_`, respectively. The old names are available
4901 with API compatibility. They new names are now completely documented.
4905 * Unify TYPE_up_ref(obj) methods signature.
4906 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4907 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4908 int (instead of void) like all others TYPE_up_ref() methods.
4909 So now these methods also check the return value of CRYPTO_atomic_add(),
4910 and the validity of object reference counter.
4912 *fdasilvayy@gmail.com*
4914 * With Windows Visual Studio builds, the .pdb files are installed
4915 alongside the installed libraries and executables. For a static
4916 library installation, ossl_static.pdb is the associate compiler
4917 generated .pdb file to be used when linking programs.
4921 * Remove openssl.spec. Packaging files belong with the packagers.
4925 * Automatic Darwin/OSX configuration has had a refresh, it will now
4926 recognise x86_64 architectures automatically. You can still decide
4927 to build for a different bitness with the environment variable
4928 KERNEL_BITS (can be 32 or 64), for example:
4930 KERNEL_BITS=32 ./config
4934 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4935 256 bit AES and HMAC with SHA256.
4939 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4943 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4947 * To enable users to have their own config files and build file templates,
4948 Configure looks in the directory indicated by the environment variable
4949 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4950 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4951 name and is used as is.
4955 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4956 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4957 X509_CERT_FILE_CTX was removed.
4961 * "shared" builds are now the default. To create only static libraries use
4962 the "no-shared" Configure option.
4966 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4967 All of these option have not worked for some while and are fundamental
4972 * Make various cleanup routines no-ops and mark them as deprecated. Most
4973 global cleanup functions are no longer required because they are handled
4974 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4975 Explicitly de-initing can cause problems (e.g. where a library that uses
4976 OpenSSL de-inits, but an application is still using it). The affected
4977 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4978 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4979 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4980 COMP_zlib_cleanup().
4984 * --strict-warnings no longer enables runtime debugging options
4985 such as REF_DEBUG. Instead, debug options are automatically
4986 enabled with '--debug' builds.
4988 *Andy Polyakov, Emilia Käsper*
4990 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4991 have been moved out of the public header files. New functions for managing
4992 these have been added.
4996 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4997 objects have been moved out of the public header files. New
4998 functions for managing these have been added.
5002 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
5003 have been moved out of the public header files. New functions for managing
5004 these have been added.
5008 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
5009 moved out of the public header files. New functions for managing these
5014 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
5018 * Removed the mk1mf build scripts.
5022 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
5023 it is always safe to #include a header now.
5027 * Removed the aged BC-32 config and all its supporting scripts
5031 * Removed support for Ultrix, Netware, and OS/2.
5035 * Add support for HKDF.
5037 *Alessandro Ghedini*
5039 * Add support for blake2b and blake2s
5043 * Added support for "pipelining". Ciphers that have the
5044 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
5045 encryptions/decryptions simultaneously. There are currently no built-in
5046 ciphers with this property but the expectation is that engines will be able
5047 to offer it to significantly improve throughput. Support has been extended
5048 into libssl so that multiple records for a single connection can be
5049 processed in one go (for >=TLS 1.1).
5053 * Added the AFALG engine. This is an async capable engine which is able to
5054 offload work to the Linux kernel. In this initial version it only supports
5055 AES128-CBC. The kernel must be version 4.1.0 or greater.
5059 * OpenSSL now uses a new threading API. It is no longer necessary to
5060 set locking callbacks to use OpenSSL in a multi-threaded environment. There
5061 are two supported threading models: pthreads and windows threads. It is
5062 also possible to configure OpenSSL at compile time for "no-threads". The
5063 old threading API should no longer be used. The functions have been
5064 replaced with "no-op" compatibility macros.
5066 *Alessandro Ghedini, Matt Caswell*
5068 * Modify behavior of ALPN to invoke callback after SNI/servername
5069 callback, such that updates to the SSL_CTX affect ALPN.
5073 * Add SSL_CIPHER queries for authentication and key-exchange.
5077 * Changes to the DEFAULT cipherlist:
5078 - Prefer (EC)DHE handshakes over plain RSA.
5079 - Prefer AEAD ciphers over legacy ciphers.
5080 - Prefer ECDSA over RSA when both certificates are available.
5081 - Prefer TLSv1.2 ciphers/PRF.
5082 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
5087 * Change the ECC default curve list to be this, in order: x25519,
5088 secp256r1, secp521r1, secp384r1.
5092 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
5093 disabled by default. They can be re-enabled using the
5094 enable-weak-ssl-ciphers option to Configure.
5098 * If the server has ALPN configured, but supports no protocols that the
5099 client advertises, send a fatal "no_application_protocol" alert.
5100 This behaviour is SHALL in RFC 7301, though it isn't universally
5101 implemented by other servers.
5105 * Add X25519 support.
5106 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
5107 for public and private key encoding using the format documented in
5108 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
5109 key generation and key derivation.
5111 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
5116 * Deprecate SRP_VBASE_get_by_user.
5117 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5118 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
5119 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
5120 seed, even if the seed is configured.
5122 Users should use SRP_VBASE_get1_by_user instead. Note that in
5123 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5124 also that even though configuring the SRP seed attempts to hide
5125 invalid usernames by continuing the handshake with fake
5126 credentials, this behaviour is not constant time and no strong
5127 guarantees are made that the handshake is indistinguishable from
5128 that of a valid user.
5132 * Configuration change; it's now possible to build dynamic engines
5133 without having to build shared libraries and vice versa. This
5134 only applies to the engines in `engines/`, those in `crypto/engine/`
5135 will always be built into libcrypto (i.e. "static").
5137 Building dynamic engines is enabled by default; to disable, use
5138 the configuration option "disable-dynamic-engine".
5140 The only requirements for building dynamic engines are the
5141 presence of the DSO module and building with position independent
5142 code, so they will also automatically be disabled if configuring
5143 with "disable-dso" or "disable-pic".
5145 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
5146 are also taken away from openssl/opensslconf.h, as they are
5151 * Configuration change; if there is a known flag to compile
5152 position independent code, it will always be applied on the
5153 libcrypto and libssl object files, and never on the application
5154 object files. This means other libraries that use routines from
5155 libcrypto / libssl can be made into shared libraries regardless
5156 of how OpenSSL was configured.
5158 If this isn't desirable, the configuration options "disable-pic"
5159 or "no-pic" can be used to disable the use of PIC. This will
5160 also disable building shared libraries and dynamic engines.
5164 * Removed JPAKE code. It was experimental and has no wide use.
5168 * The INSTALL_PREFIX Makefile variable has been renamed to
5169 DESTDIR. That makes for less confusion on what this variable
5170 is for. Also, the configuration option --install_prefix is
5175 * Heartbeat for TLS has been removed and is disabled by default
5176 for DTLS; configure with enable-heartbeats. Code that uses the
5177 old #define's might need to be updated.
5179 *Emilia Käsper, Rich Salz*
5181 * Rename REF_CHECK to REF_DEBUG.
5185 * New "unified" build system
5187 The "unified" build system is aimed to be a common system for all
5188 platforms we support. With it comes new support for VMS.
5190 This system builds supports building in a different directory tree
5191 than the source tree. It produces one Makefile (for unix family
5192 or lookalikes), or one descrip.mms (for VMS).
5194 The source of information to make the Makefile / descrip.mms is
5195 small files called 'build.info', holding the necessary
5196 information for each directory with source to compile, and a
5197 template in Configurations, like unix-Makefile.tmpl or
5200 With this change, the library names were also renamed on Windows
5201 and on VMS. They now have names that are closer to the standard
5202 on Unix, and include the major version number, and in certain
5203 cases, the architecture they are built for. See "Notes on shared
5204 libraries" in INSTALL.
5206 We rely heavily on the perl module Text::Template.
5210 * Added support for auto-initialisation and de-initialisation of the library.
5211 OpenSSL no longer requires explicit init or deinit routines to be called,
5212 except in certain circumstances. See the OPENSSL_init_crypto() and
5213 OPENSSL_init_ssl() man pages for further information.
5217 * The arguments to the DTLSv1_listen function have changed. Specifically the
5218 "peer" argument is now expected to be a BIO_ADDR object.
5220 * Rewrite of BIO networking library. The BIO library lacked consistent
5221 support of IPv6, and adding it required some more extensive
5222 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
5223 which hold all types of addresses and chains of address information.
5224 It also introduces a new API, with functions like BIO_socket,
5225 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
5226 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
5227 have been adapted accordingly.
5231 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
5236 * CRIME protection: disable compression by default, even if OpenSSL is
5237 compiled with zlib enabled. Applications can still enable compression
5238 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
5239 using the SSL_CONF library to configure compression.
5243 * The signature of the session callback configured with
5244 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
5245 was explicitly marked as `const unsigned char*` instead of
5250 * Always DPURIFY. Remove the use of uninitialized memory in the
5251 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
5255 * Removed many obsolete configuration items, including
5256 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
5257 MD2_CHAR, MD2_INT, MD2_LONG
5259 IDEA_SHORT, IDEA_LONG
5260 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5262 *Rich Salz, with advice from Andy Polyakov*
5264 * Many BN internals have been moved to an internal header file.
5266 *Rich Salz with help from Andy Polyakov*
5268 * Configuration and writing out the results from it has changed.
5269 Files such as Makefile include/openssl/opensslconf.h and are now
5270 produced through general templates, such as Makefile.in and
5271 crypto/opensslconf.h.in and some help from the perl module
5274 Also, the center of configuration information is no longer
5275 Makefile. Instead, Configure produces a perl module in
5276 configdata.pm which holds most of the config data (in the hash
5277 table %config), the target data that comes from the target
5278 configuration in one of the `Configurations/*.conf` files (in
5283 * To clarify their intended purposes, the Configure options
5284 --prefix and --openssldir change their semantics, and become more
5285 straightforward and less interdependent.
5287 --prefix shall be used exclusively to give the location INSTALLTOP
5288 where programs, scripts, libraries, include files and manuals are
5289 going to be installed. The default is now /usr/local.
5291 --openssldir shall be used exclusively to give the default
5292 location OPENSSLDIR where certificates, private keys, CRLs are
5293 managed. This is also where the default openssl.cnf gets
5295 If the directory given with this option is a relative path, the
5296 values of both the --prefix value and the --openssldir value will
5297 be combined to become OPENSSLDIR.
5298 The default for --openssldir is INSTALLTOP/ssl.
5300 Anyone who uses --openssldir to specify where OpenSSL is to be
5301 installed MUST change to use --prefix instead.
5305 * The GOST engine was out of date and therefore it has been removed. An up
5306 to date GOST engine is now being maintained in an external repository.
5307 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5308 support for GOST ciphersuites (these are only activated if a GOST engine
5313 * EGD is no longer supported by default; use enable-egd when
5316 *Ben Kaduk and Rich Salz*
5318 * The distribution now has Makefile.in files, which are used to
5319 create Makefile's when Configure is run. *Configure must be run
5320 before trying to build now.*
5324 * The return value for SSL_CIPHER_description() for error conditions
5329 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5331 Obtaining and performing DNSSEC validation of TLSA records is
5332 the application's responsibility. The application provides
5333 the TLSA records of its choice to OpenSSL, and these are then
5334 used to authenticate the peer.
5336 The TLSA records need not even come from DNS. They can, for
5337 example, be used to implement local end-entity certificate or
5338 trust-anchor "pinning", where the "pin" data takes the form
5339 of TLSA records, which can augment or replace verification
5340 based on the usual WebPKI public certification authorities.
5344 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5345 continues to support deprecated interfaces in default builds.
5346 However, applications are strongly advised to compile their
5347 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5348 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5349 or the 1.1.0 releases.
5351 In environments in which all applications have been ported to
5352 not use any deprecated interfaces OpenSSL's Configure script
5353 should be used with the --api=1.1.0 option to entirely remove
5354 support for the deprecated features from the library and
5355 unconditionally disable them in the installed headers.
5356 Essentially the same effect can be achieved with the "no-deprecated"
5357 argument to Configure, except that this will always restrict
5358 the build to just the latest API, rather than a fixed API
5361 As applications are ported to future revisions of the API,
5362 they should update their compile-time OPENSSL_API_COMPAT define
5363 accordingly, but in most cases should be able to continue to
5364 compile with later releases.
5366 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5367 0x10000000L and 0x00908000L, respectively. However those
5368 versions did not support the OPENSSL_API_COMPAT feature, and
5369 so applications are not typically tested for explicit support
5370 of just the undeprecated features of either release.
5374 * Add support for setting the minimum and maximum supported protocol.
5375 It can bet set via the SSL_set_min_proto_version() and
5376 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5377 MaxProtocol. It's recommended to use the new APIs to disable
5378 protocols instead of disabling individual protocols using
5379 SSL_set_options() or SSL_CONF's Protocol. This change also
5380 removes support for disabling TLS 1.2 in the OpenSSL TLS
5381 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5385 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5389 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5390 and integrates ECDSA and ECDH functionality into EC. Implementations can
5391 now redirect key generation and no longer need to convert to or from
5394 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5395 include the ec.h header file instead.
5399 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5400 ciphers who are no longer supported and drops support the ephemeral RSA key
5401 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5405 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5406 opaque. For HMAC_CTX, the following constructors and destructors
5409 HMAC_CTX *HMAC_CTX_new(void);
5410 void HMAC_CTX_free(HMAC_CTX *ctx);
5412 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5413 destroy such methods has been added. See EVP_MD_meth_new(3) and
5414 EVP_CIPHER_meth_new(3) for documentation.
5417 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5418 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5419 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5420 an already created structure.
5421 2) For consistency with the majority of our object creators and
5422 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5423 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5424 for deprecated builds.
5428 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5429 cryptographic operations to be performed asynchronously as long as an
5430 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5431 further details. Libssl has also had this capability integrated with the
5432 introduction of the new mode SSL_MODE_ASYNC and associated error
5433 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5434 pages. This work was developed in partnership with Intel Corp.
5438 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5439 always enabled now. If you want to disable the support you should
5440 exclude it using the list of supported ciphers. This also means that the
5441 "-no_ecdhe" option has been removed from s_server.
5445 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5446 SSL_{CTX_}set1_curves() which can set a list.
5450 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5451 curve you want to support using SSL_{CTX_}set1_curves().
5455 * State machine rewrite. The state machine code has been significantly
5456 refactored in order to remove much duplication of code and solve issues
5457 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5458 further details). This change does have some associated API changes.
5459 Notably the SSL_state() function has been removed and replaced by
5460 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5461 SSL_set_state() has been removed altogether. The previous handshake states
5462 defined in ssl.h and ssl3.h have also been removed.
5466 * All instances of the string "ssleay" in the public API were replaced
5467 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5468 Some error codes related to internal RSA_eay API's were renamed.
5472 * The demo files in crypto/threads were moved to demo/threads.
5476 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5479 *Matt Caswell, Rich Salz*
5481 * New ASN.1 embed macro.
5483 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5484 structure is not allocated: it is part of the parent. That is instead of
5492 This reduces memory fragmentation and make it impossible to accidentally
5493 set a mandatory field to NULL.
5495 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5496 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5497 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5502 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5506 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5507 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5508 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5509 DES and RC4 ciphersuites.
5513 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5514 This changes the decoding behaviour for some invalid messages,
5515 though the change is mostly in the more lenient direction, and
5516 legacy behaviour is preserved as much as possible.
5520 * Fix no-stdio build.
5521 *David Woodhouse <David.Woodhouse@intel.com> and also*
5522 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5524 * New testing framework
5525 The testing framework has been largely rewritten and is now using
5526 perl and the perl modules Test::Harness and an extended variant of
5527 Test::More called OpenSSL::Test to do its work. All test scripts in
5528 test/ have been rewritten into test recipes, and all direct calls to
5529 executables in test/Makefile have become individual recipes using the
5530 simplified testing OpenSSL::Test::Simple.
5532 For documentation on our testing modules, do:
5534 perldoc test/testlib/OpenSSL/Test/Simple.pm
5535 perldoc test/testlib/OpenSSL/Test.pm
5539 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5540 are used; the latter aborts on memory leaks (usually checked on exit).
5541 Some undocumented "set malloc, etc., hooks" functions were removed
5542 and others were changed. All are now documented.
5546 * In DSA_generate_parameters_ex, if the provided seed is too short,
5549 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5551 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5552 from RFC4279, RFC4785, RFC5487, RFC5489.
5554 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5555 original RSA_PSK patch.
5559 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5560 era flag was never set throughout the codebase (only read). Also removed
5561 SSL3_FLAGS_POP_BUFFER which was only used if
5562 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5566 * Changed the default name options in the "ca", "crl", "req" and "x509"
5567 to be "oneline" instead of "compat".
5571 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5572 not aware of clients that still exhibit this bug, and the workaround
5573 hasn't been working properly for a while.
5577 * The return type of BIO_number_read() and BIO_number_written() as well as
5578 the corresponding num_read and num_write members in the BIO structure has
5579 changed from unsigned long to uint64_t. On platforms where an unsigned
5580 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5585 * Given the pervasive nature of TLS extensions it is inadvisable to run
5586 OpenSSL without support for them. It also means that maintaining
5587 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5588 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5592 * Removed support for the two export grade static DH ciphersuites
5593 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5594 were newly added (along with a number of other static DH ciphersuites) to
5595 1.0.2. However the two export ones have *never* worked since they were
5596 introduced. It seems strange in any case to be adding new export
5597 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5601 * Version negotiation has been rewritten. In particular SSLv23_method(),
5602 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5603 and turned into macros which simply call the new preferred function names
5604 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5605 should use the new names instead. Also as part of this change the ssl23.h
5606 header file has been removed.
5610 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5611 code and the associated standard is no longer considered fit-for-purpose.
5615 * RT2547 was closed. When generating a private key, try to make the
5616 output file readable only by the owner. This behavior change might
5617 be noticeable when interacting with other software.
5619 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5624 * Added HTTP GET support to the ocsp command.
5628 * Changed default digest for the dgst and enc commands from MD5 to
5633 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5637 * Added support for TLS extended master secret from
5638 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5639 initial patch which was a great help during development.
5643 * All libssl internal structures have been removed from the public header
5644 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5645 now redundant). Users should not attempt to access internal structures
5646 directly. Instead they should use the provided API functions.
5650 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5651 Access to deprecated functions can be re-enabled by running config with
5652 "enable-deprecated". In addition applications wishing to use deprecated
5653 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5654 will, by default, disable some transitive includes that previously existed
5655 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5659 * Added support for OCB mode. OpenSSL has been granted a patent license
5660 compatible with the OpenSSL license for use of OCB. Details are available
5661 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5662 for OCB can be removed by calling config with no-ocb.
5666 * SSLv2 support has been removed. It still supports receiving an SSLv2
5667 compatible client hello.
5671 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5672 done while fixing the error code for the key-too-small case.
5674 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5676 * CA.sh has been removed; use CA.pl instead.
5680 * Removed old DES API.
5684 * Remove various unsupported platforms:
5690 Sinix/ReliantUNIX RM400
5695 16-bit platforms such as WIN16
5699 * Clean up OPENSSL_NO_xxx #define's
5700 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5701 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5702 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5703 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5704 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5705 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5706 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5707 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5708 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5709 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5713 * Cleaned up dead code
5714 Remove all but one '#ifdef undef' which is to be looked at.
5718 * Clean up calling of xxx_free routines.
5719 Just like free(), fix most of the xxx_free routines to accept
5720 NULL. Remove the non-null checks from callers. Save much code.
5724 * Add secure heap for storage of private keys (when possible).
5725 Add BIO_s_secmem(), CBIGNUM, etc.
5726 Contributed by Akamai Technologies under our Corporate CLA.
5730 * Experimental support for a new, fast, unbiased prime candidate generator,
5731 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5733 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5735 * New output format NSS in the sess_id command line tool. This allows
5736 exporting the session id and the master key in NSS keylog format.
5738 *Martin Kaiser <martin@kaiser.cx>*
5740 * Harmonize version and its documentation. -f flag is used to display
5743 *mancha <mancha1@zoho.com>*
5745 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5746 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5748 *mancha <mancha1@zoho.com>*
5750 * Fix some double frees. These are not thought to be exploitable.
5752 *mancha <mancha1@zoho.com>*
5754 * A missing bounds check in the handling of the TLS heartbeat extension
5755 can be used to reveal up to 64k of memory to a connected client or
5758 Thanks for Neel Mehta of Google Security for discovering this bug and to
5759 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5760 preparing the fix ([CVE-2014-0160])
5762 *Adam Langley, Bodo Moeller*
5764 * Fix for the attack described in the paper "Recovering OpenSSL
5765 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5766 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5767 <http://eprint.iacr.org/2014/140>
5769 Thanks to Yuval Yarom and Naomi Benger for discovering this
5770 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5772 *Yuval Yarom and Naomi Benger*
5774 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5775 this fixes a limitation in previous versions of OpenSSL.
5779 * Experimental encrypt-then-mac support.
5781 Experimental support for encrypt then mac from
5782 draft-gutmann-tls-encrypt-then-mac-02.txt
5784 To enable it set the appropriate extension number (0x42 for the test
5785 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5787 For non-compliant peers (i.e. just about everything) this should have no
5790 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5794 * Add EVP support for key wrapping algorithms, to avoid problems with
5795 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5796 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5797 algorithms and include tests cases.
5801 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5806 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5807 MGF1 digest and OAEP label.
5811 * Make openssl verify return errors.
5813 *Chris Palmer <palmer@google.com> and Ben Laurie*
5815 * New function ASN1_TIME_diff to calculate the difference between two
5816 ASN1_TIME structures or one structure and the current time.
5820 * Update fips_test_suite to support multiple command line options. New
5821 test to induce all self test errors in sequence and check expected
5826 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5827 sign or verify all in one operation.
5831 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5832 test programs and fips_test_suite. Includes functionality to parse
5833 the minimal script output of fipsalgest.pl directly.
5837 * Add authorisation parameter to FIPS_module_mode_set().
5841 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5845 * Use separate DRBG fields for internal and external flags. New function
5846 FIPS_drbg_health_check() to perform on demand health checking. Add
5847 generation tests to fips_test_suite with reduced health check interval to
5848 demonstrate periodic health checking. Add "nodh" option to
5849 fips_test_suite to skip very slow DH test.
5853 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5858 * More extensive health check for DRBG checking many more failure modes.
5859 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5860 combination: call this in fips_test_suite.
5864 * Add support for canonical generation of DSA parameter 'g'. See
5867 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5868 POST to handle HMAC cases.
5872 * Add functions FIPS_module_version() and FIPS_module_version_text()
5873 to return numerical and string versions of the FIPS module number.
5877 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5878 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5879 outside the validated module in the FIPS capable OpenSSL.
5883 * Minor change to DRBG entropy callback semantics. In some cases
5884 there is no multiple of the block length between min_len and
5885 max_len. Allow the callback to return more than max_len bytes
5886 of entropy but discard any extra: it is the callback's responsibility
5887 to ensure that the extra data discarded does not impact the
5888 requested amount of entropy.
5892 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5893 information in FIPS186-3, SP800-57 and SP800-131A.
5897 * CCM support via EVP. Interface is very similar to GCM case except we
5898 must supply all data in one chunk (i.e. no update, final) and the
5899 message length must be supplied if AAD is used. Add algorithm test
5904 * Initial version of POST overhaul. Add POST callback to allow the status
5905 of POST to be monitored and/or failures induced. Modify fips_test_suite
5906 to use callback. Always run all selftests even if one fails.
5910 * XTS support including algorithm test driver in the fips_gcmtest program.
5911 Note: this does increase the maximum key length from 32 to 64 bytes but
5912 there should be no binary compatibility issues as existing applications
5913 will never use XTS mode.
5917 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5918 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5919 performs algorithm blocking for unapproved PRNG types. Also do not
5920 set PRNG type in FIPS_mode_set(): leave this to the application.
5921 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5922 the standard OpenSSL PRNG: set additional data to a date time vector.
5926 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5927 This shouldn't present any incompatibility problems because applications
5928 shouldn't be using these directly and any that are will need to rethink
5929 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5933 * Extensive self tests and health checking required by SP800-90 DRBG.
5934 Remove strength parameter from FIPS_drbg_instantiate and always
5935 instantiate at maximum supported strength.
5939 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5943 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5947 * New function DH_compute_key_padded() to compute a DH key and pad with
5948 leading zeroes if needed: this complies with SP800-56A et al.
5952 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5953 anything, incomplete, subject to change and largely untested at present.
5957 * Modify fipscanisteronly build option to only build the necessary object
5958 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5962 * Add experimental option FIPSSYMS to give all symbols in
5963 fipscanister.o and FIPS or fips prefix. This will avoid
5964 conflicts with future versions of OpenSSL. Add perl script
5965 util/fipsas.pl to preprocess assembly language source files
5966 and rename any affected symbols.
5970 * Add selftest checks and algorithm block of non-fips algorithms in
5971 FIPS mode. Remove DES2 from selftests.
5975 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5976 return internal method without any ENGINE dependencies. Add new
5977 tiny fips sign and verify functions.
5981 * New build option no-ec2m to disable characteristic 2 code.
5985 * New build option "fipscanisteronly". This only builds fipscanister.o
5986 and (currently) associated fips utilities. Uses the file Makefile.fips
5987 instead of Makefile.org as the prototype.
5991 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5992 Update fips_gcmtest to use IV generator.
5996 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5997 setting output buffer to NULL. The `*Final` function must be
5998 called although it will not retrieve any additional data. The tag
5999 can be set or retrieved with a ctrl. The IV length is by default 12
6000 bytes (96 bits) but can be set to an alternative value. If the IV
6001 length exceeds the maximum IV length (currently 16 bytes) it cannot be
6006 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
6007 underlying do_cipher function handles all cipher semantics itself
6008 including padding and finalisation. This is useful if (for example)
6009 an ENGINE cipher handles block padding itself. The behaviour of
6010 do_cipher is subtly changed if this flag is set: the return value
6011 is the number of characters written to the output buffer (zero is
6012 no longer an error code) or a negative error code. Also if the
6013 input buffer is NULL and length 0 finalisation should be performed.
6017 * If a candidate issuer certificate is already part of the constructed
6018 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
6022 * Improve forward-security support: add functions
6024 void SSL_CTX_set_not_resumable_session_callback(
6025 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
6026 void SSL_set_not_resumable_session_callback(
6027 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
6029 for use by SSL/TLS servers; the callback function will be called whenever a
6030 new session is created, and gets to decide whether the session may be
6031 cached to make it resumable (return 0) or not (return 1). (As by the
6032 SSL/TLS protocol specifications, the session_id sent by the server will be
6033 empty to indicate that the session is not resumable; also, the server will
6034 not generate RFC 4507 (RFC 5077) session tickets.)
6036 A simple reasonable callback implementation is to return is_forward_secure.
6037 This parameter will be set to 1 or 0 depending on the ciphersuite selected
6038 by the SSL/TLS server library, indicating whether it can provide forward
6041 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
6043 * New -verify_name option in command line utilities to set verification
6048 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
6049 Add CMAC pkey methods.
6053 * Experimental renegotiation in s_server -www mode. If the client
6054 browses /reneg connection is renegotiated. If /renegcert it is
6055 renegotiated requesting a certificate.
6059 * Add an "external" session cache for debugging purposes to s_server. This
6060 should help trace issues which normally are only apparent in deployed
6061 multi-process servers.
6065 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
6066 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
6067 BIO_set_cipher() and some obscure PEM functions were changed so they
6068 can now return an error. The RAND changes required a change to the
6069 RAND_METHOD structure.
6073 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
6074 a gcc attribute to warn if the result of a function is ignored. This
6075 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
6076 whose return value is often ignored.
6080 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
6081 These allow SCTs (signed certificate timestamps) to be requested and
6082 validated when establishing a connection.
6084 *Rob Percival <robpercival@google.com>*
6089 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
6091 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
6092 used even when parsing explicit parameters, when loading a encoded key
6093 or calling `EC_GROUP_new_from_ecpkparameters()`/
6094 `EC_GROUP_new_from_ecparameters()`.
6095 This prevents bypass of security hardening and performance gains,
6096 especially for curves with specialized EC_METHODs.
6097 By default, if a key encoded with explicit parameters is loaded and later
6098 encoded, the output is still encoded with explicit parameters, even if
6099 internally a "named" EC_GROUP is used for computation.
6103 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
6104 this change, EC_GROUP_set_generator would accept order and/or cofactor as
6105 NULL. After this change, only the cofactor parameter can be NULL. It also
6106 does some minimal sanity checks on the passed order.
6111 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
6112 An attack is simple, if the first CMS_recipientInfo is valid but the
6113 second CMS_recipientInfo is chosen ciphertext. If the second
6114 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
6115 encryption key will be replaced by garbage, and the message cannot be
6116 decoded, but if the RSA decryption fails, the correct encryption key is
6117 used and the recipient will not notice the attack.
6118 As a work around for this potential attack the length of the decrypted
6119 key must be equal to the cipher default key length, in case the
6120 certificate is not given and all recipientInfo are tried out.
6121 The old behaviour can be re-enabled in the CMS code by setting the
6122 CMS_DEBUG_DECRYPT flag.
6127 * Document issue with installation paths in diverse Windows builds
6129 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
6130 binaries and run-time config file.
6135 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
6137 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
6138 This changes the size when using the `genpkey` command when no size is given.
6139 It fixes an omission in earlier changes that changed all RSA, DSA and DH
6140 generation commands to use 2048 bits by default.
6144 * Add FIPS support for Android Arm 64-bit
6146 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
6147 Module in Version 2.0.10. For some reason, the corresponding target
6148 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
6149 built with FIPS support on Android Arm 64-bit. This omission has been
6152 *Matthias St. Pierre*
6154 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
6156 * 0-byte record padding oracle
6158 If an application encounters a fatal protocol error and then calls
6159 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
6160 then OpenSSL can respond differently to the calling application if a 0 byte
6161 record is received with invalid padding compared to if a 0 byte record is
6162 received with an invalid MAC. If the application then behaves differently
6163 based on that in a way that is detectable to the remote peer, then this
6164 amounts to a padding oracle that could be used to decrypt data.
6166 In order for this to be exploitable "non-stitched" ciphersuites must be in
6167 use. Stitched ciphersuites are optimised implementations of certain
6168 commonly used ciphersuites. Also the application must call SSL_shutdown()
6169 twice even if a protocol error has occurred (applications should not do
6170 this but some do anyway).
6172 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
6173 Aviram, with additional investigation by Steven Collison and Andrew
6174 Hourselt. It was reported to OpenSSL on 10th December 2018.
6179 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
6183 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
6185 * Microarchitecture timing vulnerability in ECC scalar multiplication
6187 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
6188 shown to be vulnerable to a microarchitecture timing side channel attack.
6189 An attacker with sufficient access to mount local timing attacks during
6190 ECDSA signature generation could recover the private key.
6192 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
6193 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
6199 * Timing vulnerability in DSA signature generation
6201 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
6202 timing side channel attack. An attacker could use variations in the signing
6203 algorithm to recover the private key.
6205 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
6210 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
6211 Module, accidentally introduced while backporting security fixes from the
6212 development branch and hindering the use of ECC in FIPS mode.
6216 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
6218 * Client DoS due to large DH parameter
6220 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
6221 malicious server can send a very large prime value to the client. This will
6222 cause the client to spend an unreasonably long period of time generating a
6223 key for this prime resulting in a hang until the client has finished. This
6224 could be exploited in a Denial Of Service attack.
6226 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
6231 * Cache timing vulnerability in RSA Key Generation
6233 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
6234 a cache timing side channel attack. An attacker with sufficient access to
6235 mount cache timing attacks during the RSA key generation process could
6236 recover the private key.
6238 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
6239 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
6244 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
6245 parameter is no longer accepted, as it leads to a corrupt table. NULL
6246 pem_str is reserved for alias entries only.
6250 * Revert blinding in ECDSA sign and instead make problematic addition
6251 length-invariant. Switch even to fixed-length Montgomery multiplication.
6255 * Change generating and checking of primes so that the error rate of not
6256 being prime depends on the intended use based on the size of the input.
6257 For larger primes this will result in more rounds of Miller-Rabin.
6258 The maximal error rate for primes with more than 1080 bits is lowered
6261 *Kurt Roeckx, Annie Yousar*
6263 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6267 * Add blinding to ECDSA and DSA signatures to protect against side channel
6268 attacks discovered by Keegan Ryan (NCC Group).
6272 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6273 now allow empty (zero character) pass phrases.
6277 * Certificate time validation (X509_cmp_time) enforces stricter
6278 compliance with RFC 5280. Fractional seconds and timezone offsets
6279 are no longer allowed.
6283 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6285 * Constructed ASN.1 types with a recursive definition could exceed the stack
6287 Constructed ASN.1 types with a recursive definition (such as can be found
6288 in PKCS7) could eventually exceed the stack given malicious input with
6289 excessive recursion. This could result in a Denial Of Service attack. There
6290 are no such structures used within SSL/TLS that come from untrusted sources
6291 so this is considered safe.
6293 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6299 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6301 * Read/write after SSL object in error state
6303 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6304 mechanism. The intent was that if a fatal error occurred during a handshake
6305 then OpenSSL would move into the error state and would immediately fail if
6306 you attempted to continue the handshake. This works as designed for the
6307 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6308 SSL_connect()), however due to a bug it does not work correctly if
6309 SSL_read() or SSL_write() is called directly. In that scenario, if the
6310 handshake fails then a fatal error will be returned in the initial function
6311 call. If SSL_read()/SSL_write() is subsequently called by the application
6312 for the same SSL object then it will succeed and the data is passed without
6313 being decrypted/encrypted directly from the SSL/TLS record layer.
6315 In order to exploit this issue an application bug would have to be present
6316 that resulted in a call to SSL_read()/SSL_write() being issued after having
6317 already received a fatal error.
6319 This issue was reported to OpenSSL by David Benjamin (Google).
6324 * rsaz_1024_mul_avx2 overflow bug on x86_64
6326 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6327 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6328 Analysis suggests that attacks against RSA and DSA as a result of this
6329 defect would be very difficult to perform and are not believed likely.
6330 Attacks against DH1024 are considered just feasible, because most of the
6331 work necessary to deduce information about a private key may be performed
6332 offline. The amount of resources required for such an attack would be
6333 significant. However, for an attack on TLS to be meaningful, the server
6334 would have to share the DH1024 private key among multiple clients, which is
6335 no longer an option since CVE-2016-0701.
6337 This only affects processors that support the AVX2 but not ADX extensions
6338 like Intel Haswell (4th generation).
6340 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6341 was originally found via the OSS-Fuzz project.
6346 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6348 * bn_sqrx8x_internal carry bug on x86_64
6350 There is a carry propagating bug in the x86_64 Montgomery squaring
6351 procedure. No EC algorithms are affected. Analysis suggests that attacks
6352 against RSA and DSA as a result of this defect would be very difficult to
6353 perform and are not believed likely. Attacks against DH are considered just
6354 feasible (although very difficult) because most of the work necessary to
6355 deduce information about a private key may be performed offline. The amount
6356 of resources required for such an attack would be very significant and
6357 likely only accessible to a limited number of attackers. An attacker would
6358 additionally need online access to an unpatched system using the target
6359 private key in a scenario with persistent DH parameters and a private
6360 key that is shared between multiple clients.
6362 This only affects processors that support the BMI1, BMI2 and ADX extensions
6363 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6365 This issue was reported to OpenSSL by the OSS-Fuzz project.
6370 * Malformed X.509 IPAddressFamily could cause OOB read
6372 If an X.509 certificate has a malformed IPAddressFamily extension,
6373 OpenSSL could do a one-byte buffer overread. The most likely result
6374 would be an erroneous display of the certificate in text format.
6376 This issue was reported to OpenSSL by the OSS-Fuzz project.
6380 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6382 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6383 platform rather than 'mingw'.
6387 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6389 * Truncated packet could crash via OOB read
6391 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6392 cipher is being used, then a truncated packet can cause that host to
6393 perform an out-of-bounds read, usually resulting in a crash.
6395 This issue was reported to OpenSSL by Robert Święcki of Google.
6400 * BN_mod_exp may produce incorrect results on x86_64
6402 There is a carry propagating bug in the x86_64 Montgomery squaring
6403 procedure. No EC algorithms are affected. Analysis suggests that attacks
6404 against RSA and DSA as a result of this defect would be very difficult to
6405 perform and are not believed likely. Attacks against DH are considered just
6406 feasible (although very difficult) because most of the work necessary to
6407 deduce information about a private key may be performed offline. The amount
6408 of resources required for such an attack would be very significant and
6409 likely only accessible to a limited number of attackers. An attacker would
6410 additionally need online access to an unpatched system using the target
6411 private key in a scenario with persistent DH parameters and a private
6412 key that is shared between multiple clients. For example this can occur by
6413 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6414 similar to CVE-2015-3193 but must be treated as a separate problem.
6416 This issue was reported to OpenSSL by the OSS-Fuzz project.
6421 * Montgomery multiplication may produce incorrect results
6423 There is a carry propagating bug in the Broadwell-specific Montgomery
6424 multiplication procedure that handles input lengths divisible by, but
6425 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6426 and DH private keys are impossible. This is because the subroutine in
6427 question is not used in operations with the private key itself and an input
6428 of the attacker's direct choice. Otherwise the bug can manifest itself as
6429 transient authentication and key negotiation failures or reproducible
6430 erroneous outcome of public-key operations with specially crafted input.
6431 Among EC algorithms only Brainpool P-512 curves are affected and one
6432 presumably can attack ECDH key negotiation. Impact was not analyzed in
6433 detail, because pre-requisites for attack are considered unlikely. Namely
6434 multiple clients have to choose the curve in question and the server has to
6435 share the private key among them, neither of which is default behaviour.
6436 Even then only clients that chose the curve will be affected.
6438 This issue was publicly reported as transient failures and was not
6439 initially recognized as a security issue. Thanks to Richard Morgan for
6440 providing reproducible case.
6445 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6446 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6447 prevent issues where no progress is being made and the peer continually
6448 sends unrecognised record types, using up resources processing them.
6452 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6454 * Missing CRL sanity check
6456 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6457 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6458 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6460 This issue only affects the OpenSSL 1.0.2i
6465 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6467 * OCSP Status Request extension unbounded memory growth
6469 A malicious client can send an excessively large OCSP Status Request
6470 extension. If that client continually requests renegotiation, sending a
6471 large OCSP Status Request extension each time, then there will be unbounded
6472 memory growth on the server. This will eventually lead to a Denial Of
6473 Service attack through memory exhaustion. Servers with a default
6474 configuration are vulnerable even if they do not support OCSP. Builds using
6475 the "no-ocsp" build time option are not affected.
6477 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6482 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6485 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6491 * OOB write in MDC2_Update()
6493 An overflow can occur in MDC2_Update() either if called directly or
6494 through the EVP_DigestUpdate() function using MDC2. If an attacker
6495 is able to supply very large amounts of input data after a previous
6496 call to EVP_EncryptUpdate() with a partial block then a length check
6497 can overflow resulting in a heap corruption.
6499 The amount of data needed is comparable to SIZE_MAX which is impractical
6502 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6507 * Malformed SHA512 ticket DoS
6509 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6510 DoS attack where a malformed ticket will result in an OOB read which will
6513 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6514 a custom server callback and ticket lookup mechanism.
6516 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6521 * OOB write in BN_bn2dec()
6523 The function BN_bn2dec() does not check the return value of BN_div_word().
6524 This can cause an OOB write if an application uses this function with an
6525 overly large BIGNUM. This could be a problem if an overly large certificate
6526 or CRL is printed out from an untrusted source. TLS is not affected because
6527 record limits will reject an oversized certificate before it is parsed.
6529 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6534 * OOB read in TS_OBJ_print_bio()
6536 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6537 the total length the OID text representation would use and not the amount
6538 of data written. This will result in OOB reads when large OIDs are
6541 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6546 * Pointer arithmetic undefined behaviour
6548 Avoid some undefined pointer arithmetic
6550 A common idiom in the codebase is to check limits in the following manner:
6553 Where "p" points to some malloc'd data of SIZE bytes and
6556 "len" here could be from some externally supplied data (e.g. from a TLS
6559 The rules of C pointer arithmetic are such that "p + len" is only well
6560 defined where len <= SIZE. Therefore the above idiom is actually
6561 undefined behaviour.
6563 For example this could cause problems if some malloc implementation
6564 provides an address for "p" such that "p + len" actually overflows for
6565 values of len that are too big and therefore p + len < limit.
6567 This issue was reported to OpenSSL by Guido Vranken
6572 * Constant time flag not preserved in DSA signing
6574 Operations in the DSA signing algorithm should run in constant time in
6575 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6576 implementation means that a non-constant time codepath is followed for
6577 certain operations. This has been demonstrated through a cache-timing
6578 attack to be sufficient for an attacker to recover the private DSA key.
6580 This issue was reported by César Pereida (Aalto University), Billy Brumley
6581 (Tampere University of Technology), and Yuval Yarom (The University of
6582 Adelaide and NICTA).
6587 * DTLS buffered message DoS
6589 In a DTLS connection where handshake messages are delivered out-of-order
6590 those messages that OpenSSL is not yet ready to process will be buffered
6591 for later use. Under certain circumstances, a flaw in the logic means that
6592 those messages do not get removed from the buffer even though the handshake
6593 has been completed. An attacker could force up to approx. 15 messages to
6594 remain in the buffer when they are no longer required. These messages will
6595 be cleared when the DTLS connection is closed. The default maximum size for
6596 a message is 100k. Therefore, the attacker could force an additional 1500k
6597 to be consumed per connection. By opening many simultaneous connections an
6598 attacker could cause a DoS attack through memory exhaustion.
6600 This issue was reported to OpenSSL by Quan Luo.
6605 * DTLS replay protection DoS
6607 A flaw in the DTLS replay attack protection mechanism means that records
6608 that arrive for future epochs update the replay protection "window" before
6609 the MAC for the record has been validated. This could be exploited by an
6610 attacker by sending a record for the next epoch (which does not have to
6611 decrypt or have a valid MAC), with a very large sequence number. This means
6612 that all subsequent legitimate packets are dropped causing a denial of
6613 service for a specific DTLS connection.
6615 This issue was reported to OpenSSL by the OCAP audit team.
6620 * Certificate message OOB reads
6622 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6623 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6624 theoretical DoS risk but this has not been observed in practice on common
6627 The messages affected are client certificate, client certificate request
6628 and server certificate. As a result the attack can only be performed
6629 against a client or a server which enables client authentication.
6631 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6636 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6638 * Prevent padding oracle in AES-NI CBC MAC check
6640 A MITM attacker can use a padding oracle attack to decrypt traffic
6641 when the connection uses an AES CBC cipher and the server support
6644 This issue was introduced as part of the fix for Lucky 13 padding
6645 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6646 constant time by making sure that always the same bytes are read and
6647 compared against either the MAC or padding bytes. But it no longer
6648 checked that there was enough data to have both the MAC and padding
6651 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6655 * Fix EVP_EncodeUpdate overflow
6657 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6658 Base64 encoding of binary data. If an attacker is able to supply very large
6659 amounts of input data then a length check can overflow resulting in a heap
6662 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6663 the `PEM_write_bio*` family of functions. These are mainly used within the
6664 OpenSSL command line applications, so any application which processes data
6665 from an untrusted source and outputs it as a PEM file should be considered
6666 vulnerable to this issue. User applications that call these APIs directly
6667 with large amounts of untrusted data may also be vulnerable.
6669 This issue was reported by Guido Vranken.
6674 * Fix EVP_EncryptUpdate overflow
6676 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6677 is able to supply very large amounts of input data after a previous call to
6678 EVP_EncryptUpdate() with a partial block then a length check can overflow
6679 resulting in a heap corruption. Following an analysis of all OpenSSL
6680 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6681 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6682 the first called function after an EVP_EncryptInit(), and therefore that
6683 specific call must be safe. The second form is where the length passed to
6684 EVP_EncryptUpdate() can be seen from the code to be some small value and
6685 therefore there is no possibility of an overflow. Since all instances are
6686 one of these two forms, it is believed that there can be no overflows in
6687 internal code due to this problem. It should be noted that
6688 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6689 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6690 of these calls have also been analysed too and it is believed there are no
6691 instances in internal usage where an overflow could occur.
6693 This issue was reported by Guido Vranken.
6698 * Prevent ASN.1 BIO excessive memory allocation
6700 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6701 a short invalid encoding can cause allocation of large amounts of memory
6702 potentially consuming excessive resources or exhausting memory.
6704 Any application parsing untrusted data through d2i BIO functions is
6705 affected. The memory based functions such as d2i_X509() are *not* affected.
6706 Since the memory based functions are used by the TLS library, TLS
6707 applications are not affected.
6709 This issue was reported by Brian Carpenter.
6716 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6717 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6718 in arbitrary stack data being returned in the buffer.
6720 This issue was reported by Guido Vranken.
6725 * Modify behavior of ALPN to invoke callback after SNI/servername
6726 callback, such that updates to the SSL_CTX affect ALPN.
6730 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6735 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6736 methods are enabled and ssl2 is disabled the methods return NULL.
6740 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6742 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6743 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6744 provide any "EXPORT" or "LOW" strength ciphers.
6748 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6749 is by default disabled at build-time. Builds that are not configured with
6750 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6751 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6752 will need to explicitly call either of:
6754 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6756 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6758 as appropriate. Even if either of those is used, or the application
6759 explicitly uses the version-specific SSLv2_method() or its client and
6760 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6761 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6762 ciphers, and SSLv2 56-bit DES are no longer available.
6767 * Fix a double-free in DSA code
6769 A double free bug was discovered when OpenSSL parses malformed DSA private
6770 keys and could lead to a DoS attack or memory corruption for applications
6771 that receive DSA private keys from untrusted sources. This scenario is
6774 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6780 * Disable SRP fake user seed to address a server memory leak.
6782 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6784 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6785 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6786 was changed to ignore the "fake user" SRP seed, even if the seed
6789 Users should use SRP_VBASE_get1_by_user instead. Note that in
6790 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6791 also that even though configuring the SRP seed attempts to hide
6792 invalid usernames by continuing the handshake with fake
6793 credentials, this behaviour is not constant time and no strong
6794 guarantees are made that the handshake is indistinguishable from
6795 that of a valid user.
6800 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6802 In the BN_hex2bn function the number of hex digits is calculated using an
6803 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6804 large values of `i` this can result in `bn_expand` not allocating any
6805 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6806 field as NULL leading to a subsequent NULL ptr deref. For very large values
6807 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6808 In this case memory is allocated to the internal BIGNUM data field, but it
6809 is insufficiently sized leading to heap corruption. A similar issue exists
6810 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6811 is ever called by user applications with very large untrusted hex/dec data.
6812 This is anticipated to be a rare occurrence.
6814 All OpenSSL internal usage of these functions use data that is not expected
6815 to be untrusted, e.g. config file data or application command line
6816 arguments. If user developed applications generate config file data based
6817 on untrusted data then it is possible that this could also lead to security
6818 consequences. This is also anticipated to be rare.
6820 This issue was reported to OpenSSL by Guido Vranken.
6825 * Fix memory issues in `BIO_*printf` functions
6827 The internal `fmtstr` function used in processing a "%s" format string in
6828 the `BIO_*printf` functions could overflow while calculating the length of a
6829 string and cause an OOB read when printing very long strings.
6831 Additionally the internal `doapr_outch` function can attempt to write to an
6832 OOB memory location (at an offset from the NULL pointer) in the event of a
6833 memory allocation failure. In 1.0.2 and below this could be caused where
6834 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6835 could be in processing a very long "%s" format string. Memory leaks can
6838 The first issue may mask the second issue dependent on compiler behaviour.
6839 These problems could enable attacks where large amounts of untrusted data
6840 is passed to the `BIO_*printf` functions. If applications use these functions
6841 in this way then they could be vulnerable. OpenSSL itself uses these
6842 functions when printing out human-readable dumps of ASN.1 data. Therefore
6843 applications that print this data could be vulnerable if the data is from
6844 untrusted sources. OpenSSL command line applications could also be
6845 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6846 as command line arguments.
6848 Libssl is not considered directly vulnerable. Additionally certificates etc
6849 received via remote connections via libssl are also unlikely to be able to
6850 trigger these issues because of message size limits enforced within libssl.
6852 This issue was reported to OpenSSL Guido Vranken.
6857 * Side channel attack on modular exponentiation
6859 A side-channel attack was found which makes use of cache-bank conflicts on
6860 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6861 of RSA keys. The ability to exploit this issue is limited as it relies on
6862 an attacker who has control of code in a thread running on the same
6863 hyper-threaded core as the victim thread which is performing decryptions.
6865 This issue was reported to OpenSSL by Yuval Yarom, The University of
6866 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6867 Nadia Heninger, University of Pennsylvania with more information at
6868 <http://cachebleed.info>.
6873 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6874 if no keysize is specified with default_bits. This fixes an
6875 omission in an earlier change that changed all RSA/DSA key generation
6876 commands to use 2048 bits by default.
6880 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6882 * DH small subgroups
6884 Historically OpenSSL only ever generated DH parameters based on "safe"
6885 primes. More recently (in version 1.0.2) support was provided for
6886 generating X9.42 style parameter files such as those required for RFC 5114
6887 support. The primes used in such files may not be "safe". Where an
6888 application is using DH configured with parameters based on primes that are
6889 not "safe" then an attacker could use this fact to find a peer's private
6890 DH exponent. This attack requires that the attacker complete multiple
6891 handshakes in which the peer uses the same private DH exponent. For example
6892 this could be used to discover a TLS server's private DH exponent if it's
6893 reusing the private DH exponent or it's using a static DH ciphersuite.
6895 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6896 TLS. It is not on by default. If the option is not set then the server
6897 reuses the same private DH exponent for the life of the server process and
6898 would be vulnerable to this attack. It is believed that many popular
6899 applications do set this option and would therefore not be at risk.
6901 The fix for this issue adds an additional check where a "q" parameter is
6902 available (as is the case in X9.42 based parameters). This detects the
6903 only known attack, and is the only possible defense for static DH
6904 ciphersuites. This could have some performance impact.
6906 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6907 default and cannot be disabled. This could have some performance impact.
6909 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6914 * SSLv2 doesn't block disabled ciphers
6916 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6917 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6918 been disabled, provided that the SSLv2 protocol was not also disabled via
6921 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6922 and Sebastian Schinzel.
6927 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6929 * BN_mod_exp may produce incorrect results on x86_64
6931 There is a carry propagating bug in the x86_64 Montgomery squaring
6932 procedure. No EC algorithms are affected. Analysis suggests that attacks
6933 against RSA and DSA as a result of this defect would be very difficult to
6934 perform and are not believed likely. Attacks against DH are considered just
6935 feasible (although very difficult) because most of the work necessary to
6936 deduce information about a private key may be performed offline. The amount
6937 of resources required for such an attack would be very significant and
6938 likely only accessible to a limited number of attackers. An attacker would
6939 additionally need online access to an unpatched system using the target
6940 private key in a scenario with persistent DH parameters and a private
6941 key that is shared between multiple clients. For example this can occur by
6942 default in OpenSSL DHE based SSL/TLS ciphersuites.
6944 This issue was reported to OpenSSL by Hanno Böck.
6949 * Certificate verify crash with missing PSS parameter
6951 The signature verification routines will crash with a NULL pointer
6952 dereference if presented with an ASN.1 signature using the RSA PSS
6953 algorithm and absent mask generation function parameter. Since these
6954 routines are used to verify certificate signature algorithms this can be
6955 used to crash any certificate verification operation and exploited in a
6956 DoS attack. Any application which performs certificate verification is
6957 vulnerable including OpenSSL clients and servers which enable client
6960 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6965 * X509_ATTRIBUTE memory leak
6967 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6968 memory. This structure is used by the PKCS#7 and CMS routines so any
6969 application which reads PKCS#7 or CMS data from untrusted sources is
6970 affected. SSL/TLS is not affected.
6972 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6978 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6979 This changes the decoding behaviour for some invalid messages,
6980 though the change is mostly in the more lenient direction, and
6981 legacy behaviour is preserved as much as possible.
6985 * In DSA_generate_parameters_ex, if the provided seed is too short,
6988 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6990 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6992 * Alternate chains certificate forgery
6994 During certificate verification, OpenSSL will attempt to find an
6995 alternative certificate chain if the first attempt to build such a chain
6996 fails. An error in the implementation of this logic can mean that an
6997 attacker could cause certain checks on untrusted certificates to be
6998 bypassed, such as the CA flag, enabling them to use a valid leaf
6999 certificate to act as a CA and "issue" an invalid certificate.
7001 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7006 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
7008 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7009 incompatibility in the handling of HMAC. The previous ABI has now been
7014 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
7016 * Malformed ECParameters causes infinite loop
7018 When processing an ECParameters structure OpenSSL enters an infinite loop
7019 if the curve specified is over a specially malformed binary polynomial
7022 This can be used to perform denial of service against any
7023 system which processes public keys, certificate requests or
7024 certificates. This includes TLS clients and TLS servers with
7025 client authentication enabled.
7027 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7032 * Exploitable out-of-bounds read in X509_cmp_time
7034 X509_cmp_time does not properly check the length of the ASN1_TIME
7035 string and can read a few bytes out of bounds. In addition,
7036 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7039 An attacker can use this to craft malformed certificates and CRLs of
7040 various sizes and potentially cause a segmentation fault, resulting in
7041 a DoS on applications that verify certificates or CRLs. TLS clients
7042 that verify CRLs are affected. TLS clients and servers with client
7043 authentication enabled may be affected if they use custom verification
7046 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7047 independently by Hanno Böck.
7052 * PKCS7 crash with missing EnvelopedContent
7054 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7055 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7056 with missing content and trigger a NULL pointer dereference on parsing.
7058 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7059 structures from untrusted sources are affected. OpenSSL clients and
7060 servers are not affected.
7062 This issue was reported to OpenSSL by Michal Zalewski (Google).
7067 * CMS verify infinite loop with unknown hash function
7069 When verifying a signedData message the CMS code can enter an infinite loop
7070 if presented with an unknown hash function OID. This can be used to perform
7071 denial of service against any system which verifies signedData messages using
7073 This issue was reported to OpenSSL by Johannes Bauer.
7078 * Race condition handling NewSessionTicket
7080 If a NewSessionTicket is received by a multi-threaded client when attempting to
7081 reuse a previous ticket then a race condition can occur potentially leading to
7082 a double free of the ticket data.
7087 * Only support 256-bit or stronger elliptic curves with the
7088 'ecdh_auto' setting (server) or by default (client). Of supported
7089 curves, prefer P-256 (both).
7093 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
7095 * ClientHello sigalgs DoS fix
7097 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
7098 invalid signature algorithms extension a NULL pointer dereference will
7099 occur. This can be exploited in a DoS attack against the server.
7101 This issue was was reported to OpenSSL by David Ramos of Stanford
7105 *Stephen Henson and Matt Caswell*
7107 * Multiblock corrupted pointer fix
7109 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
7110 feature only applies on 64 bit x86 architecture platforms that support AES
7111 NI instructions. A defect in the implementation of "multiblock" can cause
7112 OpenSSL's internal write buffer to become incorrectly set to NULL when
7113 using non-blocking IO. Typically, when the user application is using a
7114 socket BIO for writing, this will only result in a failed connection.
7115 However if some other BIO is used then it is likely that a segmentation
7116 fault will be triggered, thus enabling a potential DoS attack.
7118 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
7123 * Segmentation fault in DTLSv1_listen fix
7125 The DTLSv1_listen function is intended to be stateless and processes the
7126 initial ClientHello from many peers. It is common for user code to loop
7127 over the call to DTLSv1_listen until a valid ClientHello is received with
7128 an associated cookie. A defect in the implementation of DTLSv1_listen means
7129 that state is preserved in the SSL object from one invocation to the next
7130 that can lead to a segmentation fault. Errors processing the initial
7131 ClientHello can trigger this scenario. An example of such an error could be
7132 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
7135 This issue was reported to OpenSSL by Per Allansson.
7140 * Segmentation fault in ASN1_TYPE_cmp fix
7142 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7143 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7144 certificate signature algorithm consistency this can be used to crash any
7145 certificate verification operation and exploited in a DoS attack. Any
7146 application which performs certificate verification is vulnerable including
7147 OpenSSL clients and servers which enable client authentication.
7152 * Segmentation fault for invalid PSS parameters fix
7154 The signature verification routines will crash with a NULL pointer
7155 dereference if presented with an ASN.1 signature using the RSA PSS
7156 algorithm and invalid parameters. Since these routines are used to verify
7157 certificate signature algorithms this can be used to crash any
7158 certificate verification operation and exploited in a DoS attack. Any
7159 application which performs certificate verification is vulnerable including
7160 OpenSSL clients and servers which enable client authentication.
7162 This issue was was reported to OpenSSL by Brian Carpenter.
7167 * ASN.1 structure reuse memory corruption fix
7169 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7170 memory corruption via an invalid write. Such reuse is and has been
7171 strongly discouraged and is believed to be rare.
7173 Applications that parse structures containing CHOICE or ANY DEFINED BY
7174 components may be affected. Certificate parsing (d2i_X509 and related
7175 functions) are however not affected. OpenSSL clients and servers are
7181 * PKCS7 NULL pointer dereferences fix
7183 The PKCS#7 parsing code does not handle missing outer ContentInfo
7184 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7185 missing content and trigger a NULL pointer dereference on parsing.
7187 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7188 otherwise parse PKCS#7 structures from untrusted sources are
7189 affected. OpenSSL clients and servers are not affected.
7191 This issue was reported to OpenSSL by Michal Zalewski (Google).
7196 * DoS via reachable assert in SSLv2 servers fix
7198 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7199 servers that both support SSLv2 and enable export cipher suites by sending
7200 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7202 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7203 (OpenSSL development team).
7208 * Empty CKE with client auth and DHE fix
7210 If client auth is used then a server can seg fault in the event of a DHE
7211 ciphersuite being selected and a zero length ClientKeyExchange message
7212 being sent by the client. This could be exploited in a DoS attack.
7217 * Handshake with unseeded PRNG fix
7219 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
7220 with an unseeded PRNG. The conditions are:
7221 - The client is on a platform where the PRNG has not been seeded
7222 automatically, and the user has not seeded manually
7223 - A protocol specific client method version has been used (i.e. not
7224 SSL_client_methodv23)
7225 - A ciphersuite is used that does not require additional random data from
7226 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
7228 If the handshake succeeds then the client random that has been used will
7229 have been generated from a PRNG with insufficient entropy and therefore the
7230 output may be predictable.
7232 For example using the following command with an unseeded openssl will
7233 succeed on an unpatched platform:
7235 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
7240 * Use After Free following d2i_ECPrivatekey error fix
7242 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7243 could cause a use after free condition. This, in turn, could cause a double
7244 free in several private key parsing functions (such as d2i_PrivateKey
7245 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7246 for applications that receive EC private keys from untrusted
7247 sources. This scenario is considered rare.
7249 This issue was discovered by the BoringSSL project and fixed in their
7255 * X509_to_X509_REQ NULL pointer deref fix
7257 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7258 the certificate key is invalid. This function is rarely used in practice.
7260 This issue was discovered by Brian Carpenter.
7265 * Removed the export ciphers from the DEFAULT ciphers
7269 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7271 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7272 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7273 So far those who have to target multiple platforms would compromise
7274 and argue that binary targeting say ARMv5 would still execute on
7275 ARMv8. "Universal" build resolves this compromise by providing
7276 near-optimal performance even on newer platforms.
7280 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7281 (other platforms pending).
7283 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7285 * Add support for the SignedCertificateTimestampList certificate and
7286 OCSP response extensions from RFC6962.
7290 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7291 for corner cases. (Certain input points at infinity could lead to
7292 bogus results, with non-infinity inputs mapped to infinity too.)
7296 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7297 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7298 common cases are optimized and there still is room for further
7299 improvements. Vector Permutation AES for Altivec is also added.
7303 * Add support for little-endian ppc64 Linux target.
7305 *Marcelo Cerri (IBM)*
7307 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7308 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7309 are optimized and there still is room for further improvements.
7310 Both 32- and 64-bit modes are supported.
7312 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7314 * Improved ARMv7 NEON support.
7318 * Support for SPARC Architecture 2011 crypto extensions, first
7319 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7320 SHA256/512, MD5, GHASH and modular exponentiation.
7322 *Andy Polyakov, David Miller*
7324 * Accelerated modular exponentiation for Intel processors, a.k.a.
7327 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7329 * Support for new and upcoming Intel processors, including AVX2,
7330 BMI and SHA ISA extensions. This includes additional "stitched"
7331 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7334 This work was sponsored by Intel Corp.
7338 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7339 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7340 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7344 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7345 this fixes a limitation in previous versions of OpenSSL.
7349 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7350 MGF1 digest and OAEP label.
7354 * Add EVP support for key wrapping algorithms, to avoid problems with
7355 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7356 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7357 algorithms and include tests cases.
7361 * Add functions to allocate and set the fields of an ECDSA_METHOD
7364 *Douglas E. Engert, Steve Henson*
7366 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7367 difference in days and seconds between two tm or ASN1_TIME structures.
7371 * Add -rev test option to s_server to just reverse order of characters
7372 received by client and send back to server. Also prints an abbreviated
7373 summary of the connection parameters.
7377 * New option -brief for s_client and s_server to print out a brief summary
7378 of connection parameters.
7382 * Add callbacks for arbitrary TLS extensions.
7384 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7386 * New option -crl_download in several openssl utilities to download CRLs
7387 from CRLDP extension in certificates.
7391 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7395 * New function X509_CRL_diff to generate a delta CRL from the difference
7396 of two full CRLs. Add support to "crl" utility.
7400 * New functions to set lookup_crls function and to retrieve
7401 X509_STORE from X509_STORE_CTX.
7405 * Print out deprecated issuer and subject unique ID fields in
7410 * Extend OCSP I/O functions so they can be used for simple general purpose
7411 HTTP as well as OCSP. New wrapper function which can be used to download
7412 CRLs using the OCSP API.
7416 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7420 * `SSL_CONF*` functions. These provide a common framework for application
7421 configuration using configuration files or command lines.
7425 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7426 message callback and prints the results. Needs compile time option
7427 "enable-ssl-trace". New options to s_client and s_server to enable
7432 * New ctrl and macro to retrieve supported points extensions.
7433 Print out extension in s_server and s_client.
7437 * New functions to retrieve certificate signature and signature
7442 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7447 * New Suite B modes for TLS code. These use and enforce the requirements
7448 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7449 only use Suite B curves. The Suite B modes can be set by using the
7450 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7454 * New chain verification flags for Suite B levels of security. Check
7455 algorithms are acceptable when flags are set in X509_verify_cert.
7459 * Make tls1_check_chain return a set of flags indicating checks passed
7460 by a certificate chain. Add additional tests to handle client
7461 certificates: checks for matching certificate type and issuer name
7466 * If an attempt is made to use a signature algorithm not in the peer
7467 preference list abort the handshake. If client has no suitable
7468 signature algorithms in response to a certificate request do not
7469 use the certificate.
7473 * If server EC tmp key is not in client preference list abort handshake.
7477 * Add support for certificate stores in CERT structure. This makes it
7478 possible to have different stores per SSL structure or one store in
7479 the parent SSL_CTX. Include distinct stores for certificate chain
7480 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7481 to build and store a certificate chain in CERT structure: returning
7482 an error if the chain cannot be built: this will allow applications
7483 to test if a chain is correctly configured.
7485 Note: if the CERT based stores are not set then the parent SSL_CTX
7486 store is used to retain compatibility with existing behaviour.
7490 * New function ssl_set_client_disabled to set a ciphersuite disabled
7491 mask based on the current session, check mask when sending client
7492 hello and checking the requested ciphersuite.
7496 * New ctrls to retrieve and set certificate types in a certificate
7497 request message. Print out received values in s_client. If certificate
7498 types is not set with custom values set sensible values based on
7499 supported signature algorithms.
7503 * Support for distinct client and server supported signature algorithms.
7507 * Add certificate callback. If set this is called whenever a certificate
7508 is required by client or server. An application can decide which
7509 certificate chain to present based on arbitrary criteria: for example
7510 supported signature algorithms. Add very simple example to s_server.
7511 This fixes many of the problems and restrictions of the existing client
7512 certificate callback: for example you can now clear an existing
7513 certificate and specify the whole chain.
7517 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7518 the certificate can be used for (if anything). Set valid_flags field
7519 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7520 to have similar checks in it.
7522 Add new "cert_flags" field to CERT structure and include a "strict mode".
7523 This enforces some TLS certificate requirements (such as only permitting
7524 certificate signature algorithms contained in the supported algorithms
7525 extension) which some implementations ignore: this option should be used
7526 with caution as it could cause interoperability issues.
7530 * Update and tidy signature algorithm extension processing. Work out
7531 shared signature algorithms based on preferences and peer algorithms
7532 and print them out in s_client and s_server. Abort handshake if no
7533 shared signature algorithms.
7537 * Add new functions to allow customised supported signature algorithms
7538 for SSL and SSL_CTX structures. Add options to s_client and s_server
7543 * New function SSL_certs_clear() to delete all references to certificates
7544 from an SSL structure. Before this once a certificate had been added
7545 it couldn't be removed.
7549 * Integrate hostname, email address and IP address checking with certificate
7550 verification. New verify options supporting checking in openssl utility.
7554 * Fixes and wildcard matching support to hostname and email checking
7555 functions. Add manual page.
7557 *Florian Weimer (Red Hat Product Security Team)*
7559 * New functions to check a hostname email or IP address against a
7560 certificate. Add options x509 utility to print results of checks against
7565 * Fix OCSP checking.
7567 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7569 * Initial experimental support for explicitly trusted non-root CAs.
7570 OpenSSL still tries to build a complete chain to a root but if an
7571 intermediate CA has a trust setting included that is used. The first
7572 setting is used: whether to trust (e.g., -addtrust option to the x509
7577 * Add -trusted_first option which attempts to find certificates in the
7578 trusted store even if an untrusted chain is also supplied.
7582 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7583 platform support for Linux and Android.
7587 * Support for linux-x32, ILP32 environment in x86_64 framework.
7591 * Experimental multi-implementation support for FIPS capable OpenSSL.
7592 When in FIPS mode the approved implementations are used as normal,
7593 when not in FIPS mode the internal unapproved versions are used instead.
7594 This means that the FIPS capable OpenSSL isn't forced to use the
7595 (often lower performance) FIPS implementations outside FIPS mode.
7599 * Transparently support X9.42 DH parameters when calling
7600 PEM_read_bio_DHparameters. This means existing applications can handle
7601 the new parameter format automatically.
7605 * Initial experimental support for X9.42 DH parameter format: mainly
7606 to support use of 'q' parameter for RFC5114 parameters.
7610 * Add DH parameters from RFC5114 including test data to dhtest.
7614 * Support for automatic EC temporary key parameter selection. If enabled
7615 the most preferred EC parameters are automatically used instead of
7616 hardcoded fixed parameters. Now a server just has to call:
7617 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7618 support ECDH and use the most appropriate parameters.
7622 * Enhance and tidy EC curve and point format TLS extension code. Use
7623 static structures instead of allocation if default values are used.
7624 New ctrls to set curves we wish to support and to retrieve shared curves.
7625 Print out shared curves in s_server. New options to s_server and s_client
7626 to set list of supported curves.
7630 * New ctrls to retrieve supported signature algorithms and
7631 supported curve values as an array of NIDs. Extend openssl utility
7632 to print out received values.
7636 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7637 between NIDs and the more common NIST names such as "P-256". Enhance
7638 ecparam utility and ECC method to recognise the NIST names for curves.
7642 * Enhance SSL/TLS certificate chain handling to support different
7643 chains for each certificate instead of one chain in the parent SSL_CTX.
7647 * Support for fixed DH ciphersuite client authentication: where both
7648 server and client use DH certificates with common parameters.
7652 * Support for fixed DH ciphersuites: those requiring DH server
7657 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7659 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7660 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7661 X509_CINF_get_signature were reverted post internal team review.
7666 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7668 * OCSP Status Request extension unbounded memory growth
7670 A malicious client can send an excessively large OCSP Status Request
7671 extension. If that client continually requests renegotiation, sending a
7672 large OCSP Status Request extension each time, then there will be unbounded
7673 memory growth on the server. This will eventually lead to a Denial Of
7674 Service attack through memory exhaustion. Servers with a default
7675 configuration are vulnerable even if they do not support OCSP. Builds using
7676 the "no-ocsp" build time option are not affected.
7678 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7683 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7686 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7692 * OOB write in MDC2_Update()
7694 An overflow can occur in MDC2_Update() either if called directly or
7695 through the EVP_DigestUpdate() function using MDC2. If an attacker
7696 is able to supply very large amounts of input data after a previous
7697 call to EVP_EncryptUpdate() with a partial block then a length check
7698 can overflow resulting in a heap corruption.
7700 The amount of data needed is comparable to SIZE_MAX which is impractical
7703 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7708 * Malformed SHA512 ticket DoS
7710 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7711 DoS attack where a malformed ticket will result in an OOB read which will
7714 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7715 a custom server callback and ticket lookup mechanism.
7717 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7722 * OOB write in BN_bn2dec()
7724 The function BN_bn2dec() does not check the return value of BN_div_word().
7725 This can cause an OOB write if an application uses this function with an
7726 overly large BIGNUM. This could be a problem if an overly large certificate
7727 or CRL is printed out from an untrusted source. TLS is not affected because
7728 record limits will reject an oversized certificate before it is parsed.
7730 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7735 * OOB read in TS_OBJ_print_bio()
7737 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7738 the total length the OID text representation would use and not the amount
7739 of data written. This will result in OOB reads when large OIDs are
7742 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7747 * Pointer arithmetic undefined behaviour
7749 Avoid some undefined pointer arithmetic
7751 A common idiom in the codebase is to check limits in the following manner:
7754 Where "p" points to some malloc'd data of SIZE bytes and
7757 "len" here could be from some externally supplied data (e.g. from a TLS
7760 The rules of C pointer arithmetic are such that "p + len" is only well
7761 defined where len <= SIZE. Therefore, the above idiom is actually
7762 undefined behaviour.
7764 For example this could cause problems if some malloc implementation
7765 provides an address for "p" such that "p + len" actually overflows for
7766 values of len that are too big and therefore p + len < limit.
7768 This issue was reported to OpenSSL by Guido Vranken
7773 * Constant time flag not preserved in DSA signing
7775 Operations in the DSA signing algorithm should run in constant time in
7776 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7777 implementation means that a non-constant time codepath is followed for
7778 certain operations. This has been demonstrated through a cache-timing
7779 attack to be sufficient for an attacker to recover the private DSA key.
7781 This issue was reported by César Pereida (Aalto University), Billy Brumley
7782 (Tampere University of Technology), and Yuval Yarom (The University of
7783 Adelaide and NICTA).
7788 * DTLS buffered message DoS
7790 In a DTLS connection where handshake messages are delivered out-of-order
7791 those messages that OpenSSL is not yet ready to process will be buffered
7792 for later use. Under certain circumstances, a flaw in the logic means that
7793 those messages do not get removed from the buffer even though the handshake
7794 has been completed. An attacker could force up to approx. 15 messages to
7795 remain in the buffer when they are no longer required. These messages will
7796 be cleared when the DTLS connection is closed. The default maximum size for
7797 a message is 100k. Therefore, the attacker could force an additional 1500k
7798 to be consumed per connection. By opening many simultaneous connections an
7799 attacker could cause a DoS attack through memory exhaustion.
7801 This issue was reported to OpenSSL by Quan Luo.
7806 * DTLS replay protection DoS
7808 A flaw in the DTLS replay attack protection mechanism means that records
7809 that arrive for future epochs update the replay protection "window" before
7810 the MAC for the record has been validated. This could be exploited by an
7811 attacker by sending a record for the next epoch (which does not have to
7812 decrypt or have a valid MAC), with a very large sequence number. This means
7813 that all subsequent legitimate packets are dropped causing a denial of
7814 service for a specific DTLS connection.
7816 This issue was reported to OpenSSL by the OCAP audit team.
7821 * Certificate message OOB reads
7823 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7824 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7825 theoretical DoS risk but this has not been observed in practice on common
7828 The messages affected are client certificate, client certificate request
7829 and server certificate. As a result the attack can only be performed
7830 against a client or a server which enables client authentication.
7832 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7837 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7839 * Prevent padding oracle in AES-NI CBC MAC check
7841 A MITM attacker can use a padding oracle attack to decrypt traffic
7842 when the connection uses an AES CBC cipher and the server support
7845 This issue was introduced as part of the fix for Lucky 13 padding
7846 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7847 constant time by making sure that always the same bytes are read and
7848 compared against either the MAC or padding bytes. But it no longer
7849 checked that there was enough data to have both the MAC and padding
7852 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7857 * Fix EVP_EncodeUpdate overflow
7859 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7860 Base64 encoding of binary data. If an attacker is able to supply very large
7861 amounts of input data then a length check can overflow resulting in a heap
7864 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7865 the `PEM_write_bio*` family of functions. These are mainly used within the
7866 OpenSSL command line applications, so any application which processes data
7867 from an untrusted source and outputs it as a PEM file should be considered
7868 vulnerable to this issue. User applications that call these APIs directly
7869 with large amounts of untrusted data may also be vulnerable.
7871 This issue was reported by Guido Vranken.
7876 * Fix EVP_EncryptUpdate overflow
7878 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7879 is able to supply very large amounts of input data after a previous call to
7880 EVP_EncryptUpdate() with a partial block then a length check can overflow
7881 resulting in a heap corruption. Following an analysis of all OpenSSL
7882 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7883 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7884 the first called function after an EVP_EncryptInit(), and therefore that
7885 specific call must be safe. The second form is where the length passed to
7886 EVP_EncryptUpdate() can be seen from the code to be some small value and
7887 therefore there is no possibility of an overflow. Since all instances are
7888 one of these two forms, it is believed that there can be no overflows in
7889 internal code due to this problem. It should be noted that
7890 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7891 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7892 of these calls have also been analysed too and it is believed there are no
7893 instances in internal usage where an overflow could occur.
7895 This issue was reported by Guido Vranken.
7900 * Prevent ASN.1 BIO excessive memory allocation
7902 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7903 a short invalid encoding can casuse allocation of large amounts of memory
7904 potentially consuming excessive resources or exhausting memory.
7906 Any application parsing untrusted data through d2i BIO functions is
7907 affected. The memory based functions such as d2i_X509() are *not* affected.
7908 Since the memory based functions are used by the TLS library, TLS
7909 applications are not affected.
7911 This issue was reported by Brian Carpenter.
7918 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7919 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7920 in arbitrary stack data being returned in the buffer.
7922 This issue was reported by Guido Vranken.
7927 * Modify behavior of ALPN to invoke callback after SNI/servername
7928 callback, such that updates to the SSL_CTX affect ALPN.
7932 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7937 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7938 methods are enabled and ssl2 is disabled the methods return NULL.
7942 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7944 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7945 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7946 provide any "EXPORT" or "LOW" strength ciphers.
7950 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7951 is by default disabled at build-time. Builds that are not configured with
7952 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7953 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7954 will need to explicitly call either of:
7956 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7958 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7960 as appropriate. Even if either of those is used, or the application
7961 explicitly uses the version-specific SSLv2_method() or its client and
7962 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7963 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7964 ciphers, and SSLv2 56-bit DES are no longer available.
7969 * Fix a double-free in DSA code
7971 A double free bug was discovered when OpenSSL parses malformed DSA private
7972 keys and could lead to a DoS attack or memory corruption for applications
7973 that receive DSA private keys from untrusted sources. This scenario is
7976 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7982 * Disable SRP fake user seed to address a server memory leak.
7984 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7986 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7987 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7988 was changed to ignore the "fake user" SRP seed, even if the seed
7991 Users should use SRP_VBASE_get1_by_user instead. Note that in
7992 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7993 also that even though configuring the SRP seed attempts to hide
7994 invalid usernames by continuing the handshake with fake
7995 credentials, this behaviour is not constant time and no strong
7996 guarantees are made that the handshake is indistinguishable from
7997 that of a valid user.
8002 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
8004 In the BN_hex2bn function the number of hex digits is calculated using an
8005 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
8006 large values of `i` this can result in `bn_expand` not allocating any
8007 memory because `i * 4` is negative. This can leave the internal BIGNUM data
8008 field as NULL leading to a subsequent NULL ptr deref. For very large values
8009 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
8010 In this case memory is allocated to the internal BIGNUM data field, but it
8011 is insufficiently sized leading to heap corruption. A similar issue exists
8012 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
8013 is ever called by user applications with very large untrusted hex/dec data.
8014 This is anticipated to be a rare occurrence.
8016 All OpenSSL internal usage of these functions use data that is not expected
8017 to be untrusted, e.g. config file data or application command line
8018 arguments. If user developed applications generate config file data based
8019 on untrusted data then it is possible that this could also lead to security
8020 consequences. This is also anticipated to be rare.
8022 This issue was reported to OpenSSL by Guido Vranken.
8027 * Fix memory issues in `BIO_*printf` functions
8029 The internal `fmtstr` function used in processing a "%s" format string in
8030 the `BIO_*printf` functions could overflow while calculating the length of a
8031 string and cause an OOB read when printing very long strings.
8033 Additionally the internal `doapr_outch` function can attempt to write to an
8034 OOB memory location (at an offset from the NULL pointer) in the event of a
8035 memory allocation failure. In 1.0.2 and below this could be caused where
8036 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
8037 could be in processing a very long "%s" format string. Memory leaks can
8040 The first issue may mask the second issue dependent on compiler behaviour.
8041 These problems could enable attacks where large amounts of untrusted data
8042 is passed to the `BIO_*printf` functions. If applications use these functions
8043 in this way then they could be vulnerable. OpenSSL itself uses these
8044 functions when printing out human-readable dumps of ASN.1 data. Therefore
8045 applications that print this data could be vulnerable if the data is from
8046 untrusted sources. OpenSSL command line applications could also be
8047 vulnerable where they print out ASN.1 data, or if untrusted data is passed
8048 as command line arguments.
8050 Libssl is not considered directly vulnerable. Additionally certificates etc
8051 received via remote connections via libssl are also unlikely to be able to
8052 trigger these issues because of message size limits enforced within libssl.
8054 This issue was reported to OpenSSL Guido Vranken.
8059 * Side channel attack on modular exponentiation
8061 A side-channel attack was found which makes use of cache-bank conflicts on
8062 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
8063 of RSA keys. The ability to exploit this issue is limited as it relies on
8064 an attacker who has control of code in a thread running on the same
8065 hyper-threaded core as the victim thread which is performing decryptions.
8067 This issue was reported to OpenSSL by Yuval Yarom, The University of
8068 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
8069 Nadia Heninger, University of Pennsylvania with more information at
8070 <http://cachebleed.info>.
8075 * Change the req command to generate a 2048-bit RSA/DSA key by default,
8076 if no keysize is specified with default_bits. This fixes an
8077 omission in an earlier change that changed all RSA/DSA key generation
8078 commands to use 2048 bits by default.
8082 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
8084 * Protection for DH small subgroup attacks
8086 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
8087 switched on by default and cannot be disabled. This could have some
8092 * SSLv2 doesn't block disabled ciphers
8094 A malicious client can negotiate SSLv2 ciphers that have been disabled on
8095 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
8096 been disabled, provided that the SSLv2 protocol was not also disabled via
8099 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
8100 and Sebastian Schinzel.
8105 * Reject DH handshakes with parameters shorter than 1024 bits.
8109 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
8111 * Certificate verify crash with missing PSS parameter
8113 The signature verification routines will crash with a NULL pointer
8114 dereference if presented with an ASN.1 signature using the RSA PSS
8115 algorithm and absent mask generation function parameter. Since these
8116 routines are used to verify certificate signature algorithms this can be
8117 used to crash any certificate verification operation and exploited in a
8118 DoS attack. Any application which performs certificate verification is
8119 vulnerable including OpenSSL clients and servers which enable client
8122 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
8127 * X509_ATTRIBUTE memory leak
8129 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8130 memory. This structure is used by the PKCS#7 and CMS routines so any
8131 application which reads PKCS#7 or CMS data from untrusted sources is
8132 affected. SSL/TLS is not affected.
8134 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8140 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
8141 This changes the decoding behaviour for some invalid messages,
8142 though the change is mostly in the more lenient direction, and
8143 legacy behaviour is preserved as much as possible.
8147 * In DSA_generate_parameters_ex, if the provided seed is too short,
8148 use a random seed, as already documented.
8150 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
8152 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
8154 * Alternate chains certificate forgery
8156 During certificate verification, OpenSSL will attempt to find an
8157 alternative certificate chain if the first attempt to build such a chain
8158 fails. An error in the implementation of this logic can mean that an
8159 attacker could cause certain checks on untrusted certificates to be
8160 bypassed, such as the CA flag, enabling them to use a valid leaf
8161 certificate to act as a CA and "issue" an invalid certificate.
8163 This issue was reported to OpenSSL by Adam Langley/David Benjamin
8169 * Race condition handling PSK identify hint
8171 If PSK identity hints are received by a multi-threaded client then
8172 the values are wrongly updated in the parent SSL_CTX structure. This can
8173 result in a race condition potentially leading to a double free of the
8179 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
8181 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
8182 incompatibility in the handling of HMAC. The previous ABI has now been
8185 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
8187 * Malformed ECParameters causes infinite loop
8189 When processing an ECParameters structure OpenSSL enters an infinite loop
8190 if the curve specified is over a specially malformed binary polynomial
8193 This can be used to perform denial of service against any
8194 system which processes public keys, certificate requests or
8195 certificates. This includes TLS clients and TLS servers with
8196 client authentication enabled.
8198 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8203 * Exploitable out-of-bounds read in X509_cmp_time
8205 X509_cmp_time does not properly check the length of the ASN1_TIME
8206 string and can read a few bytes out of bounds. In addition,
8207 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8210 An attacker can use this to craft malformed certificates and CRLs of
8211 various sizes and potentially cause a segmentation fault, resulting in
8212 a DoS on applications that verify certificates or CRLs. TLS clients
8213 that verify CRLs are affected. TLS clients and servers with client
8214 authentication enabled may be affected if they use custom verification
8217 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8218 independently by Hanno Böck.
8223 * PKCS7 crash with missing EnvelopedContent
8225 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8226 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8227 with missing content and trigger a NULL pointer dereference on parsing.
8229 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8230 structures from untrusted sources are affected. OpenSSL clients and
8231 servers are not affected.
8233 This issue was reported to OpenSSL by Michal Zalewski (Google).
8238 * CMS verify infinite loop with unknown hash function
8240 When verifying a signedData message the CMS code can enter an infinite loop
8241 if presented with an unknown hash function OID. This can be used to perform
8242 denial of service against any system which verifies signedData messages using
8244 This issue was reported to OpenSSL by Johannes Bauer.
8249 * Race condition handling NewSessionTicket
8251 If a NewSessionTicket is received by a multi-threaded client when attempting to
8252 reuse a previous ticket then a race condition can occur potentially leading to
8253 a double free of the ticket data.
8258 * Reject DH handshakes with parameters shorter than 768 bits.
8260 *Kurt Roeckx and Emilia Kasper*
8262 * dhparam: generate 2048-bit parameters by default.
8264 *Kurt Roeckx and Emilia Kasper*
8266 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8268 * Segmentation fault in ASN1_TYPE_cmp fix
8270 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8271 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8272 certificate signature algorithm consistency this can be used to crash any
8273 certificate verification operation and exploited in a DoS attack. Any
8274 application which performs certificate verification is vulnerable including
8275 OpenSSL clients and servers which enable client authentication.
8280 * ASN.1 structure reuse memory corruption fix
8282 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8283 memory corruption via an invalid write. Such reuse is and has been
8284 strongly discouraged and is believed to be rare.
8286 Applications that parse structures containing CHOICE or ANY DEFINED BY
8287 components may be affected. Certificate parsing (d2i_X509 and related
8288 functions) are however not affected. OpenSSL clients and servers are
8294 * PKCS7 NULL pointer dereferences fix
8296 The PKCS#7 parsing code does not handle missing outer ContentInfo
8297 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8298 missing content and trigger a NULL pointer dereference on parsing.
8300 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8301 otherwise parse PKCS#7 structures from untrusted sources are
8302 affected. OpenSSL clients and servers are not affected.
8304 This issue was reported to OpenSSL by Michal Zalewski (Google).
8309 * DoS via reachable assert in SSLv2 servers fix
8311 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8312 servers that both support SSLv2 and enable export cipher suites by sending
8313 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8315 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8316 (OpenSSL development team).
8321 * Use After Free following d2i_ECPrivatekey error fix
8323 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8324 could cause a use after free condition. This, in turn, could cause a double
8325 free in several private key parsing functions (such as d2i_PrivateKey
8326 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8327 for applications that receive EC private keys from untrusted
8328 sources. This scenario is considered rare.
8330 This issue was discovered by the BoringSSL project and fixed in their
8336 * X509_to_X509_REQ NULL pointer deref fix
8338 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8339 the certificate key is invalid. This function is rarely used in practice.
8341 This issue was discovered by Brian Carpenter.
8346 * Removed the export ciphers from the DEFAULT ciphers
8350 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8352 * Build fixes for the Windows and OpenVMS platforms
8354 *Matt Caswell and Richard Levitte*
8356 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8358 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8359 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8360 dereference. This could lead to a Denial Of Service attack. Thanks to
8361 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8366 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8367 dtls1_buffer_record function under certain conditions. In particular this
8368 could occur if an attacker sent repeated DTLS records with the same
8369 sequence number but for the next epoch. The memory leak could be exploited
8370 by an attacker in a Denial of Service attack through memory exhaustion.
8371 Thanks to Chris Mueller for reporting this issue.
8376 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8377 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8378 method would be set to NULL which could later result in a NULL pointer
8379 dereference. Thanks to Frank Schmirler for reporting this issue.
8384 * Abort handshake if server key exchange message is omitted for ephemeral
8387 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8388 reporting this issue.
8393 * Remove non-export ephemeral RSA code on client and server. This code
8394 violated the TLS standard by allowing the use of temporary RSA keys in
8395 non-export ciphersuites and could be used by a server to effectively
8396 downgrade the RSA key length used to a value smaller than the server
8397 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8398 INRIA or reporting this issue.
8403 * Fixed issue where DH client certificates are accepted without verification.
8404 An OpenSSL server will accept a DH certificate for client authentication
8405 without the certificate verify message. This effectively allows a client to
8406 authenticate without the use of a private key. This only affects servers
8407 which trust a client certificate authority which issues certificates
8408 containing DH keys: these are extremely rare and hardly ever encountered.
8409 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8415 * Ensure that the session ID context of an SSL is updated when its
8416 SSL_CTX is updated via SSL_set_SSL_CTX.
8418 The session ID context is typically set from the parent SSL_CTX,
8419 and can vary with the CTX.
8423 * Fix various certificate fingerprint issues.
8425 By using non-DER or invalid encodings outside the signed portion of a
8426 certificate the fingerprint can be changed without breaking the signature.
8427 Although no details of the signed portion of the certificate can be changed
8428 this can cause problems with some applications: e.g. those using the
8429 certificate fingerprint for blacklists.
8431 1. Reject signatures with non zero unused bits.
8433 If the BIT STRING containing the signature has non zero unused bits reject
8434 the signature. All current signature algorithms require zero unused bits.
8436 2. Check certificate algorithm consistency.
8438 Check the AlgorithmIdentifier inside TBS matches the one in the
8439 certificate signature. NB: this will result in signature failure
8440 errors for some broken certificates.
8442 Thanks to Konrad Kraszewski from Google for reporting this issue.
8444 3. Check DSA/ECDSA signatures use DER.
8446 Re-encode DSA/ECDSA signatures and compare with the original received
8447 signature. Return an error if there is a mismatch.
8449 This will reject various cases including garbage after signature
8450 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8451 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8452 (negative or with leading zeroes).
8454 Further analysis was conducted and fixes were developed by Stephen Henson
8455 of the OpenSSL core team.
8461 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8462 results on some platforms, including x86_64. This bug occurs at random
8463 with a very low probability, and is not known to be exploitable in any
8464 way, though its exact impact is difficult to determine. Thanks to Pieter
8465 Wuille (Blockstream) who reported this issue and also suggested an initial
8466 fix. Further analysis was conducted by the OpenSSL development team and
8467 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8468 the OpenSSL core team.
8473 * Do not resume sessions on the server if the negotiated protocol
8474 version does not match the session's version. Resuming with a different
8475 version, while not strictly forbidden by the RFC, is of questionable
8476 sanity and breaks all known clients.
8478 *David Benjamin, Emilia Käsper*
8480 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8481 early CCS messages during renegotiation. (Note that because
8482 renegotiation is encrypted, this early CCS was not exploitable.)
8486 * Tighten client-side session ticket handling during renegotiation:
8487 ensure that the client only accepts a session ticket if the server sends
8488 the extension anew in the ServerHello. Previously, a TLS client would
8489 reuse the old extension state and thus accept a session ticket if one was
8490 announced in the initial ServerHello.
8492 Similarly, ensure that the client requires a session ticket if one
8493 was advertised in the ServerHello. Previously, a TLS client would
8494 ignore a missing NewSessionTicket message.
8498 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8502 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8503 sends a carefully crafted handshake message, to cause OpenSSL to fail
8504 to free up to 64k of memory causing a memory leak. This could be
8505 exploited in a Denial Of Service attack. This issue affects OpenSSL
8506 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8507 whether SRTP is used or configured. Implementations of OpenSSL that
8508 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8510 The fix was developed by the OpenSSL team.
8515 * Session Ticket Memory Leak.
8517 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8518 integrity of that ticket is first verified. In the event of a session
8519 ticket integrity check failing, OpenSSL will fail to free memory
8520 causing a memory leak. By sending a large number of invalid session
8521 tickets an attacker could exploit this issue in a Denial Of Service
8527 * Build option no-ssl3 is incomplete.
8529 When OpenSSL is configured with "no-ssl3" as a build option, servers
8530 could accept and complete an SSL 3.0 handshake, and clients could be
8531 configured to send them.
8534 *Akamai and the OpenSSL team*
8536 * Add support for TLS_FALLBACK_SCSV.
8537 Client applications doing fallback retries should call
8538 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8541 *Adam Langley, Bodo Moeller*
8543 * Add additional DigestInfo checks.
8545 Re-encode DigestInto in DER and check against the original when
8546 verifying RSA signature: this will reject any improperly encoded
8547 DigestInfo structures.
8549 Note: this is a precautionary measure and no attacks are currently known.
8553 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8555 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8556 SRP code can be overrun an internal buffer. Add sanity check that
8557 g, A, B < N to SRP code.
8559 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8560 Group for discovering this issue.
8565 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8566 TLS 1.0 instead of higher protocol versions when the ClientHello message
8567 is badly fragmented. This allows a man-in-the-middle attacker to force a
8568 downgrade to TLS 1.0 even if both the server and the client support a
8569 higher protocol version, by modifying the client's TLS records.
8571 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8572 researching this issue.
8577 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8578 to a denial of service attack. A malicious server can crash the client
8579 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8580 ciphersuite and sending carefully crafted handshake messages.
8582 Thanks to Felix Gröbert (Google) for discovering and researching this
8588 * By sending carefully crafted DTLS packets an attacker could cause openssl
8589 to leak memory. This can be exploited through a Denial of Service attack.
8590 Thanks to Adam Langley for discovering and researching this issue.
8595 * An attacker can force openssl to consume large amounts of memory whilst
8596 processing DTLS handshake messages. This can be exploited through a
8597 Denial of Service attack.
8598 Thanks to Adam Langley for discovering and researching this issue.
8603 * An attacker can force an error condition which causes openssl to crash
8604 whilst processing DTLS packets due to memory being freed twice. This
8605 can be exploited through a Denial of Service attack.
8606 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8612 * If a multithreaded client connects to a malicious server using a resumed
8613 session and the server sends an ec point format extension it could write
8614 up to 255 bytes to freed memory.
8616 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8622 * A malicious server can crash an OpenSSL client with a null pointer
8623 dereference (read) by specifying an SRP ciphersuite even though it was not
8624 properly negotiated with the client. This can be exploited through a
8625 Denial of Service attack.
8627 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8628 discovering and researching this issue.
8633 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8634 X509_name_oneline, X509_name_print_ex et al. to leak some information
8635 from the stack. Applications may be affected if they echo pretty printing
8636 output to the attacker.
8638 Thanks to Ivan Fratric (Google) for discovering this issue.
8641 *Emilia Käsper, and Steve Henson*
8643 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8644 for corner cases. (Certain input points at infinity could lead to
8645 bogus results, with non-infinity inputs mapped to infinity too.)
8649 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8651 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8652 handshake can force the use of weak keying material in OpenSSL
8653 SSL/TLS clients and servers.
8655 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8656 researching this issue. ([CVE-2014-0224])
8658 *KIKUCHI Masashi, Steve Henson*
8660 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8661 OpenSSL DTLS client the code can be made to recurse eventually crashing
8664 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8667 *Imre Rad, Steve Henson*
8669 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8670 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8671 client or server. This is potentially exploitable to run arbitrary
8672 code on a vulnerable client or server.
8674 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8676 *Jüri Aedla, Steve Henson*
8678 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8679 are subject to a denial of service attack.
8681 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8682 this issue. ([CVE-2014-3470])
8684 *Felix Gröbert, Ivan Fratric, Steve Henson*
8686 * Harmonize version and its documentation. -f flag is used to display
8689 *mancha <mancha1@zoho.com>*
8691 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8692 in i2d_ECPrivateKey.
8694 *mancha <mancha1@zoho.com>*
8696 * Fix some double frees. These are not thought to be exploitable.
8698 *mancha <mancha1@zoho.com>*
8700 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8702 * A missing bounds check in the handling of the TLS heartbeat extension
8703 can be used to reveal up to 64k of memory to a connected client or
8706 Thanks for Neel Mehta of Google Security for discovering this bug and to
8707 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8708 preparing the fix ([CVE-2014-0160])
8710 *Adam Langley, Bodo Moeller*
8712 * Fix for the attack described in the paper "Recovering OpenSSL
8713 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8714 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8715 <http://eprint.iacr.org/2014/140>
8717 Thanks to Yuval Yarom and Naomi Benger for discovering this
8718 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8720 *Yuval Yarom and Naomi Benger*
8722 * TLS pad extension: draft-agl-tls-padding-03
8724 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8725 TLS client Hello record length value would otherwise be > 255 and
8726 less that 512 pad with a dummy extension containing zeroes so it
8727 is at least 512 bytes long.
8729 *Adam Langley, Steve Henson*
8731 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8733 * Fix for TLS record tampering bug. A carefully crafted invalid
8734 handshake could crash OpenSSL with a NULL pointer exception.
8735 Thanks to Anton Johansson for reporting this issues.
8738 * Keep original DTLS digest and encryption contexts in retransmission
8739 structures so we can use the previous session parameters if they need
8740 to be resent. ([CVE-2013-6450])
8744 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8745 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8746 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8747 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8748 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8749 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8751 *Rob Stradling, Adam Langley*
8753 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8755 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8756 supporting platforms or when small records were transferred.
8758 *Andy Polyakov, Steve Henson*
8760 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8762 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8764 This addresses the flaw in CBC record processing discovered by
8765 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8766 at: <http://www.isg.rhul.ac.uk/tls/>
8768 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8769 Security Group at Royal Holloway, University of London
8770 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8771 Emilia Käsper for the initial patch.
8774 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8776 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8777 ciphersuites which can be exploited in a denial of service attack.
8778 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8779 and detecting this bug and to Wolfgang Ettlinger
8780 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8785 * Return an error when checking OCSP signatures when key is NULL.
8786 This fixes a DoS attack. ([CVE-2013-0166])
8790 * Make openssl verify return errors.
8792 *Chris Palmer <palmer@google.com> and Ben Laurie*
8794 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8795 the right response is stapled. Also change SSL_get_certificate()
8796 so it returns the certificate actually sent.
8797 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8799 *Rob Stradling <rob.stradling@comodo.com>*
8801 * Fix possible deadlock when decoding public keys.
8805 * Don't use TLS 1.0 record version number in initial client hello
8810 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8812 * Sanity check record length before skipping explicit IV in TLS
8813 1.2, 1.1 and DTLS to fix DoS attack.
8815 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8816 fuzzing as a service testing platform.
8821 * Initialise tkeylen properly when encrypting CMS messages.
8822 Thanks to Solar Designer of Openwall for reporting this issue.
8826 * In FIPS mode don't try to use composite ciphers as they are not
8831 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8833 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8834 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8835 mean any application compiled against OpenSSL 1.0.0 headers setting
8836 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8837 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8838 0x10000000L Any application which was previously compiled against
8839 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8840 will need to be recompiled as a result. Letting be results in
8841 inability to disable specifically TLS 1.1 and in client context,
8842 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8846 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8847 disable just protocol X, but all protocols above X *if* there are
8848 protocols *below* X still enabled. In more practical terms it means
8849 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8850 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8851 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8856 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8858 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8859 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8860 in CRYPTO_realloc_clean.
8862 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8863 issue and to Adam Langley <agl@chromium.org> for fixing it.
8866 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8868 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8872 * Workarounds for some broken servers that "hang" if a client hello
8873 record length exceeds 255 bytes.
8875 1. Do not use record version number > TLS 1.0 in initial client
8876 hello: some (but not all) hanging servers will now work.
8877 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8878 the number of ciphers sent in the client hello. This should be
8879 set to an even number, such as 50, for example by passing:
8880 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8881 Most broken servers should now work.
8882 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8883 TLS 1.2 client support entirely.
8887 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8891 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8893 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8894 STRING form instead of a DigestInfo.
8898 * The format used for MDC2 RSA signatures is inconsistent between EVP
8899 and the RSA_sign/RSA_verify functions. This was made more apparent when
8900 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8901 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8902 the correct format in RSA_verify so both forms transparently work.
8906 * Some servers which support TLS 1.0 can choke if we initially indicate
8907 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8908 encrypted premaster secret. As a workaround use the maximum permitted
8909 client version in client hello, this should keep such servers happy
8910 and still work with previous versions of OpenSSL.
8914 * Add support for TLS/DTLS heartbeats.
8916 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8918 * Add support for SCTP.
8920 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8922 * Improved PRNG seeding for VOS.
8924 *Paul Green <Paul.Green@stratus.com>*
8926 * Extensive assembler packs updates, most notably:
8928 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8929 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8930 - x86_64: bit-sliced AES implementation;
8931 - ARM: NEON support, contemporary platforms optimizations;
8932 - s390x: z196 support;
8933 - `*`: GHASH and GF(2^m) multiplication implementations;
8937 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8938 (removal of unnecessary code)
8940 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8942 * Add TLS key material exporter from RFC 5705.
8946 * Add DTLS-SRTP negotiation from RFC 5764.
8950 * Add Next Protocol Negotiation,
8951 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8952 disabled with a no-npn flag to config or Configure. Code donated
8955 *Adam Langley <agl@google.com> and Ben Laurie*
8957 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8958 NIST-P256, NIST-P521, with constant-time single point multiplication on
8959 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8960 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8961 Code made available under Apache License version 2.0.
8963 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8964 line to include this in your build of OpenSSL, and run "make depend" (or
8965 "make update"). This enables the following EC_METHODs:
8967 EC_GFp_nistp224_method()
8968 EC_GFp_nistp256_method()
8969 EC_GFp_nistp521_method()
8971 EC_GROUP_new_by_curve_name() will automatically use these (while
8972 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8975 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8977 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8978 all platforms. Move ssize_t definition from e_os.h to the public
8979 header file e_os2.h as it now appears in public header file cms.h
8983 * New -sigopt option to the ca, req and x509 utilities. Additional
8984 signature parameters can be passed using this option and in
8989 * Add RSA PSS signing function. This will generate and set the
8990 appropriate AlgorithmIdentifiers for PSS based on those in the
8991 corresponding EVP_MD_CTX structure. No application support yet.
8995 * Support for companion algorithm specific ASN1 signing routines.
8996 New function ASN1_item_sign_ctx() signs a pre-initialised
8997 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8998 the appropriate parameters.
9002 * Add new algorithm specific ASN1 verification initialisation function
9003 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
9004 handling will be the same no matter what EVP_PKEY_METHOD is used.
9005 Add a PSS handler to support verification of PSS signatures: checked
9006 against a number of sample certificates.
9010 * Add signature printing for PSS. Add PSS OIDs.
9012 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
9014 * Add algorithm specific signature printing. An individual ASN1 method
9015 can now print out signatures instead of the standard hex dump.
9017 More complex signatures (e.g. PSS) can print out more meaningful
9018 information. Include DSA version that prints out the signature
9023 * Password based recipient info support for CMS library: implementing
9028 * Split password based encryption into PBES2 and PBKDF2 functions. This
9029 neatly separates the code into cipher and PBE sections and is required
9030 for some algorithms that split PBES2 into separate pieces (such as
9031 password based CMS).
9035 * Session-handling fixes:
9036 - Fix handling of connections that are resuming with a session ID,
9037 but also support Session Tickets.
9038 - Fix a bug that suppressed issuing of a new ticket if the client
9039 presented a ticket with an expired session.
9040 - Try to set the ticket lifetime hint to something reasonable.
9041 - Make tickets shorter by excluding irrelevant information.
9042 - On the client side, don't ignore renewed tickets.
9044 *Adam Langley, Bodo Moeller (Google)*
9046 * Fix PSK session representation.
9050 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
9052 This work was sponsored by Intel.
9056 * Add GCM support to TLS library. Some custom code is needed to split
9057 the IV between the fixed (from PRF) and explicit (from TLS record)
9058 portions. This adds all GCM ciphersuites supported by RFC5288 and
9059 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
9060 add a special AESGCM string for GCM only.
9064 * Expand range of ctrls for AES GCM. Permit setting invocation
9065 field on decrypt and retrieval of invocation field only on encrypt.
9069 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
9070 As required by RFC5289 these ciphersuites cannot be used if for
9071 versions of TLS earlier than 1.2.
9075 * For FIPS capable OpenSSL interpret a NULL default public key method
9076 as unset and return the appropriate default but do *not* set the default.
9077 This means we can return the appropriate method in applications that
9078 switch between FIPS and non-FIPS modes.
9082 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
9083 ENGINE is used then we cannot handle that in the FIPS module so we
9084 keep original code iff non-FIPS operations are allowed.
9088 * Add -attime option to openssl utilities.
9090 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
9092 * Redirect DSA and DH operations to FIPS module in FIPS mode.
9096 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
9097 FIPS EC methods unconditionally for now.
9101 * New build option no-ec2m to disable characteristic 2 code.
9105 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
9106 all cases can be covered as some introduce binary incompatibilities.
9110 * Redirect RSA operations to FIPS module including keygen,
9111 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
9115 * Add similar low-level API blocking to ciphers.
9119 * low-level digest APIs are not approved in FIPS mode: any attempt
9120 to use these will cause a fatal error. Applications that *really* want
9121 to use them can use the `private_*` version instead.
9125 * Redirect cipher operations to FIPS module for FIPS builds.
9129 * Redirect digest operations to FIPS module for FIPS builds.
9133 * Update build system to add "fips" flag which will link in fipscanister.o
9134 for static and shared library builds embedding a signature if needed.
9138 * Output TLS supported curves in preference order instead of numerical
9139 order. This is currently hardcoded for the highest order curves first.
9140 This should be configurable so applications can judge speed vs strength.
9144 * Add TLS v1.2 server support for client authentication.
9148 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
9153 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
9154 FIPS modules versions.
9158 * Add TLS v1.2 client side support for client authentication. Keep cache
9159 of handshake records longer as we don't know the hash algorithm to use
9160 until after the certificate request message is received.
9164 * Initial TLS v1.2 client support. Add a default signature algorithms
9165 extension including all the algorithms we support. Parse new signature
9166 format in client key exchange. Relax some ECC signing restrictions for
9167 TLS v1.2 as indicated in RFC5246.
9171 * Add server support for TLS v1.2 signature algorithms extension. Switch
9172 to new signature format when needed using client digest preference.
9173 All server ciphersuites should now work correctly in TLS v1.2. No client
9174 support yet and no support for client certificates.
9178 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
9179 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
9180 ciphersuites. At present only RSA key exchange ciphersuites work with
9181 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
9182 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
9183 and version checking.
9187 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
9188 with this defined it will not be affected by any changes to ssl internal
9189 structures. Add several utility functions to allow openssl application
9190 to work with OPENSSL_NO_SSL_INTERN defined.
9194 * A long standing patch to add support for SRP from EdelWeb (Peter
9195 Sylvester and Christophe Renou) was integrated.
9196 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
9197 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
9200 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
9204 * Permit abbreviated handshakes when renegotiating using the function
9205 SSL_renegotiate_abbreviated().
9207 *Robin Seggelmann <seggelmann@fh-muenster.de>*
9209 * Add call to ENGINE_register_all_complete() to
9210 ENGINE_load_builtin_engines(), so some implementations get used
9211 automatically instead of needing explicit application support.
9215 * Add support for TLS key exporter as described in RFC5705.
9217 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
9219 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
9220 a few changes are required:
9222 Add SSL_OP_NO_TLSv1_1 flag.
9223 Add TLSv1_1 methods.
9224 Update version checking logic to handle version 1.1.
9225 Add explicit IV handling (ported from DTLS code).
9226 Add command line options to s_client/s_server.
9233 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
9235 * X509_ATTRIBUTE memory leak
9237 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
9238 memory. This structure is used by the PKCS#7 and CMS routines so any
9239 application which reads PKCS#7 or CMS data from untrusted sources is
9240 affected. SSL/TLS is not affected.
9242 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
9248 * Race condition handling PSK identify hint
9250 If PSK identity hints are received by a multi-threaded client then
9251 the values are wrongly updated in the parent SSL_CTX structure. This can
9252 result in a race condition potentially leading to a double free of the
9258 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
9260 * Malformed ECParameters causes infinite loop
9262 When processing an ECParameters structure OpenSSL enters an infinite loop
9263 if the curve specified is over a specially malformed binary polynomial
9266 This can be used to perform denial of service against any
9267 system which processes public keys, certificate requests or
9268 certificates. This includes TLS clients and TLS servers with
9269 client authentication enabled.
9271 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9276 * Exploitable out-of-bounds read in X509_cmp_time
9278 X509_cmp_time does not properly check the length of the ASN1_TIME
9279 string and can read a few bytes out of bounds. In addition,
9280 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9283 An attacker can use this to craft malformed certificates and CRLs of
9284 various sizes and potentially cause a segmentation fault, resulting in
9285 a DoS on applications that verify certificates or CRLs. TLS clients
9286 that verify CRLs are affected. TLS clients and servers with client
9287 authentication enabled may be affected if they use custom verification
9290 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9291 independently by Hanno Böck.
9296 * PKCS7 crash with missing EnvelopedContent
9298 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9299 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9300 with missing content and trigger a NULL pointer dereference on parsing.
9302 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9303 structures from untrusted sources are affected. OpenSSL clients and
9304 servers are not affected.
9306 This issue was reported to OpenSSL by Michal Zalewski (Google).
9311 * CMS verify infinite loop with unknown hash function
9313 When verifying a signedData message the CMS code can enter an infinite loop
9314 if presented with an unknown hash function OID. This can be used to perform
9315 denial of service against any system which verifies signedData messages using
9317 This issue was reported to OpenSSL by Johannes Bauer.
9322 * Race condition handling NewSessionTicket
9324 If a NewSessionTicket is received by a multi-threaded client when attempting to
9325 reuse a previous ticket then a race condition can occur potentially leading to
9326 a double free of the ticket data.
9331 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9333 * Segmentation fault in ASN1_TYPE_cmp fix
9335 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9336 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9337 certificate signature algorithm consistency this can be used to crash any
9338 certificate verification operation and exploited in a DoS attack. Any
9339 application which performs certificate verification is vulnerable including
9340 OpenSSL clients and servers which enable client authentication.
9345 * ASN.1 structure reuse memory corruption fix
9347 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9348 memory corruption via an invalid write. Such reuse is and has been
9349 strongly discouraged and is believed to be rare.
9351 Applications that parse structures containing CHOICE or ANY DEFINED BY
9352 components may be affected. Certificate parsing (d2i_X509 and related
9353 functions) are however not affected. OpenSSL clients and servers are
9359 * PKCS7 NULL pointer dereferences fix
9361 The PKCS#7 parsing code does not handle missing outer ContentInfo
9362 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9363 missing content and trigger a NULL pointer dereference on parsing.
9365 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9366 otherwise parse PKCS#7 structures from untrusted sources are
9367 affected. OpenSSL clients and servers are not affected.
9369 This issue was reported to OpenSSL by Michal Zalewski (Google).
9374 * DoS via reachable assert in SSLv2 servers fix
9376 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9377 servers that both support SSLv2 and enable export cipher suites by sending
9378 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9380 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9381 (OpenSSL development team).
9386 * Use After Free following d2i_ECPrivatekey error fix
9388 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9389 could cause a use after free condition. This, in turn, could cause a double
9390 free in several private key parsing functions (such as d2i_PrivateKey
9391 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9392 for applications that receive EC private keys from untrusted
9393 sources. This scenario is considered rare.
9395 This issue was discovered by the BoringSSL project and fixed in their
9401 * X509_to_X509_REQ NULL pointer deref fix
9403 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9404 the certificate key is invalid. This function is rarely used in practice.
9406 This issue was discovered by Brian Carpenter.
9411 * Removed the export ciphers from the DEFAULT ciphers
9415 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9417 * Build fixes for the Windows and OpenVMS platforms
9419 *Matt Caswell and Richard Levitte*
9421 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9423 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9424 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9425 dereference. This could lead to a Denial Of Service attack. Thanks to
9426 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9431 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9432 dtls1_buffer_record function under certain conditions. In particular this
9433 could occur if an attacker sent repeated DTLS records with the same
9434 sequence number but for the next epoch. The memory leak could be exploited
9435 by an attacker in a Denial of Service attack through memory exhaustion.
9436 Thanks to Chris Mueller for reporting this issue.
9441 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9442 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9443 method would be set to NULL which could later result in a NULL pointer
9444 dereference. Thanks to Frank Schmirler for reporting this issue.
9449 * Abort handshake if server key exchange message is omitted for ephemeral
9452 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9453 reporting this issue.
9458 * Remove non-export ephemeral RSA code on client and server. This code
9459 violated the TLS standard by allowing the use of temporary RSA keys in
9460 non-export ciphersuites and could be used by a server to effectively
9461 downgrade the RSA key length used to a value smaller than the server
9462 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9463 INRIA or reporting this issue.
9468 * Fixed issue where DH client certificates are accepted without verification.
9469 An OpenSSL server will accept a DH certificate for client authentication
9470 without the certificate verify message. This effectively allows a client to
9471 authenticate without the use of a private key. This only affects servers
9472 which trust a client certificate authority which issues certificates
9473 containing DH keys: these are extremely rare and hardly ever encountered.
9474 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9480 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9481 results on some platforms, including x86_64. This bug occurs at random
9482 with a very low probability, and is not known to be exploitable in any
9483 way, though its exact impact is difficult to determine. Thanks to Pieter
9484 Wuille (Blockstream) who reported this issue and also suggested an initial
9485 fix. Further analysis was conducted by the OpenSSL development team and
9486 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9487 the OpenSSL core team.
9492 * Fix various certificate fingerprint issues.
9494 By using non-DER or invalid encodings outside the signed portion of a
9495 certificate the fingerprint can be changed without breaking the signature.
9496 Although no details of the signed portion of the certificate can be changed
9497 this can cause problems with some applications: e.g. those using the
9498 certificate fingerprint for blacklists.
9500 1. Reject signatures with non zero unused bits.
9502 If the BIT STRING containing the signature has non zero unused bits reject
9503 the signature. All current signature algorithms require zero unused bits.
9505 2. Check certificate algorithm consistency.
9507 Check the AlgorithmIdentifier inside TBS matches the one in the
9508 certificate signature. NB: this will result in signature failure
9509 errors for some broken certificates.
9511 Thanks to Konrad Kraszewski from Google for reporting this issue.
9513 3. Check DSA/ECDSA signatures use DER.
9515 Re-encode DSA/ECDSA signatures and compare with the original received
9516 signature. Return an error if there is a mismatch.
9518 This will reject various cases including garbage after signature
9519 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9520 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9521 (negative or with leading zeroes).
9523 Further analysis was conducted and fixes were developed by Stephen Henson
9524 of the OpenSSL core team.
9530 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9532 * Session Ticket Memory Leak.
9534 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9535 integrity of that ticket is first verified. In the event of a session
9536 ticket integrity check failing, OpenSSL will fail to free memory
9537 causing a memory leak. By sending a large number of invalid session
9538 tickets an attacker could exploit this issue in a Denial Of Service
9544 * Build option no-ssl3 is incomplete.
9546 When OpenSSL is configured with "no-ssl3" as a build option, servers
9547 could accept and complete an SSL 3.0 handshake, and clients could be
9548 configured to send them.
9551 *Akamai and the OpenSSL team*
9553 * Add support for TLS_FALLBACK_SCSV.
9554 Client applications doing fallback retries should call
9555 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9558 *Adam Langley, Bodo Moeller*
9560 * Add additional DigestInfo checks.
9562 Re-encode DigestInto in DER and check against the original when
9563 verifying RSA signature: this will reject any improperly encoded
9564 DigestInfo structures.
9566 Note: this is a precautionary measure and no attacks are currently known.
9570 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9572 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9573 to a denial of service attack. A malicious server can crash the client
9574 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9575 ciphersuite and sending carefully crafted handshake messages.
9577 Thanks to Felix Gröbert (Google) for discovering and researching this
9583 * By sending carefully crafted DTLS packets an attacker could cause openssl
9584 to leak memory. This can be exploited through a Denial of Service attack.
9585 Thanks to Adam Langley for discovering and researching this issue.
9590 * An attacker can force openssl to consume large amounts of memory whilst
9591 processing DTLS handshake messages. This can be exploited through a
9592 Denial of Service attack.
9593 Thanks to Adam Langley for discovering and researching this issue.
9598 * An attacker can force an error condition which causes openssl to crash
9599 whilst processing DTLS packets due to memory being freed twice. This
9600 can be exploited through a Denial of Service attack.
9601 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9607 * If a multithreaded client connects to a malicious server using a resumed
9608 session and the server sends an ec point format extension it could write
9609 up to 255 bytes to freed memory.
9611 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9617 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9618 X509_name_oneline, X509_name_print_ex et al. to leak some information
9619 from the stack. Applications may be affected if they echo pretty printing
9620 output to the attacker.
9622 Thanks to Ivan Fratric (Google) for discovering this issue.
9625 *Emilia Käsper, and Steve Henson*
9627 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9628 for corner cases. (Certain input points at infinity could lead to
9629 bogus results, with non-infinity inputs mapped to infinity too.)
9633 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9635 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9636 handshake can force the use of weak keying material in OpenSSL
9637 SSL/TLS clients and servers.
9639 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9640 researching this issue. ([CVE-2014-0224])
9642 *KIKUCHI Masashi, Steve Henson*
9644 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9645 OpenSSL DTLS client the code can be made to recurse eventually crashing
9648 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9651 *Imre Rad, Steve Henson*
9653 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9654 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9655 client or server. This is potentially exploitable to run arbitrary
9656 code on a vulnerable client or server.
9658 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9660 *Jüri Aedla, Steve Henson*
9662 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9663 are subject to a denial of service attack.
9665 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9666 this issue. ([CVE-2014-3470])
9668 *Felix Gröbert, Ivan Fratric, Steve Henson*
9670 * Harmonize version and its documentation. -f flag is used to display
9673 *mancha <mancha1@zoho.com>*
9675 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9676 in i2d_ECPrivateKey.
9678 *mancha <mancha1@zoho.com>*
9680 * Fix some double frees. These are not thought to be exploitable.
9682 *mancha <mancha1@zoho.com>*
9684 * Fix for the attack described in the paper "Recovering OpenSSL
9685 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9686 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9687 <http://eprint.iacr.org/2014/140>
9689 Thanks to Yuval Yarom and Naomi Benger for discovering this
9690 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9692 *Yuval Yarom and Naomi Benger*
9694 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9696 * Keep original DTLS digest and encryption contexts in retransmission
9697 structures so we can use the previous session parameters if they need
9698 to be resent. ([CVE-2013-6450])
9702 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9703 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9704 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9705 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9706 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9707 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9709 *Rob Stradling, Adam Langley*
9711 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9713 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9715 This addresses the flaw in CBC record processing discovered by
9716 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9717 at: <http://www.isg.rhul.ac.uk/tls/>
9719 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9720 Security Group at Royal Holloway, University of London
9721 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9722 Emilia Käsper for the initial patch.
9725 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9727 * Return an error when checking OCSP signatures when key is NULL.
9728 This fixes a DoS attack. ([CVE-2013-0166])
9732 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9733 the right response is stapled. Also change SSL_get_certificate()
9734 so it returns the certificate actually sent.
9735 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9736 (This is a backport)
9738 *Rob Stradling <rob.stradling@comodo.com>*
9740 * Fix possible deadlock when decoding public keys.
9744 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9746 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9749 * Sanity check record length before skipping explicit IV in DTLS
9752 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9753 fuzzing as a service testing platform.
9758 * Initialise tkeylen properly when encrypting CMS messages.
9759 Thanks to Solar Designer of Openwall for reporting this issue.
9763 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9765 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9766 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9767 in CRYPTO_realloc_clean.
9769 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9770 issue and to Adam Langley <agl@chromium.org> for fixing it.
9773 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9775 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9777 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9778 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9779 content decryption and always return the same error. Note: this attack
9780 needs on average 2^20 messages so it only affects automated senders. The
9781 old behaviour can be re-enabled in the CMS code by setting the
9782 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9783 an MMA defence is not necessary.
9784 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9785 this issue. ([CVE-2012-0884])
9789 * Fix CVE-2011-4619: make sure we really are receiving a
9790 client hello before rejecting multiple SGC restarts. Thanks to
9791 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9795 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9797 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9798 Thanks to Antonio Martin, Enterprise Secure Access Research and
9799 Development, Cisco Systems, Inc. for discovering this bug and
9800 preparing a fix. ([CVE-2012-0050])
9804 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9806 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9807 of the Vaudenay padding oracle attack on CBC mode encryption
9808 which enables an efficient plaintext recovery attack against
9809 the OpenSSL implementation of DTLS. Their attack exploits timing
9810 differences arising during decryption processing. A research
9811 paper describing this attack can be found at:
9812 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9813 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9814 Security Group at Royal Holloway, University of London
9815 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9816 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9817 for preparing the fix. ([CVE-2011-4108])
9819 *Robin Seggelmann, Michael Tuexen*
9821 * Clear bytes used for block padding of SSL 3.0 records.
9824 *Adam Langley (Google)*
9826 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9827 Kadianakis <desnacked@gmail.com> for discovering this issue and
9828 Adam Langley for preparing the fix. ([CVE-2011-4619])
9830 *Adam Langley (Google)*
9832 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9834 *Andrey Kulikov <amdeich@gmail.com>*
9836 * Prevent malformed RFC3779 data triggering an assertion failure.
9837 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9838 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9840 *Rob Austein <sra@hactrn.net>*
9842 * Improved PRNG seeding for VOS.
9844 *Paul Green <Paul.Green@stratus.com>*
9846 * Fix ssl_ciph.c set-up race.
9848 *Adam Langley (Google)*
9850 * Fix spurious failures in ecdsatest.c.
9852 *Emilia Käsper (Google)*
9854 * Fix the BIO_f_buffer() implementation (which was mixing different
9855 interpretations of the `..._len` fields).
9857 *Adam Langley (Google)*
9859 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9860 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9861 threads won't reuse the same blinding coefficients.
9863 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9864 lock to call BN_BLINDING_invert_ex, and avoids one use of
9865 BN_BLINDING_update for each BN_BLINDING structure (previously,
9866 the last update always remained unused).
9868 *Emilia Käsper (Google)*
9870 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9872 *Bob Buckholz (Google)*
9874 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9876 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9877 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9879 *Kaspar Brand <ossl@velox.ch>*
9881 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9882 for multi-threaded use of ECDH. ([CVE-2011-3210])
9884 *Adam Langley (Google)*
9886 * Fix x509_name_ex_d2i memory leak on bad inputs.
9890 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9891 signature public key algorithm by using OID xref utilities instead.
9892 Before this you could only use some ECC ciphersuites with SHA1 only.
9896 * Add protection against ECDSA timing attacks as mentioned in the paper
9897 by Billy Bob Brumley and Nicola Tuveri, see:
9898 <http://eprint.iacr.org/2011/232.pdf>
9900 *Billy Bob Brumley and Nicola Tuveri*
9902 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9904 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9906 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9908 * Fix bug in string printing code: if *any* escaping is enabled we must
9909 escape the escape character (backslash) or the resulting string is
9914 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9916 * Disable code workaround for ancient and obsolete Netscape browsers
9917 and servers: an attacker can use it in a ciphersuite downgrade attack.
9918 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9922 * Fixed J-PAKE implementation error, originally discovered by
9923 Sebastien Martini, further info and confirmation from Stefan
9924 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9928 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9930 * Fix extension code to avoid race conditions which can result in a buffer
9931 overrun vulnerability: resumed sessions must not be modified as they can
9932 be shared by multiple threads. CVE-2010-3864
9936 * Fix WIN32 build system to correctly link an ENGINE directory into
9941 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9943 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9946 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9948 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9950 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9951 context. The operation can be customised via the ctrl mechanism in
9952 case ENGINEs want to include additional functionality.
9956 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9960 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9961 output hashes compatible with older versions of OpenSSL.
9963 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9965 * Fix compression algorithm handling: if resuming a session use the
9966 compression algorithm of the resumed session instead of determining
9967 it from client hello again. Don't allow server to change algorithm.
9971 * Add load_crls() function to commands tidying load_certs() too. Add option
9972 to verify utility to allow additional CRLs to be included.
9976 * Update OCSP request code to permit adding custom headers to the request:
9977 some responders need this.
9981 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9984 *Julia Lawall <julia@diku.dk>*
9986 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9987 needlessly dereferenced structures, used obsolete functions and
9988 didn't handle all updated verify codes correctly.
9992 * Disable MD2 in the default configuration.
9996 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9997 indicate the initial BIO being pushed or popped. This makes it possible
9998 to determine whether the BIO is the one explicitly called or as a result
9999 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
10000 it handles reference counts correctly and doesn't zero out the I/O bio
10001 when it is not being explicitly popped. WARNING: applications which
10002 included workarounds for the old buggy behaviour will need to be modified
10003 or they could free up already freed BIOs.
10007 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
10008 renaming to all platforms (within the 0.9.8 branch, this was
10009 done conditionally on Netware platforms to avoid a name clash).
10011 *Guenter <lists@gknw.net>*
10013 * Add ECDHE and PSK support to DTLS.
10015 *Michael Tuexen <tuexen@fh-muenster.de>*
10017 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
10022 * Add "missing" function EVP_MD_flags() (without this the only way to
10023 retrieve a digest flags is by accessing the structure directly. Update
10024 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
10025 or cipher is registered as in the "from" argument. Print out all
10026 registered digests in the dgst usage message instead of manually
10027 attempting to work them out.
10031 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
10032 this allows the use of compression and extensions. Change default cipher
10033 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
10034 by default unless an application cipher string requests it.
10038 * Alter match criteria in PKCS12_parse(). It used to try to use local
10039 key ids to find matching certificates and keys but some PKCS#12 files
10040 don't follow the (somewhat unwritten) rules and this strategy fails.
10041 Now just gather all certificates together and the first private key
10042 then look for the first certificate that matches the key.
10046 * Support use of registered digest and cipher names for dgst and cipher
10047 commands instead of having to add each one as a special case. So now
10054 openssl dgst -sha256 foo
10056 and this works for ENGINE based algorithms too.
10060 * Update Gost ENGINE to support parameter files.
10062 *Victor B. Wagner <vitus@cryptocom.ru>*
10064 * Support GeneralizedTime in ca utility.
10066 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
10068 * Enhance the hash format used for certificate directory links. The new
10069 form uses the canonical encoding (meaning equivalent names will work
10070 even if they aren't identical) and uses SHA1 instead of MD5. This form
10071 is incompatible with the older format and as a result c_rehash should
10072 be used to rebuild symbolic links.
10076 * Make PKCS#8 the default write format for private keys, replacing the
10077 traditional format. This form is standardised, more secure and doesn't
10078 include an implicit MD5 dependency.
10082 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
10083 committed to OpenSSL should pass this lot as a minimum.
10087 * Add session ticket override functionality for use by EAP-FAST.
10089 *Jouni Malinen <j@w1.fi>*
10091 * Modify HMAC functions to return a value. Since these can be implemented
10092 in an ENGINE errors can occur.
10096 * Type-checked OBJ_bsearch_ex.
10100 * Type-checked OBJ_bsearch. Also some constification necessitated
10101 by type-checking. Still to come: TXT_DB, bsearch(?),
10102 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
10107 * New function OPENSSL_gmtime_adj() to add a specific number of days and
10108 seconds to a tm structure directly, instead of going through OS
10109 specific date routines. This avoids any issues with OS routines such
10110 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
10111 and X509_time_adj_ex() to cover the extended range. The existing
10112 X509_time_adj() is still usable and will no longer have any date issues.
10116 * Delta CRL support. New use deltas option which will attempt to locate
10117 and search any appropriate delta CRLs available.
10119 This work was sponsored by Google.
10123 * Support for CRLs partitioned by reason code. Reorganise CRL processing
10124 code and add additional score elements. Validate alternate CRL paths
10125 as part of the CRL checking and indicate a new error "CRL path validation
10126 error" in this case. Applications wanting additional details can use
10127 the verify callback and check the new "parent" field. If this is not
10128 NULL CRL path validation is taking place. Existing applications won't
10129 see this because it requires extended CRL support which is off by
10132 This work was sponsored by Google.
10136 * Support for freshest CRL extension.
10138 This work was sponsored by Google.
10142 * Initial indirect CRL support. Currently only supported in the CRLs
10143 passed directly and not via lookup. Process certificate issuer
10144 CRL entry extension and lookup CRL entries by bother issuer name
10145 and serial number. Check and process CRL issuer entry in IDP extension.
10147 This work was sponsored by Google.
10151 * Add support for distinct certificate and CRL paths. The CRL issuer
10152 certificate is validated separately in this case. Only enabled if
10153 an extended CRL support flag is set: this flag will enable additional
10154 CRL functionality in future.
10156 This work was sponsored by Google.
10160 * Add support for policy mappings extension.
10162 This work was sponsored by Google.
10166 * Fixes to pathlength constraint, self issued certificate handling,
10167 policy processing to align with RFC3280 and PKITS tests.
10169 This work was sponsored by Google.
10173 * Support for name constraints certificate extension. DN, email, DNS
10174 and URI types are currently supported.
10176 This work was sponsored by Google.
10180 * To cater for systems that provide a pointer-based thread ID rather
10181 than numeric, deprecate the current numeric thread ID mechanism and
10182 replace it with a structure and associated callback type. This
10183 mechanism allows a numeric "hash" to be extracted from a thread ID in
10184 either case, and on platforms where pointers are larger than 'long',
10185 mixing is done to help ensure the numeric 'hash' is usable even if it
10186 can't be guaranteed unique. The default mechanism is to use "&errno"
10187 as a pointer-based thread ID to distinguish between threads.
10189 Applications that want to provide their own thread IDs should now use
10190 CRYPTO_THREADID_set_callback() to register a callback that will call
10191 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
10193 Note that ERR_remove_state() is now deprecated, because it is tied
10194 to the assumption that thread IDs are numeric. ERR_remove_state(0)
10195 to free the current thread's error state should be replaced by
10196 ERR_remove_thread_state(NULL).
10198 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
10199 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
10200 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
10201 application was previously providing a numeric thread callback that
10202 was inappropriate for distinguishing threads, then uniqueness might
10203 have been obtained with &errno that happened immediately in the
10204 intermediate development versions of OpenSSL; this is no longer the
10205 case, the numeric thread callback will now override the automatic use
10208 *Geoff Thorpe, with help from Bodo Moeller*
10210 * Initial support for different CRL issuing certificates. This covers a
10211 simple case where the self issued certificates in the chain exist and
10212 the real CRL issuer is higher in the existing chain.
10214 This work was sponsored by Google.
10218 * Removed effectively defunct crypto/store from the build.
10222 * Revamp of STACK to provide stronger type-checking. Still to come:
10223 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
10224 ASN1_STRING, CONF_VALUE.
10228 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
10229 RAM on SSL connections. This option can save about 34k per idle SSL.
10233 * Revamp of LHASH to provide stronger type-checking. Still to come:
10234 STACK, TXT_DB, bsearch, qsort.
10238 * Initial support for Cryptographic Message Syntax (aka CMS) based
10239 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
10240 support for data, signedData, compressedData, digestedData and
10241 encryptedData, envelopedData types included. Scripts to check against
10242 RFC4134 examples draft and interop and consistency checks of many
10243 content types and variants.
10247 * Add options to enc utility to support use of zlib compression BIO.
10251 * Extend mk1mf to support importing of options and assembly language
10252 files from Configure script, currently only included in VC-WIN32.
10253 The assembly language rules can now optionally generate the source
10254 files from the associated perl scripts.
10258 * Implement remaining functionality needed to support GOST ciphersuites.
10259 Interop testing has been performed using CryptoPro implementations.
10261 *Victor B. Wagner <vitus@cryptocom.ru>*
10263 * s390x assembler pack.
10267 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10272 * Implement Opaque PRF Input TLS extension as specified in
10273 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10274 official specification yet and no extension type assignment by
10275 IANA exists, this extension (for now) will have to be explicitly
10276 enabled when building OpenSSL by providing the extension number
10277 to use. For example, specify an option
10279 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10281 to the "config" or "Configure" script to enable the extension,
10282 assuming extension number 0x9527 (which is a completely arbitrary
10283 and unofficial assignment based on the MD5 hash of the Internet
10284 Draft). Note that by doing so, you potentially lose
10285 interoperability with other TLS implementations since these might
10286 be using the same extension number for other purposes.
10288 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10289 opaque PRF input value to use in the handshake. This will create
10290 an internal copy of the length-'len' string at 'src', and will
10291 return non-zero for success.
10293 To get more control and flexibility, provide a callback function
10296 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10297 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10301 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10304 Callback function 'cb' will be called in handshakes, and is
10305 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10306 Argument 'arg' is for application purposes (the value as given to
10307 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10308 be provided to the callback function). The callback function
10309 has to return non-zero to report success: usually 1 to use opaque
10310 PRF input just if possible, or 2 to enforce use of the opaque PRF
10311 input. In the latter case, the library will abort the handshake
10312 if opaque PRF input is not successfully negotiated.
10314 Arguments 'peerinput' and 'len' given to the callback function
10315 will always be NULL and 0 in the case of a client. A server will
10316 see the client's opaque PRF input through these variables if
10317 available (NULL and 0 otherwise). Note that if the server
10318 provides an opaque PRF input, the length must be the same as the
10319 length of the client's opaque PRF input.
10321 Note that the callback function will only be called when creating
10322 a new session (session resumption can resume whatever was
10323 previously negotiated), and will not be called in SSL 2.0
10324 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10325 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10326 for applications that need to enforce opaque PRF input.
10330 * Update ssl code to support digests other than SHA1+MD5 for handshake
10333 *Victor B. Wagner <vitus@cryptocom.ru>*
10335 * Add RFC4507 support to OpenSSL. This includes the corrections in
10336 RFC4507bis. The encrypted ticket format is an encrypted encoded
10337 SSL_SESSION structure, that way new session features are automatically
10340 If a client application caches session in an SSL_SESSION structure
10341 support is transparent because tickets are now stored in the encoded
10344 The SSL_CTX structure automatically generates keys for ticket
10345 protection in servers so again support should be possible
10346 with no application modification.
10348 If a client or server wishes to disable RFC4507 support then the option
10349 SSL_OP_NO_TICKET can be set.
10351 Add a TLS extension debugging callback to allow the contents of any client
10352 or server extensions to be examined.
10354 This work was sponsored by Google.
10358 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10359 OpenSSL should now compile cleanly on gcc 4.2
10361 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10363 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10364 support including streaming MAC support: this is required for GOST
10365 ciphersuite support.
10367 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10369 * Add option -stream to use PKCS#7 streaming in smime utility. New
10370 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10371 to output in BER and PEM format.
10375 * Experimental support for use of HMAC via EVP_PKEY interface. This
10376 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10377 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10378 ENGINE support for HMAC keys which are unextractable. New -mac and
10379 -macopt options to dgst utility.
10383 * New option -sigopt to dgst utility. Update dgst to use
10384 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10385 alternative signing parameters such as X9.31 or PSS in the dgst
10390 * Change ssl_cipher_apply_rule(), the internal function that does
10391 the work each time a ciphersuite string requests enabling
10392 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10393 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10394 the order of disabled ciphersuites such that those ciphersuites
10395 that most recently went from enabled to disabled not only stay
10396 in order with respect to each other, but also have higher priority
10397 than other disabled ciphersuites the next time ciphersuites are
10400 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10401 the same ciphersuites as with "HIGH" alone, but in a specific
10402 order where the PSK ciphersuites come first (since they are the
10403 most recently disabled ciphersuites when "HIGH" is parsed).
10405 Also, change ssl_create_cipher_list() (using this new
10406 functionality) such that between otherwise identical
10407 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10412 * Change ssl_create_cipher_list() so that it automatically
10413 arranges the ciphersuites in reasonable order before starting
10414 to process the rule string. Thus, the definition for "DEFAULT"
10415 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10416 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10417 This makes it much easier to arrive at a reasonable default order
10418 in applications for which anonymous ciphers are OK (meaning
10419 that you can't actually use DEFAULT).
10421 *Bodo Moeller; suggested by Victor Duchovni*
10423 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10424 processing) into multiple integers instead of setting
10425 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10426 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10427 (These masks as well as the individual bit definitions are hidden
10428 away into the non-exported interface ssl/ssl_locl.h, so this
10429 change to the definition of the SSL_CIPHER structure shouldn't
10430 affect applications.) This give us more bits for each of these
10431 categories, so there is no longer a need to coagulate AES128 and
10432 AES256 into a single algorithm bit, and to coagulate Camellia128
10433 and Camellia256 into a single algorithm bit, which has led to all
10436 Thus, among other things, the kludge introduced in 0.9.7m and
10437 0.9.8e for masking out AES256 independently of AES128 or masking
10438 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10440 With the change, we also introduce new ciphersuite aliases that
10441 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10446 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10447 Use the leftmost N bytes of the signature input if the input is
10448 larger than the prime q (with N being the size in bytes of q).
10452 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10453 it yet and it is largely untested.
10457 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10461 * Initial incomplete changes to avoid need for function casts in OpenSSL
10462 some compilers (gcc 4.2 and later) reject their use. Safestack is
10463 reimplemented. Update ASN1 to avoid use of legacy functions.
10467 * Win32/64 targets are linked with Winsock2.
10471 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10472 to external functions. This can be used to increase CRL handling
10473 efficiency especially when CRLs are very large by (for example) storing
10474 the CRL revoked certificates in a database.
10478 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10479 new CRLs added to a directory can be used. New command line option
10480 -verify_return_error to s_client and s_server. This causes real errors
10481 to be returned by the verify callback instead of carrying on no matter
10482 what. This reflects the way a "real world" verify callback would behave.
10486 * GOST engine, supporting several GOST algorithms and public key formats.
10487 Kindly donated by Cryptocom.
10491 * Partial support for Issuing Distribution Point CRL extension. CRLs
10492 partitioned by DP are handled but no indirect CRL or reason partitioning
10493 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10494 selected via a scoring technique which handles IDP and AKID in CRLs.
10498 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10499 will ultimately be used for all verify operations: this will remove the
10500 X509_STORE dependency on certificate verification and allow alternative
10501 lookup methods. X509_STORE based implementations of these two callbacks.
10505 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10506 Modify get_crl() to find a valid (unexpired) CRL if possible.
10510 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10511 this would be called X509_CRL_cmp() but that name is already used by
10512 a function that just compares CRL issuer names. Cache several CRL
10513 extensions in X509_CRL structure and cache CRLDP in X509.
10517 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10518 this maps equivalent X509_NAME structures into a consistent structure.
10519 Name comparison can then be performed rapidly using memcmp().
10523 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10528 * Allow digests to supply their own micalg string for S/MIME type using
10529 the ctrl EVP_MD_CTRL_MICALG.
10533 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10534 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10535 ctrl. It can then customise the structure before and/or after signing
10540 * New function OBJ_add_sigid() to allow application defined signature OIDs
10541 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10542 to free up any added signature OIDs.
10546 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10547 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10548 digest and cipher tables. New options added to openssl utility:
10549 list-message-digest-algorithms and list-cipher-algorithms.
10553 * Change the array representation of binary polynomials: the list
10554 of degrees of non-zero coefficients is now terminated with -1.
10555 Previously it was terminated with 0, which was also part of the
10556 value; thus, the array representation was not applicable to
10557 polynomials where t^0 has coefficient zero. This change makes
10558 the array representation useful in a more general context.
10562 * Various modifications and fixes to SSL/TLS cipher string
10563 handling. For ECC, the code now distinguishes between fixed ECDH
10564 with RSA certificates on the one hand and with ECDSA certificates
10565 on the other hand, since these are separate ciphersuites. The
10566 unused code for Fortezza ciphersuites has been removed.
10568 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10569 (not "ECDHE"). For consistency with the code for DH
10570 certificates, use of ECDH certificates is now considered ECDH
10571 authentication, not RSA or ECDSA authentication (the latter is
10572 merely the CA's signing algorithm and not actively used in the
10575 The temporary ciphersuite alias "ECCdraft" is no longer
10576 available, and ECC ciphersuites are no longer excluded from "ALL"
10577 and "DEFAULT". The following aliases now exist for RFC 4492
10578 ciphersuites, most of these by analogy with the DH case:
10580 kECDHr - ECDH cert, signed with RSA
10581 kECDHe - ECDH cert, signed with ECDSA
10582 kECDH - ECDH cert (signed with either RSA or ECDSA)
10583 kEECDH - ephemeral ECDH
10584 ECDH - ECDH cert or ephemeral ECDH
10587 aECDSA - ECDSA cert
10590 AECDH - anonymous ECDH
10591 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10595 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10596 Use correct micalg parameters depending on digest(s) in signed message.
10600 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10601 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10605 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10606 an engine to register a method. Add ENGINE lookups for methods and
10607 functional reference processing.
10611 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10612 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10617 * New -resign option to smime utility. This adds one or more signers
10618 to an existing PKCS#7 signedData structure. Also -md option to use an
10619 alternative message digest algorithm for signing.
10623 * Tidy up PKCS#7 routines and add new functions to make it easier to
10624 create PKCS7 structures containing multiple signers. Update smime
10625 application to support multiple signers.
10629 * New -macalg option to pkcs12 utility to allow setting of an alternative
10634 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10635 Reorganize PBE internals to lookup from a static table using NIDs,
10636 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10637 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10638 PRF which will be automatically used with PBES2.
10642 * Replace the algorithm specific calls to generate keys in "req" with the
10647 * Update PKCS#7 enveloped data routines to use new API. This is now
10648 supported by any public key method supporting the encrypt operation. A
10649 ctrl is added to allow the public key algorithm to examine or modify
10650 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10655 * Add a ctrl to asn1 method to allow a public key algorithm to express
10656 a default digest type to use. In most cases this will be SHA1 but some
10657 algorithms (such as GOST) need to specify an alternative digest. The
10658 return value indicates how strong the preference is 1 means optional and
10659 2 is mandatory (that is it is the only supported type). Modify
10660 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10661 use the default md. Update openssl utilities to use the default digest
10662 type for signing if it is not explicitly indicated.
10666 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10667 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10668 signing method from the key type. This effectively removes the link
10669 between digests and public key types.
10673 * Add an OID cross reference table and utility functions. Its purpose is to
10674 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10675 rsaEncryption. This will allow some of the algorithm specific hackery
10676 needed to use the correct OID to be removed.
10680 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10681 structures for PKCS7_sign(). They are now set up by the relevant public
10686 * Add provisional EC pkey method with support for ECDSA and ECDH.
10690 * Add support for key derivation (agreement) in the API, DH method and
10695 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10696 public and private key formats. As a side effect these add additional
10697 command line functionality not previously available: DSA signatures can be
10698 generated and verified using pkeyutl and DH key support and generation in
10705 *Oliver Tappe <zooey@hirschkaefer.de>*
10707 * New make target "install_html_docs" installs HTML renditions of the
10710 *Oliver Tappe <zooey@hirschkaefer.de>*
10712 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10713 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10714 support key and parameter generation and add initial key generation
10715 functionality for RSA.
10719 * Add functions for main EVP_PKEY_method operations. The undocumented
10720 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10721 `EVP_PKEY_{encrypt,decrypt}_old`.
10725 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10726 key API, doesn't do much yet.
10730 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10731 public key algorithms. New option to openssl utility:
10732 "list-public-key-algorithms" to print out info.
10736 * Implement the Supported Elliptic Curves Extension for
10737 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10741 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10742 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10746 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10747 utilities such as rsa, dsa, dsaparam etc except they process any key
10752 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10753 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10754 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10759 * Initial support for pluggable public key ASN1.
10760 De-spaghettify the public key ASN1 handling. Move public and private
10761 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10762 algorithm specific handling to a single module within the relevant
10763 algorithm directory. Add functions to allow (near) opaque processing
10764 of public and private key structures.
10768 * Implement the Supported Point Formats Extension for
10769 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10773 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10774 for the psk identity [hint] and the psk callback functions to the
10775 SSL_SESSION, SSL and SSL_CTX structure.
10778 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10782 SSL_CTX_use_psk_identity_hint
10783 SSL_get_psk_identity_hint
10784 SSL_get_psk_identity
10785 SSL_use_psk_identity_hint
10787 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10789 * Add RFC 3161 compliant time stamp request creation, response generation
10790 and response verification functionality.
10792 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10794 * Add initial support for TLS extensions, specifically for the server_name
10795 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10796 have new members for a hostname. The SSL data structure has an
10797 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10798 stored in that context to allow for session resumption, even after the
10799 SSL has been switched to a new SSL_CTX in reaction to a client's
10800 server_name extension.
10802 New functions (subject to change):
10804 SSL_get_servername()
10805 SSL_get_servername_type()
10808 New CTRL codes and macros (subject to change):
10810 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10811 - SSL_CTX_set_tlsext_servername_callback()
10812 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10813 - SSL_CTX_set_tlsext_servername_arg()
10814 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10816 openssl s_client has a new '-servername ...' option.
10818 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10819 '-key2 ...', '-servername_fatal' (subject to change). This allows
10820 testing the HostName extension for a specific single hostname ('-cert'
10821 and '-key' remain fallbacks for handshakes without HostName
10822 negotiation). If the unrecognized_name alert has to be sent, this by
10823 default is a warning; it becomes fatal with the '-servername_fatal'
10826 *Peter Sylvester, Remy Allais, Christophe Renou*
10828 * Whirlpool hash implementation is added.
10832 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10833 bn(64,32). Because of instruction set limitations it doesn't have
10834 any negative impact on performance. This was done mostly in order
10835 to make it possible to share assembler modules, such as bn_mul_mont
10836 implementations, between 32- and 64-bit builds without hassle.
10840 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10841 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10846 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10847 dedicated Montgomery multiplication procedure, is introduced.
10848 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10849 "64-bit" performance on certain 32-bit targets.
10853 * New option SSL_OP_NO_COMP to disable use of compression selectively
10854 in SSL structures. New SSL ctrl to set maximum send fragment size.
10855 Save memory by setting the I/O buffer sizes dynamically instead of
10856 using the maximum available value.
10860 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10861 in addition to the text details.
10865 * Very, very preliminary EXPERIMENTAL support for printing of general
10866 ASN1 structures. This currently produces rather ugly output and doesn't
10867 handle several customised structures at all.
10871 * Integrated support for PVK file format and some related formats such
10872 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10873 these in the 'rsa' and 'dsa' utilities.
10877 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10881 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10882 place for the (very old) "NETSCAPE" format certificates which are now
10883 handled using new ASN1 code equivalents.
10887 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10888 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10889 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10893 * Modify CRL distribution points extension code to print out previously
10894 unsupported fields. Enhance extension setting code to allow setting of
10899 * Add print and set support for Issuing Distribution Point CRL extension.
10903 * Change 'Configure' script to enable Camellia by default.
10910 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10912 * When rejecting SSL/TLS records due to an incorrect version number, never
10913 update s->server with a new major version number. As of
10914 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10915 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10916 the previous behavior could result in a read attempt at NULL when
10917 receiving specific incorrect SSL/TLS records once record payload
10918 protection is active. ([CVE-2010-0740])
10920 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10922 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10923 could be crashed if the relevant tables were not present (e.g. chrooted).
10925 *Tomas Hoger <thoger@redhat.com>*
10927 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10929 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10931 *Martin Olsson, Neel Mehta*
10933 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10934 accommodate for stack sorting, always a write lock!).
10938 * On some versions of WIN32 Heap32Next is very slow. This can cause
10939 excessive delays in the RAND_poll(): over a minute. As a workaround
10940 include a time check in the inner Heap32Next loop too.
10944 * The code that handled flushing of data in SSL/TLS originally used the
10945 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10946 the problem outlined in PR#1949. The fix suggested there however can
10947 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10948 of Apache). So instead simplify the code to flush unconditionally.
10949 This should be fine since flushing with no data to flush is a no op.
10953 * Handle TLS versions 2.0 and later properly and correctly use the
10954 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10955 off ancient servers have a habit of sticking around for a while...
10959 * Modify compression code so it frees up structures without using the
10960 ex_data callbacks. This works around a problem where some applications
10961 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10962 restarting) then use compression (e.g. SSL with compression) later.
10963 This results in significant per-connection memory leaks and
10964 has caused some security issues including CVE-2008-1678 and
10969 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10970 change when encrypting or decrypting.
10974 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10975 connect and renegotiate with servers which do not support RI.
10976 Until RI is more widely deployed this option is enabled by default.
10980 * Add "missing" ssl ctrls to clear options and mode.
10984 * If client attempts to renegotiate and doesn't support RI respond with
10985 a no_renegotiation alert as required by RFC5746. Some renegotiating
10986 TLS clients will continue a connection gracefully when they receive
10987 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10988 waiting for a server hello which it will never receive. Now we treat a
10989 received no_renegotiation alert as a fatal error. This is because
10990 applications requesting a renegotiation might well expect it to succeed
10991 and would have no code in place to handle the server denying it so the
10992 only safe thing to do is to terminate the connection.
10996 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10997 peer supports secure renegotiation and 0 otherwise. Print out peer
10998 renegotiation support in s_client/s_server.
11002 * Replace the highly broken and deprecated SPKAC certification method with
11003 the updated NID creation version. This should correctly handle UTF8.
11007 * Implement RFC5746. Re-enable renegotiation but require the extension
11008 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
11009 turns out to be a bad idea. It has been replaced by
11010 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
11011 SSL_CTX_set_options(). This is really not recommended unless you
11012 know what you are doing.
11014 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
11016 * Fixes to stateless session resumption handling. Use initial_ctx when
11017 issuing and attempting to decrypt tickets in case it has changed during
11018 servername handling. Use a non-zero length session ID when attempting
11019 stateless session resumption: this makes it possible to determine if
11020 a resumption has occurred immediately after receiving server hello
11021 (several places in OpenSSL subtly assume this) instead of later in
11026 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
11027 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
11028 fixes for a few places where the return code is not checked
11031 *Julia Lawall <julia@diku.dk>*
11033 * Add --strict-warnings option to Configure script to include devteam
11034 warnings in other configurations.
11038 * Add support for --libdir option and LIBDIR variable in makefiles. This
11039 makes it possible to install openssl libraries in locations which
11040 have names other than "lib", for example "/usr/lib64" which some
11043 *Steve Henson, based on patch from Jeremy Utley*
11045 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
11046 X690 8.9.12 and can produce some misleading textual output of OIDs.
11048 *Steve Henson, reported by Dan Kaminsky*
11050 * Delete MD2 from algorithm tables. This follows the recommendation in
11051 several standards that it is not used in new applications due to
11052 several cryptographic weaknesses. For binary compatibility reasons
11053 the MD2 API is still compiled in by default.
11057 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
11062 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
11063 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
11066 *Guenter <lists@gknw.net>*
11068 * Fix the server certificate chain building code to use X509_verify_cert(),
11069 it used to have an ad-hoc builder which was unable to cope with anything
11070 other than a simple chain.
11072 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
11074 * Don't check self signed certificate signatures in X509_verify_cert()
11075 by default (a flag can override this): it just wastes time without
11076 adding any security. As a useful side effect self signed root CAs
11077 with non-FIPS digests are now usable in FIPS mode.
11081 * In dtls1_process_out_of_seq_message() the check if the current message
11082 is already buffered was missing. For every new message was memory
11083 allocated, allowing an attacker to perform an denial of service attack
11084 with sending out of seq handshake messages until there is no memory
11085 left. Additionally every future message was buffered, even if the
11086 sequence number made no sense and would be part of another handshake.
11087 So only messages with sequence numbers less than 10 in advance will be
11088 buffered. ([CVE-2009-1378])
11090 *Robin Seggelmann, discovered by Daniel Mentz*
11092 * Records are buffered if they arrive with a future epoch to be
11093 processed after finishing the corresponding handshake. There is
11094 currently no limitation to this buffer allowing an attacker to perform
11095 a DOS attack with sending records with future epochs until there is no
11096 memory left. This patch adds the pqueue_size() function to determine
11097 the size of a buffer and limits the record buffer to 100 entries.
11100 *Robin Seggelmann, discovered by Daniel Mentz*
11102 * Keep a copy of frag->msg_header.frag_len so it can be used after the
11103 parent structure is freed. ([CVE-2009-1379])
11107 * Handle non-blocking I/O properly in SSL_shutdown() call.
11109 *Darryl Miles <darryl-mailinglists@netbauds.net>*
11111 * Add `2.5.4.*` OIDs
11113 *Ilya O. <vrghost@gmail.com>*
11115 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
11117 * Disable renegotiation completely - this fixes a severe security
11118 problem ([CVE-2009-3555]) at the cost of breaking all
11119 renegotiation. Renegotiation can be re-enabled by setting
11120 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
11121 run-time. This is really not recommended unless you know what
11126 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
11128 * Don't set val to NULL when freeing up structures, it is freed up by
11129 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
11130 zeroing past the valid field. ([CVE-2009-0789])
11132 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
11134 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
11135 checked correctly. This would allow some invalid signed attributes to
11136 appear to verify correctly. ([CVE-2009-0591])
11138 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11140 * Reject UniversalString and BMPString types with invalid lengths. This
11141 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
11142 a legal length. ([CVE-2009-0590])
11146 * Set S/MIME signing as the default purpose rather than setting it
11147 unconditionally. This allows applications to override it at the store
11152 * Permit restricted recursion of ASN1 strings. This is needed in practice
11153 to handle some structures.
11157 * Improve efficiency of mem_gets: don't search whole buffer each time
11160 *Jeremy Shapiro <jnshapir@us.ibm.com>*
11162 * New -hex option for openssl rand.
11166 * Print out UTF8String and NumericString when parsing ASN1.
11170 * Support NumericString type for name components.
11174 * Allow CC in the environment to override the automatically chosen
11175 compiler. Note that nothing is done to ensure flags work with the
11180 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
11182 * Properly check EVP_VerifyFinal() and similar return values
11185 *Ben Laurie, Bodo Moeller, Google Security Team*
11187 * Enable TLS extensions by default.
11191 * Allow the CHIL engine to be loaded, whether the application is
11192 multithreaded or not. (This does not release the developer from the
11193 obligation to set up the dynamic locking callbacks.)
11195 *Sander Temme <sander@temme.net>*
11197 * Use correct exit code if there is an error in dgst command.
11199 *Steve Henson; problem pointed out by Roland Dirlewanger*
11201 * Tweak Configure so that you need to say "experimental-jpake" to enable
11202 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
11206 * Add experimental JPAKE support, including demo authentication in
11207 s_client and s_server.
11211 * Set the comparison function in v3_addr_canonize().
11213 *Rob Austein <sra@hactrn.net>*
11215 * Add support for XMPP STARTTLS in s_client.
11217 *Philip Paeps <philip@freebsd.org>*
11219 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
11220 to ensure that even with this option, only ciphersuites in the
11221 server's preference list will be accepted. (Note that the option
11222 applies only when resuming a session, so the earlier behavior was
11223 just about the algorithm choice for symmetric cryptography.)
11227 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
11229 * Fix NULL pointer dereference if a DTLS server received
11230 ChangeCipherSpec as first record ([CVE-2009-1386]).
11234 * Fix a state transition in s3_srvr.c and d1_srvr.c
11235 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
11237 *Nagendra Modadugu*
11239 * The fix in 0.9.8c that supposedly got rid of unsafe
11240 double-checked locking was incomplete for RSA blinding,
11241 addressing just one layer of what turns out to have been
11242 doubly unsafe triple-checked locking.
11244 So now fix this for real by retiring the MONT_HELPER macro
11245 in crypto/rsa/rsa_eay.c.
11247 *Bodo Moeller; problem pointed out by Marius Schilder*
11249 * Various precautionary measures:
11251 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
11253 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
11254 (NB: This would require knowledge of the secret session ticket key
11255 to exploit, in which case you'd be SOL either way.)
11257 - Change bn_nist.c so that it will properly handle input BIGNUMs
11258 outside the expected range.
11260 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11263 *Neel Mehta, Bodo Moeller*
11265 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11266 the load fails. Useful for distros.
11268 *Ben Laurie and the FreeBSD team*
11270 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11274 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11278 * Expand ENGINE to support engine supplied SSL client certificate functions.
11280 This work was sponsored by Logica.
11284 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11285 keystores. Support for SSL/TLS client authentication too.
11286 Not compiled unless enable-capieng specified to Configure.
11288 This work was sponsored by Logica.
11292 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11293 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11294 attribute creation routines such as certificate requests and PKCS#12
11299 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11301 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11302 handshake which could lead to a client crash as found using the
11303 Codenomicon TLS test suite ([CVE-2008-1672])
11305 *Steve Henson, Mark Cox*
11307 * Fix double free in TLS server name extensions which could lead to
11308 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11312 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11314 Clear the error queue to ensure that error entries left from
11315 older function calls do not interfere with the correct operation.
11317 *Lutz Jaenicke, Erik de Castro Lopo*
11319 * Remove root CA certificates of commercial CAs:
11321 The OpenSSL project does not recommend any specific CA and does not
11322 have any policy with respect to including or excluding any CA.
11323 Therefore, it does not make any sense to ship an arbitrary selection
11324 of root CA certificates with the OpenSSL software.
11328 * RSA OAEP patches to fix two separate invalid memory reads.
11329 The first one involves inputs when 'lzero' is greater than
11330 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11331 before the beginning of from). The second one involves inputs where
11332 the 'db' section contains nothing but zeroes (there is a one-byte
11333 invalid read after the end of 'db').
11335 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11337 * Partial backport from 0.9.9-dev:
11339 Introduce bn_mul_mont (dedicated Montgomery multiplication
11340 procedure) as a candidate for BIGNUM assembler implementation.
11341 While 0.9.9-dev uses assembler for various architectures, only
11342 x86_64 is available by default here in the 0.9.8 branch, and
11343 32-bit x86 is available through a compile-time setting.
11345 To try the 32-bit x86 assembler implementation, use Configure
11346 option "enable-montasm" (which exists only for this backport).
11348 As "enable-montasm" for 32-bit x86 disclaims code stability
11349 anyway, in this constellation we activate additional code
11350 backported from 0.9.9-dev for further performance improvements,
11351 namely BN_from_montgomery_word. (To enable this otherwise,
11352 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11354 *Andy Polyakov (backport partially by Bodo Moeller)*
11356 * Add TLS session ticket callback. This allows an application to set
11357 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11358 values. This is useful for key rollover for example where several key
11359 sets may exist with different names.
11363 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11364 This was broken until now in 0.9.8 releases, such that the only way
11365 a registered ENGINE could be used (assuming it initialises
11366 successfully on the host) was to explicitly set it as the default
11367 for the relevant algorithms. This is in contradiction with 0.9.7
11368 behaviour and the documentation. With this fix, when an ENGINE is
11369 registered into a given algorithm's table of implementations, the
11370 'uptodate' flag is reset so that auto-discovery will be used next
11371 time a new context for that algorithm attempts to select an
11374 *Ian Lister (tweaked by Geoff Thorpe)*
11376 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11377 implementation in the following ways:
11379 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11382 Lack of BER streaming support means one pass streaming processing is
11383 only supported if data is detached: setting the streaming flag is
11384 ignored for embedded content.
11386 CMS support is disabled by default and must be explicitly enabled
11387 with the enable-cms configuration option.
11391 * Update the GMP engine glue to do direct copies between BIGNUM and
11392 mpz_t when openssl and GMP use the same limb size. Otherwise the
11393 existing "conversion via a text string export" trick is still used.
11395 *Paul Sheer <paulsheer@gmail.com>*
11397 * Zlib compression BIO. This is a filter BIO which compressed and
11398 uncompresses any data passed through it.
11402 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11403 RFC3394 compatible AES key wrapping.
11407 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11408 sets string data without copying. X509_ALGOR_set0() and
11409 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11410 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11411 from an X509_ATTRIBUTE structure optionally checking it occurs only
11412 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11417 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11418 to get the expected BN_FLG_CONSTTIME behavior.
11420 *Bodo Moeller (Google)*
11424 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11425 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11426 - added some more tests to do_tests.pl
11427 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11428 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11429 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11430 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11431 - various changes to netware.pl to enable gcc-cross builds on Win32
11433 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11434 - various changes to fix missing prototype warnings
11435 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11436 - added AES, WHIRLPOOL and CPUID assembler code to build files
11437 - added missing AES assembler make rules to mk1mf.pl
11438 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11440 *Guenter Knauf <eflash@gmx.net>*
11442 * Implement certificate status request TLS extension defined in RFC3546.
11443 A client can set the appropriate parameters and receive the encoded
11444 OCSP response via a callback. A server can query the supplied parameters
11445 and set the encoded OCSP response in the callback. Add simplified examples
11446 to s_client and s_server.
11450 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11452 * Fix various bugs:
11453 + Binary incompatibility of ssl_ctx_st structure
11454 + DTLS interoperation with non-compliant servers
11455 + Don't call get_session_cb() without proposed session
11456 + Fix ia64 assembler code
11458 *Andy Polyakov, Steve Henson*
11460 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11462 * DTLS Handshake overhaul. There were longstanding issues with
11463 OpenSSL DTLS implementation, which were making it impossible for
11464 RFC 4347 compliant client to communicate with OpenSSL server.
11465 Unfortunately just fixing these incompatibilities would "cut off"
11466 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11467 server keeps tolerating non RFC compliant syntax. The opposite is
11468 not true, 0.9.8f client can not communicate with earlier server.
11469 This update even addresses CVE-2007-4995.
11473 * Changes to avoid need for function casts in OpenSSL: some compilers
11474 (gcc 4.2 and later) reject their use.
11475 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11478 * Add RFC4507 support to OpenSSL. This includes the corrections in
11479 RFC4507bis. The encrypted ticket format is an encrypted encoded
11480 SSL_SESSION structure, that way new session features are automatically
11483 If a client application caches session in an SSL_SESSION structure
11484 support is transparent because tickets are now stored in the encoded
11487 The SSL_CTX structure automatically generates keys for ticket
11488 protection in servers so again support should be possible
11489 with no application modification.
11491 If a client or server wishes to disable RFC4507 support then the option
11492 SSL_OP_NO_TICKET can be set.
11494 Add a TLS extension debugging callback to allow the contents of any client
11495 or server extensions to be examined.
11497 This work was sponsored by Google.
11501 * Add initial support for TLS extensions, specifically for the server_name
11502 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11503 have new members for a hostname. The SSL data structure has an
11504 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11505 stored in that context to allow for session resumption, even after the
11506 SSL has been switched to a new SSL_CTX in reaction to a client's
11507 server_name extension.
11509 New functions (subject to change):
11511 SSL_get_servername()
11512 SSL_get_servername_type()
11515 New CTRL codes and macros (subject to change):
11517 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11518 - SSL_CTX_set_tlsext_servername_callback()
11519 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11520 - SSL_CTX_set_tlsext_servername_arg()
11521 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11523 openssl s_client has a new '-servername ...' option.
11525 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11526 '-key2 ...', '-servername_fatal' (subject to change). This allows
11527 testing the HostName extension for a specific single hostname ('-cert'
11528 and '-key' remain fallbacks for handshakes without HostName
11529 negotiation). If the unrecognized_name alert has to be sent, this by
11530 default is a warning; it becomes fatal with the '-servername_fatal'
11533 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11535 * Add AES and SSE2 assembly language support to VC++ build.
11539 * Mitigate attack on final subtraction in Montgomery reduction.
11543 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11544 (which previously caused an internal error).
11548 * Squeeze another 10% out of IGE mode when in != out.
11552 * AES IGE mode speedup.
11554 *Dean Gaudet (Google)*
11556 * Add the Korean symmetric 128-bit cipher SEED (see
11557 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11558 add SEED ciphersuites from RFC 4162:
11560 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11561 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11562 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11563 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11565 To minimize changes between patchlevels in the OpenSSL 0.9.8
11566 series, SEED remains excluded from compilation unless OpenSSL
11567 is configured with 'enable-seed'.
11569 *KISA, Bodo Moeller*
11571 * Mitigate branch prediction attacks, which can be practical if a
11572 single processor is shared, allowing a spy process to extract
11573 information. For detailed background information, see
11574 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11575 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11576 and Necessary Software Countermeasures"). The core of the change
11577 are new versions BN_div_no_branch() and
11578 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11579 respectively, which are slower, but avoid the security-relevant
11580 conditional branches. These are automatically called by BN_div()
11581 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11582 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11583 remove a conditional branch.
11585 BN_FLG_CONSTTIME is the new name for the previous
11586 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11587 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11588 in the exponent causes BN_mod_exp_mont() to use the alternative
11589 implementation in BN_mod_exp_mont_consttime().) The old name
11590 remains as a deprecated alias.
11592 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11593 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11594 constant-time implementations for more than just exponentiation.
11595 Here too the old name is kept as a deprecated alias.
11597 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11598 the BN_BLINDING structure gets an independent copy of the
11599 modulus. This means that the previous `BIGNUM *m` argument to
11600 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11601 essentially becomes `const BIGNUM *m`, although we can't actually
11602 change this in the header file before 0.9.9. It allows
11603 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11604 enable BN_FLG_CONSTTIME.
11606 *Matthew D Wood (Intel Corp)*
11608 * In the SSL/TLS server implementation, be strict about session ID
11609 context matching (which matters if an application uses a single
11610 external cache for different purposes). Previously,
11611 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11612 set. This did ensure strict client verification, but meant that,
11613 with applications using a single external cache for quite
11614 different requirements, clients could circumvent ciphersuite
11615 restrictions for a given session ID context by starting a session
11616 in a different context.
11620 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11621 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11622 authentication-only ciphersuites.
11626 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11627 not complete and could lead to a possible single byte overflow
11628 ([CVE-2007-5135]) [Ben Laurie]
11630 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11632 * Since AES128 and AES256 (and similarly Camellia128 and
11633 Camellia256) share a single mask bit in the logic of
11634 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11635 kludge to work properly if AES128 is available and AES256 isn't
11636 (or if Camellia128 is available and Camellia256 isn't).
11640 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11641 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11642 When a point or a seed is encoded in a BIT STRING, we need to
11643 prevent the removal of trailing zero bits to get the proper DER
11644 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11645 of a NamedBitList, for which trailing 0 bits need to be removed.)
11649 * Have SSL/TLS server implementation tolerate "mismatched" record
11650 protocol version while receiving ClientHello even if the
11651 ClientHello is fragmented. (The server can't insist on the
11652 particular protocol version it has chosen before the ServerHello
11653 message has informed the client about his choice.)
11657 * Add RFC 3779 support.
11659 *Rob Austein for ARIN, Ben Laurie*
11661 * Load error codes if they are not already present instead of using a
11662 static variable. This allows them to be cleanly unloaded and reloaded.
11663 Improve header file function name parsing.
11667 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11668 or CAPABILITY handshake as required by RFCs.
11670 *Goetz Babin-Ebell*
11672 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11674 * Introduce limits to prevent malicious keys being able to
11675 cause a denial of service. ([CVE-2006-2940])
11677 *Steve Henson, Bodo Moeller*
11679 * Fix ASN.1 parsing of certain invalid structures that can result
11680 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11682 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11683 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11685 * Fix SSL client code which could crash if connecting to a
11686 malicious SSLv2 server. ([CVE-2006-4343])
11688 *Tavis Ormandy and Will Drewry, Google Security Team*
11690 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11691 match only those. Before that, "AES256-SHA" would be interpreted
11692 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11693 the same strength classification in 0.9.7h) as we currently only
11694 have a single AES bit in the ciphersuite description bitmap.
11695 That change, however, also applied to ciphersuite strings such as
11696 "RC4-MD5" that intentionally matched multiple ciphersuites --
11697 namely, SSL 2.0 ciphersuites in addition to the more common ones
11698 from SSL 3.0/TLS 1.0.
11700 So we change the selection algorithm again: Naming an explicit
11701 ciphersuite selects this one ciphersuite, and any other similar
11702 ciphersuite (same bitmap) from *other* protocol versions.
11703 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11704 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11706 Since SSL 2.0 does not have any ciphersuites for which the
11707 128/256 bit distinction would be relevant, this works for now.
11708 The proper fix will be to use different bits for AES128 and
11709 AES256, which would have avoided the problems from the beginning;
11710 however, bits are scarce, so we can only do this in a new release
11711 (not just a patchlevel) when we can change the SSL_CIPHER
11712 definition to split the single 'unsigned long mask' bitmap into
11713 multiple values to extend the available space.
11717 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11719 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11720 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11722 * Add AES IGE and biIGE modes.
11726 * Change the Unix randomness entropy gathering to use poll() when
11727 possible instead of select(), since the latter has some
11728 undesirable limitations.
11730 *Darryl Miles via Richard Levitte and Bodo Moeller*
11732 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11733 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11734 cannot be implicitly activated as part of, e.g., the "AES" alias.
11735 However, please upgrade to OpenSSL 0.9.9[-dev] for
11736 non-experimental use of the ECC ciphersuites to get TLS extension
11737 support, which is required for curve and point format negotiation
11738 to avoid potential handshake problems.
11742 * Disable rogue ciphersuites:
11744 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11745 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11746 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11748 The latter two were purportedly from
11749 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11752 Also deactivate the remaining ciphersuites from
11753 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11754 unofficial, and the ID has long expired.
11758 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11759 dual-core machines) and other potential thread-safety issues.
11763 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11764 versions), which is now available for royalty-free use
11765 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11766 Also, add Camellia TLS ciphersuites from RFC 4132.
11768 To minimize changes between patchlevels in the OpenSSL 0.9.8
11769 series, Camellia remains excluded from compilation unless OpenSSL
11770 is configured with 'enable-camellia'.
11774 * Disable the padding bug check when compression is in use. The padding
11775 bug check assumes the first packet is of even length, this is not
11776 necessarily true if compression is enabled and can result in false
11777 positives causing handshake failure. The actual bug test is ancient
11778 code so it is hoped that implementations will either have fixed it by
11779 now or any which still have the bug do not support compression.
11783 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11785 * When applying a cipher rule check to see if string match is an explicit
11786 cipher suite and only match that one cipher suite if it is.
11790 * Link in manifests for VC++ if needed.
11792 *Austin Ziegler <halostatue@gmail.com>*
11794 * Update support for ECC-based TLS ciphersuites according to
11795 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11796 TLS extensions, which are supported starting with the 0.9.9
11797 branch, not in the OpenSSL 0.9.8 branch).
11801 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11802 opaque EVP_CIPHER_CTX handling.
11806 * Fixes and enhancements to zlib compression code. We now only use
11807 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11808 to conform with the standards mentioned here:
11809 <http://www.zlib.net/DLL_FAQ.txt>
11810 Static zlib linking now works on Windows and the new --with-zlib-include
11811 --with-zlib-lib options to Configure can be used to supply the location
11812 of the headers and library. Gracefully handle case where zlib library
11817 * Several fixes and enhancements to the OID generation code. The old code
11818 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11819 handle numbers larger than ULONG_MAX, truncated printing and had a
11820 non standard OBJ_obj2txt() behaviour.
11824 * Add support for building of engines under engine/ as shared libraries
11825 under VC++ build system.
11829 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11830 Hopefully, we will not see any false combination of paths any more.
11834 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11836 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11837 (part of SSL_OP_ALL). This option used to disable the
11838 countermeasure against man-in-the-middle protocol-version
11839 rollback in the SSL 2.0 server implementation, which is a bad
11840 idea. ([CVE-2005-2969])
11842 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11843 for Information Security, National Institute of Advanced Industrial
11844 Science and Technology [AIST], Japan)*
11846 * Add two function to clear and return the verify parameter flags.
11850 * Keep cipherlists sorted in the source instead of sorting them at
11851 runtime, thus removing the need for a lock.
11855 * Avoid some small subgroup attacks in Diffie-Hellman.
11857 *Nick Mathewson and Ben Laurie*
11859 * Add functions for well-known primes.
11863 * Extended Windows CE support.
11865 *Satoshi Nakamura and Andy Polyakov*
11867 * Initialize SSL_METHOD structures at compile time instead of during
11868 runtime, thus removing the need for a lock.
11872 * Make PKCS7_decrypt() work even if no certificate is supplied by
11873 attempting to decrypt each encrypted key in turn. Add support to
11878 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11880 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11883 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11887 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11888 key into the same file any more.
11892 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11896 * Add -utf8 command line and config file option to 'ca'.
11898 *Stefan <stf@udoma.org*
11900 * Removed the macro des_crypt(), as it seems to conflict with some
11901 libraries. Use DES_crypt().
11905 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11906 involves renaming the source and generated shared-libs for
11907 both. The engines will accept the corrected or legacy ids
11908 ('ncipher' and '4758_cca' respectively) when binding. NB,
11909 this only applies when building 'shared'.
11911 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11913 * Add attribute functions to EVP_PKEY structure. Modify
11914 PKCS12_create() to recognize a CSP name attribute and
11915 use it. Make -CSP option work again in pkcs12 utility.
11919 * Add new functionality to the bn blinding code:
11920 - automatic re-creation of the BN_BLINDING parameters after
11921 a fixed number of uses (currently 32)
11922 - add new function for parameter creation
11923 - introduce flags to control the update behaviour of the
11924 BN_BLINDING parameters
11925 - hide BN_BLINDING structure
11926 Add a second BN_BLINDING slot to the RSA structure to improve
11927 performance when a single RSA object is shared among several
11932 * Add support for DTLS.
11934 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11936 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11937 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11941 * Remove buggy and incomplete DH cert support from
11942 ssl/ssl_rsa.c and ssl/s3_both.c
11946 * Use SHA-1 instead of MD5 as the default digest algorithm for
11947 the `apps/openssl` commands.
11951 * Compile clean with "-Wall -Wmissing-prototypes
11952 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11953 DEBUG_SAFESTACK must also be set.
11957 * Change ./Configure so that certain algorithms can be disabled by default.
11958 The new counterpiece to "no-xxx" is "enable-xxx".
11960 The patented RC5 and MDC2 algorithms will now be disabled unless
11961 "enable-rc5" and "enable-mdc2", respectively, are specified.
11963 (IDEA remains enabled despite being patented. This is because IDEA
11964 is frequently required for interoperability, and there is no license
11965 fee for non-commercial use. As before, "no-idea" can be used to
11966 avoid this algorithm.)
11970 * Add processing of proxy certificates (see RFC 3820). This work was
11971 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11972 EGEE (Enabling Grids for E-science in Europe).
11976 * RC4 performance overhaul on modern architectures/implementations, such
11977 as Intel P4, IA-64 and AMD64.
11981 * New utility extract-section.pl. This can be used specify an alternative
11982 section number in a pod file instead of having to treat each file as
11983 a separate case in Makefile. This can be done by adding two lines to the
11986 =for comment openssl_section:XXX
11988 The blank line is mandatory.
11992 * New arguments -certform, -keyform and -pass for s_client and s_server
11993 to allow alternative format key and certificate files and passphrase
11998 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11999 update associated structures and add various utility functions.
12001 Add new policy related verify parameters, include policy checking in
12002 standard verify code. Enhance 'smime' application with extra parameters
12003 to support policy checking and print out.
12007 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
12008 Nehemiah processors. These extensions support AES encryption in hardware
12009 as well as RNG (though RNG support is currently disabled).
12011 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
12013 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
12017 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
12019 *Andy Polyakov and a number of other people*
12021 * Improved PowerPC platform support. Most notably BIGNUM assembler
12022 implementation contributed by IBM.
12024 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
12026 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
12027 exponent rather than 'unsigned long'. There is a corresponding change to
12028 the new 'rsa_keygen' element of the RSA_METHOD structure.
12030 *Jelte Jansen, Geoff Thorpe*
12032 * Functionality for creating the initial serial number file is now
12033 moved from CA.pl to the 'ca' utility with a new option -create_serial.
12035 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
12036 number file to 1, which is bound to cause problems. To avoid
12037 the problems while respecting compatibility between different 0.9.7
12038 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
12039 CA.pl for serial number initialization. With the new release 0.9.8,
12040 we can fix the problem directly in the 'ca' utility.)
12044 * Reduced header interdependencies by declaring more opaque objects in
12045 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
12046 give fewer recursive includes, which could break lazy source code - so
12047 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
12048 developers should define this symbol when building and using openssl to
12049 ensure they track the recommended behaviour, interfaces, [etc], but
12050 backwards-compatible behaviour prevails when this isn't defined.
12054 * New function X509_POLICY_NODE_print() which prints out policy nodes.
12058 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
12059 This will generate a random key of the appropriate length based on the
12060 cipher context. The EVP_CIPHER can provide its own random key generation
12061 routine to support keys of a specific form. This is used in the des and
12062 3des routines to generate a key of the correct parity. Update S/MIME
12063 code to use new functions and hence generate correct parity DES keys.
12064 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
12065 valid (weak or incorrect parity).
12069 * Add a local set of CRLs that can be used by X509_verify_cert() as well
12070 as looking them up. This is useful when the verified structure may contain
12071 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
12072 present unless the new PKCS7_NO_CRL flag is asserted.
12076 * Extend ASN1 oid configuration module. It now additionally accepts the
12079 shortName = some long name, 1.2.3.4
12083 * Reimplemented the BN_CTX implementation. There is now no more static
12084 limitation on the number of variables it can handle nor the depth of the
12085 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
12086 information can now expand as required, and rather than having a single
12087 static array of bignums, BN_CTX now uses a linked-list of such arrays
12088 allowing it to expand on demand whilst maintaining the usefulness of
12089 BN_CTX's "bundling".
12093 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
12094 to allow all RSA operations to function using a single BN_CTX.
12098 * Preliminary support for certificate policy evaluation and checking. This
12099 is initially intended to pass the tests outlined in "Conformance Testing
12100 of Relying Party Client Certificate Path Processing Logic" v1.07.
12104 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
12105 remained unused and not that useful. A variety of other little bignum
12106 tweaks and fixes have also been made continuing on from the audit (see
12111 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
12112 associated ASN1, EVP and SSL functions and old ASN1 macros.
12116 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
12117 and this should never fail. So the return value from the use of
12118 BN_set_word() (which can fail due to needless expansion) is now deprecated;
12119 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
12123 * BN_CTX_get() should return zero-valued bignums, providing the same
12124 initialised value as BN_new().
12126 *Geoff Thorpe, suggested by Ulf Möller*
12128 * Support for inhibitAnyPolicy certificate extension.
12132 * An audit of the BIGNUM code is underway, for which debugging code is
12133 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
12134 is considered valid when processing BIGNUMs, and causes execution to
12135 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
12136 further steps are taken to deliberately pollute unused data in BIGNUM
12137 structures to try and expose faulty code further on. For now, openssl will
12138 (in its default mode of operation) continue to tolerate the inconsistent
12139 forms that it has tolerated in the past, but authors and packagers should
12140 consider trying openssl and their own applications when compiled with
12141 these debugging symbols defined. It will help highlight potential bugs in
12142 their own code, and will improve the test coverage for OpenSSL itself. At
12143 some point, these tighter rules will become openssl's default to improve
12144 maintainability, though the assert()s and other overheads will remain only
12145 in debugging configurations. See bn.h for more details.
12147 *Geoff Thorpe, Nils Larsch, Ulf Möller*
12149 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
12150 that can only be obtained through BN_CTX_new() (which implicitly
12151 initialises it). The presence of this function only made it possible
12152 to overwrite an existing structure (and cause memory leaks).
12156 * Because of the callback-based approach for implementing LHASH as a
12157 template type, lh_insert() adds opaque objects to hash-tables and
12158 lh_doall() or lh_doall_arg() are typically used with a destructor callback
12159 to clean up those corresponding objects before destroying the hash table
12160 (and losing the object pointers). So some over-zealous constifications in
12161 LHASH have been relaxed so that lh_insert() does not take (nor store) the
12162 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
12163 prototyped to have "const" restrictions on the object pointers they are
12164 given (and so aren't required to cast them away any more).
12168 * The tmdiff.h API was so ugly and minimal that our own timing utility
12169 (speed) prefers to use its own implementation. The two implementations
12170 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
12171 its object type properly exposed (MS_TM) instead of casting to/from
12172 `char *`. This may still change yet if someone realises MS_TM and
12174 aren't necessarily the greatest nomenclatures - but this is what was used
12175 internally to the implementation so I've used that for now.
12179 * Ensure that deprecated functions do not get compiled when
12180 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
12181 the self-tests were still using deprecated key-generation functions so
12182 these have been updated also.
12186 * Reorganise PKCS#7 code to separate the digest location functionality
12187 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
12188 New function PKCS7_set_digest() to set the digest type for PKCS#7
12189 digestedData type. Add additional code to correctly generate the
12190 digestedData type and add support for this type in PKCS7 initialization
12195 * New function PKCS7_set0_type_other() this initializes a PKCS7
12196 structure of type "other".
12200 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
12201 sure the loop does correctly stop and breaking ("division by zero")
12202 modulus operations are not performed. The (pre-generated) prime
12203 table crypto/bn/bn_prime.h was already correct, but it could not be
12204 re-generated on some platforms because of the "division by zero"
12205 situation in the script.
12207 *Ralf S. Engelschall*
12209 * Update support for ECC-based TLS ciphersuites according to
12210 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
12211 SHA-1 now is only used for "small" curves (where the
12212 representation of a field element takes up to 24 bytes); for
12213 larger curves, the field element resulting from ECDH is directly
12214 used as premaster secret.
12216 *Douglas Stebila (Sun Microsystems Laboratories)*
12218 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
12219 curve secp160r1 to the tests.
12221 *Douglas Stebila (Sun Microsystems Laboratories)*
12223 * Add the possibility to load symbols globally with DSO.
12225 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
12227 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
12228 control of the error stack.
12232 * Add support for STORE in ENGINE.
12236 * Add the STORE type. The intention is to provide a common interface
12237 to certificate and key stores, be they simple file-based stores, or
12238 HSM-type store, or LDAP stores, or...
12239 NOTE: The code is currently UNTESTED and isn't really used anywhere.
12243 * Add a generic structure called OPENSSL_ITEM. This can be used to
12244 pass a list of arguments to any function as well as provide a way
12245 for a function to pass data back to the caller.
12249 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
12250 works like BUF_strdup() but can be used to duplicate a portion of
12251 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
12256 * Add the function sk_find_ex() which works like sk_find(), but will
12257 return an index to an element even if an exact match couldn't be
12258 found. The index is guaranteed to point at the element where the
12259 searched-for key would be inserted to preserve sorting order.
12263 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12264 takes an extra flags argument for optional functionality. Currently,
12265 the following flags are defined:
12267 OBJ_BSEARCH_VALUE_ON_NOMATCH
12268 This one gets OBJ_bsearch_ex() to return a pointer to the first
12269 element where the comparing function returns a negative or zero
12272 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12273 This one gets OBJ_bsearch_ex() to return a pointer to the first
12274 element where the comparing function returns zero. This is useful
12275 if there are more than one element where the comparing function
12280 * Make it possible to create self-signed certificates with 'openssl ca'
12281 in such a way that the self-signed certificate becomes part of the
12282 CA database and uses the same mechanisms for serial number generation
12283 as all other certificate signing. The new flag '-selfsign' enables
12284 this functionality. Adapt CA.sh and CA.pl.in.
12288 * Add functionality to check the public key of a certificate request
12289 against a given private. This is useful to check that a certificate
12290 request can be signed by that key (self-signing).
12294 * Make it possible to have multiple active certificates with the same
12295 subject in the CA index file. This is done only if the keyword
12296 'unique_subject' is set to 'no' in the main CA section (default
12297 if 'CA_default') of the configuration file. The value is saved
12298 with the database itself in a separate index attribute file,
12299 named like the index file with '.attr' appended to the name.
12303 * Generate multi-valued AVAs using '+' notation in config files for
12308 * Support for nameConstraints certificate extension.
12312 * Support for policyConstraints certificate extension.
12316 * Support for policyMappings certificate extension.
12320 * Make sure the default DSA_METHOD implementation only uses its
12321 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12322 and change its own handlers to be NULL so as to remove unnecessary
12323 indirection. This lets alternative implementations fallback to the
12324 default implementation more easily.
12328 * Support for directoryName in GeneralName related extensions
12333 * Make it possible to link applications using Makefile.shared.
12334 Make that possible even when linking against static libraries!
12338 * Support for single pass processing for S/MIME signing. This now
12339 means that S/MIME signing can be done from a pipe, in addition
12340 cleartext signing (multipart/signed type) is effectively streaming
12341 and the signed data does not need to be all held in memory.
12343 This is done with a new flag PKCS7_STREAM. When this flag is set
12344 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12345 is done after the data is output (and digests calculated) in
12346 SMIME_write_PKCS7().
12350 * Add full support for -rpath/-R, both in shared libraries and
12351 applications, at least on the platforms where it's known how
12356 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12357 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12358 will now compute a table of multiples of the generator that
12359 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12360 faster (notably in the case of a single point multiplication,
12361 scalar * generator).
12363 *Nils Larsch, Bodo Moeller*
12365 * IPv6 support for certificate extensions. The various extensions
12366 which use the IP:a.b.c.d can now take IPv6 addresses using the
12367 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12372 * Added an ENGINE that implements RSA by performing private key
12373 exponentiations with the GMP library. The conversions to and from
12374 GMP's mpz_t format aren't optimised nor are any montgomery forms
12375 cached, and on x86 it appears OpenSSL's own performance has caught up.
12376 However there are likely to be other architectures where GMP could
12377 provide a boost. This ENGINE is not built in by default, but it can be
12378 specified at Configure time and should be accompanied by the necessary
12379 linker additions, eg;
12380 ./config -DOPENSSL_USE_GMP -lgmp
12384 * "openssl engine" will not display ENGINE/DSO load failure errors when
12385 testing availability of engines with "-t" - the old behaviour is
12386 produced by increasing the feature's verbosity with "-tt".
12390 * ECDSA routines: under certain error conditions uninitialized BN objects
12391 could be freed. Solution: make sure initialization is performed early
12392 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12397 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12398 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12399 software implementations. For DSA and DH, parameter generation can
12400 also be overridden by providing the appropriate method callbacks.
12404 * Change the "progress" mechanism used in key-generation and
12405 primality testing to functions that take a new BN_GENCB pointer in
12406 place of callback/argument pairs. The new API functions have `_ex`
12407 postfixes and the older functions are reimplemented as wrappers for
12408 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12409 declarations of the old functions to help (graceful) attempts to
12410 migrate to the new functions. Also, the new key-generation API
12411 functions operate on a caller-supplied key-structure and return
12412 success/failure rather than returning a key or NULL - this is to
12413 help make "keygen" another member function of RSA_METHOD etc.
12415 Example for using the new callback interface:
12417 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12418 void *my_arg = ...;
12421 BN_GENCB_set(&my_cb, my_callback, my_arg);
12423 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12424 /* For the meaning of a, b in calls to my_callback(), see the
12425 * documentation of the function that calls the callback.
12426 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12427 * my_callback should return 1 if it wants BN_is_prime_ex()
12428 * to continue, or 0 to stop.
12433 * Change the ZLIB compression method to be stateful, and make it
12434 available to TLS with the number defined in
12435 draft-ietf-tls-compression-04.txt.
12439 * Add the ASN.1 structures and functions for CertificatePair, which
12440 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12442 CertificatePair ::= SEQUENCE {
12443 forward [0] Certificate OPTIONAL,
12444 reverse [1] Certificate OPTIONAL,
12445 -- at least one of the pair shall be present -- }
12447 Also implement the PEM functions to read and write certificate
12448 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12450 This needed to be defined, mostly for the sake of the LDAP
12451 attribute crossCertificatePair, but may prove useful elsewhere as
12456 * Make it possible to inhibit symlinking of shared libraries in
12457 Makefile.shared, for Cygwin's sake.
12461 * Extend the BIGNUM API by creating a function
12462 void BN_set_negative(BIGNUM *a, int neg);
12463 and a macro that behave like
12464 int BN_is_negative(const BIGNUM *a);
12466 to avoid the need to access 'a->neg' directly in applications.
12470 * Implement fast modular reduction for pseudo-Mersenne primes
12471 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12472 EC_GROUP_new_curve_GFp() will now automatically use this
12475 *Nils Larsch <nla@trustcenter.de>*
12477 * Add new lock type (CRYPTO_LOCK_BN).
12481 * Change the ENGINE framework to automatically load engines
12482 dynamically from specific directories unless they could be
12483 found to already be built in or loaded. Move all the
12484 current engines except for the cryptodev one to a new
12485 directory engines/.
12486 The engines in engines/ are built as shared libraries if
12487 the "shared" options was given to ./Configure or ./config.
12488 Otherwise, they are inserted in libcrypto.a.
12489 /usr/local/ssl/engines is the default directory for dynamic
12490 engines, but that can be overridden at configure time through
12491 the usual use of --prefix and/or --openssldir, and at run
12492 time with the environment variable OPENSSL_ENGINES.
12494 *Geoff Thorpe and Richard Levitte*
12496 * Add Makefile.shared, a helper makefile to build shared
12497 libraries. Adapt Makefile.org.
12501 * Add version info to Win32 DLLs.
12503 *Peter 'Luna' Runestig" <peter@runestig.com>*
12505 * Add new 'medium level' PKCS#12 API. Certificates and keys
12506 can be added using this API to created arbitrary PKCS#12
12507 files while avoiding the low-level API.
12509 New options to PKCS12_create(), key or cert can be NULL and
12510 will then be omitted from the output file. The encryption
12511 algorithm NIDs can be set to -1 for no encryption, the mac
12512 iteration count can be set to 0 to omit the mac.
12514 Enhance pkcs12 utility by making the -nokeys and -nocerts
12515 options work when creating a PKCS#12 file. New option -nomac
12516 to omit the mac, NONE can be set for an encryption algorithm.
12517 New code is modified to use the enhanced PKCS12_create()
12518 instead of the low-level API.
12522 * Extend ASN1 encoder to support indefinite length constructed
12523 encoding. This can output sequences tags and octet strings in
12524 this form. Modify pk7_asn1.c to support indefinite length
12525 encoding. This is experimental and needs additional code to
12526 be useful, such as an ASN1 bio and some enhanced streaming
12529 Extend template encode functionality so that tagging is passed
12530 down to the template encoder.
12534 * Let 'openssl req' fail if an argument to '-newkey' is not
12535 recognized instead of using RSA as a default.
12539 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12540 As these are not official, they are not included in "ALL";
12541 the "ECCdraft" ciphersuite group alias can be used to select them.
12543 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12545 * Add ECDH engine support.
12547 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12549 * Add ECDH in new directory crypto/ecdh/.
12551 *Douglas Stebila (Sun Microsystems Laboratories)*
12553 * Let BN_rand_range() abort with an error after 100 iterations
12554 without success (which indicates a broken PRNG).
12558 * Change BN_mod_sqrt() so that it verifies that the input value
12559 is really the square of the return value. (Previously,
12560 BN_mod_sqrt would show GIGO behaviour.)
12564 * Add named elliptic curves over binary fields from X9.62, SECG,
12565 and WAP/WTLS; add OIDs that were still missing.
12567 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12569 * Extend the EC library for elliptic curves over binary fields
12570 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12573 EC_GF2m_simple_method
12577 EC_GROUP_new_curve_GF2m
12578 EC_GROUP_set_curve_GF2m
12579 EC_GROUP_get_curve_GF2m
12580 EC_POINT_set_affine_coordinates_GF2m
12581 EC_POINT_get_affine_coordinates_GF2m
12582 EC_POINT_set_compressed_coordinates_GF2m
12584 Point compression for binary fields is disabled by default for
12585 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12588 As binary polynomials are represented as BIGNUMs, various members
12589 of the EC_GROUP and EC_POINT data structures can be shared
12590 between the implementations for prime fields and binary fields;
12591 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12592 are essentially identical to their `..._GFp` counterparts.
12593 (For simplicity, the `..._GFp` prefix has been dropped from
12594 various internal method names.)
12596 An internal 'field_div' method (similar to 'field_mul' and
12597 'field_sqr') has been added; this is used only for binary fields.
12599 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12601 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12602 through methods ('mul', 'precompute_mult').
12604 The generic implementations (now internally called 'ec_wNAF_mul'
12605 and 'ec_wNAF_precomputed_mult') remain the default if these
12606 methods are undefined.
12608 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12610 * New function EC_GROUP_get_degree, which is defined through
12611 EC_METHOD. For curves over prime fields, this returns the bit
12612 length of the modulus.
12614 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12616 * New functions EC_GROUP_dup, EC_POINT_dup.
12617 (These simply call ..._new and ..._copy).
12619 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12621 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12622 Polynomials are represented as BIGNUMs (where the sign bit is not
12623 used) in the following functions [macros]:
12626 BN_GF2m_sub [= BN_GF2m_add]
12627 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12628 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12629 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12631 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12632 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12633 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12634 BN_GF2m_cmp [= BN_ucmp]
12636 (Note that only the 'mod' functions are actually for fields GF(2^m).
12637 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12639 For some functions, an the irreducible polynomial defining a
12640 field can be given as an 'unsigned int[]' with strictly
12641 decreasing elements giving the indices of those bits that are set;
12642 i.e., p[] represents the polynomial
12643 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12645 p[0] > p[1] > ... > p[k] = 0.
12646 This applies to the following functions:
12649 BN_GF2m_mod_mul_arr
12650 BN_GF2m_mod_sqr_arr
12651 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12652 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12653 BN_GF2m_mod_exp_arr
12654 BN_GF2m_mod_sqrt_arr
12655 BN_GF2m_mod_solve_quad_arr
12659 Conversion can be performed by the following functions:
12664 bntest.c has additional tests for binary polynomial arithmetic.
12666 Two implementations for BN_GF2m_mod_div() are available.
12667 The default algorithm simply uses BN_GF2m_mod_inv() and
12668 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12669 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12670 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12672 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12674 * Add new error code 'ERR_R_DISABLED' that can be used when some
12675 functionality is disabled at compile-time.
12677 *Douglas Stebila <douglas.stebila@sun.com>*
12679 * Change default behaviour of 'openssl asn1parse' so that more
12680 information is visible when viewing, e.g., a certificate:
12682 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12683 mode the content of non-printable OCTET STRINGs is output in a
12684 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12685 avoid the appearance of a printable string.
12687 *Nils Larsch <nla@trustcenter.de>*
12689 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12691 EC_GROUP_set_asn1_flag()
12692 EC_GROUP_get_asn1_flag()
12693 EC_GROUP_set_point_conversion_form()
12694 EC_GROUP_get_point_conversion_form()
12695 These control ASN1 encoding details:
12696 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12697 has been set to OPENSSL_EC_NAMED_CURVE.
12698 - Points are encoded in uncompressed form by default; options for
12699 asn1_for are as for point2oct, namely
12700 POINT_CONVERSION_COMPRESSED
12701 POINT_CONVERSION_UNCOMPRESSED
12702 POINT_CONVERSION_HYBRID
12704 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12706 EC_GROUP_set_seed()
12707 EC_GROUP_get0_seed()
12708 EC_GROUP_get_seed_len()
12709 This is used only for ASN1 purposes (so far).
12711 *Nils Larsch <nla@trustcenter.de>*
12713 * Add 'field_type' member to EC_METHOD, which holds the NID
12714 of the appropriate field type OID. The new function
12715 EC_METHOD_get_field_type() returns this value.
12717 *Nils Larsch <nla@trustcenter.de>*
12720 EC_POINT_point2bn()
12721 EC_POINT_bn2point()
12722 EC_POINT_point2hex()
12723 EC_POINT_hex2point()
12724 providing useful interfaces to EC_POINT_point2oct() and
12725 EC_POINT_oct2point().
12727 *Nils Larsch <nla@trustcenter.de>*
12729 * Change internals of the EC library so that the functions
12730 EC_GROUP_set_generator()
12731 EC_GROUP_get_generator()
12732 EC_GROUP_get_order()
12733 EC_GROUP_get_cofactor()
12734 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12735 to methods, which would lead to unnecessary code duplication when
12736 adding different types of curves.
12738 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12740 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12741 arithmetic, and such that modified wNAFs are generated
12742 (which avoid length expansion in many cases).
12746 * Add a function EC_GROUP_check_discriminant() (defined via
12747 EC_METHOD) that verifies that the curve discriminant is non-zero.
12749 Add a function EC_GROUP_check() that makes some sanity tests
12750 on a EC_GROUP, its generator and order. This includes
12751 EC_GROUP_check_discriminant().
12753 *Nils Larsch <nla@trustcenter.de>*
12755 * Add ECDSA in new directory crypto/ecdsa/.
12757 Add applications 'openssl ecparam' and 'openssl ecdsa'
12758 (these are based on 'openssl dsaparam' and 'openssl dsa').
12760 ECDSA support is also included in various other files across the
12761 library. Most notably,
12762 - 'openssl req' now has a '-newkey ecdsa:file' option;
12763 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12764 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12765 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12766 them suitable for ECDSA where domain parameters must be
12767 extracted before the specific public key;
12768 - ECDSA engine support has been added.
12770 *Nils Larsch <nla@trustcenter.de>*
12772 * Include some named elliptic curves, and add OIDs from X9.62,
12773 SECG, and WAP/WTLS. Each curve can be obtained from the new
12775 EC_GROUP_new_by_curve_name(),
12776 and the list of available named curves can be obtained with
12777 EC_get_builtin_curves().
12778 Also add a 'curve_name' member to EC_GROUP objects, which can be
12780 EC_GROUP_set_curve_name()
12781 EC_GROUP_get_curve_name()
12783 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12785 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12786 was actually never needed) and in BN_mul(). The removal in BN_mul()
12787 required a small change in bn_mul_part_recursive() and the addition
12788 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12789 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12790 bn_sub_words() and bn_add_words() except they take arrays with
12795 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12797 * Cleanse PEM buffers before freeing them since they may contain
12800 *Benjamin Bennett <ben@psc.edu>*
12802 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12803 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12804 authentication-only ciphersuites.
12808 * Since AES128 and AES256 share a single mask bit in the logic of
12809 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12810 kludge to work properly if AES128 is available and AES256 isn't.
12814 * Expand security boundary to match 1.1.1 module.
12818 * Remove redundant features: hash file source, editing of test vectors
12819 modify fipsld to use external fips_premain.c signature.
12823 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12824 run algorithm test programs.
12828 * Make algorithm test programs more tolerant of whitespace.
12832 * Have SSL/TLS server implementation tolerate "mismatched" record
12833 protocol version while receiving ClientHello even if the
12834 ClientHello is fragmented. (The server can't insist on the
12835 particular protocol version it has chosen before the ServerHello
12836 message has informed the client about his choice.)
12840 * Load error codes if they are not already present instead of using a
12841 static variable. This allows them to be cleanly unloaded and reloaded.
12845 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12847 * Introduce limits to prevent malicious keys being able to
12848 cause a denial of service. ([CVE-2006-2940])
12850 *Steve Henson, Bodo Moeller*
12852 * Fix ASN.1 parsing of certain invalid structures that can result
12853 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12855 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12856 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12858 * Fix SSL client code which could crash if connecting to a
12859 malicious SSLv2 server. ([CVE-2006-4343])
12861 *Tavis Ormandy and Will Drewry, Google Security Team*
12863 * Change ciphersuite string processing so that an explicit
12864 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12865 will no longer include "AES128-SHA"), and any other similar
12866 ciphersuite (same bitmap) from *other* protocol versions (so that
12867 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12868 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12869 changes from 0.9.8b and 0.9.8d.
12873 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12875 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12876 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12878 * Change the Unix randomness entropy gathering to use poll() when
12879 possible instead of select(), since the latter has some
12880 undesirable limitations.
12882 *Darryl Miles via Richard Levitte and Bodo Moeller*
12884 * Disable rogue ciphersuites:
12886 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12887 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12888 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12890 The latter two were purportedly from
12891 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12894 Also deactivate the remaining ciphersuites from
12895 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12896 unofficial, and the ID has long expired.
12900 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12901 dual-core machines) and other potential thread-safety issues.
12905 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12907 * Adapt fipsld and the build system to link against the validated FIPS
12908 module in FIPS mode.
12912 * Fixes for VC++ 2005 build under Windows.
12916 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12917 from a Windows bash shell such as MSYS. It is autodetected from the
12918 "config" script when run from a VC++ environment. Modify standard VC++
12919 build to use fipscanister.o from the GNU make build.
12923 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12925 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12926 The value now differs depending on if you build for FIPS or not.
12927 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12928 safely run with a non-FIPSed libcrypto, as it may crash because of
12929 the difference induced by this change.
12933 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12935 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12936 (part of SSL_OP_ALL). This option used to disable the
12937 countermeasure against man-in-the-middle protocol-version
12938 rollback in the SSL 2.0 server implementation, which is a bad
12939 idea. ([CVE-2005-2969])
12941 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12942 for Information Security, National Institute of Advanced Industrial
12943 Science and Technology [AIST, Japan)]*
12945 * Minimal support for X9.31 signatures and PSS padding modes. This is
12946 mainly for FIPS compliance and not fully integrated at this stage.
12950 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12951 the exponentiation using a fixed-length exponent. (Otherwise,
12952 the information leaked through timing could expose the secret key
12953 after many signatures; cf. Bleichenbacher's attack on DSA with
12958 * Make a new fixed-window mod_exp implementation the default for
12959 RSA, DSA, and DH private-key operations so that the sequence of
12960 squares and multiplies and the memory access pattern are
12961 independent of the particular secret key. This will mitigate
12962 cache-timing and potential related attacks.
12964 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12965 and this is automatically used by BN_mod_exp_mont() if the new flag
12966 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12967 will use this BN flag for private exponents unless the flag
12968 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12969 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12971 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12973 * Change the client implementation for SSLv23_method() and
12974 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12975 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12976 (Previously, the SSL 2.0 backwards compatible Client Hello
12977 message format would be used even with SSL_OP_NO_SSLv2.)
12981 * Add support for smime-type MIME parameter in S/MIME messages which some
12986 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12987 a threadsafe manner. Modify rsa code to use new function and add calls
12988 to dsa and dh code (which had race conditions before).
12992 * Include the fixed error library code in the C error file definitions
12993 instead of fixing them up at runtime. This keeps the error code
12994 structures constant.
12998 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
13000 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
13003 * Fixes for newer kerberos headers. NB: the casts are needed because
13004 the 'length' field is signed on one version and unsigned on another
13005 with no (?) obvious way to tell the difference, without these VC++
13006 complains. Also the "definition" of FAR (blank) is no longer included
13007 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
13008 some needed definitions.
13012 * Undo Cygwin change.
13016 * Added support for proxy certificates according to RFC 3820.
13017 Because they may be a security thread to unaware applications,
13018 they must be explicitly allowed in run-time. See
13019 docs/HOWTO/proxy_certificates.txt for further information.
13023 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
13025 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
13026 server and client random values. Previously
13027 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
13028 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
13030 This change has negligible security impact because:
13032 1. Server and client random values still have 24 bytes of pseudo random
13035 2. Server and client random values are sent in the clear in the initial
13038 3. The master secret is derived using the premaster secret (48 bytes in
13039 size for static RSA ciphersuites) as well as client server and random
13042 The OpenSSL team would like to thank the UK NISCC for bringing this issue
13045 *Stephen Henson, reported by UK NISCC*
13047 * Use Windows randomness collection on Cygwin.
13051 * Fix hang in EGD/PRNGD query when communication socket is closed
13052 prematurely by EGD/PRNGD.
13054 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
13056 * Prompt for pass phrases when appropriate for PKCS12 input format.
13060 * Back-port of selected performance improvements from development
13061 branch, as well as improved support for PowerPC platforms.
13065 * Add lots of checks for memory allocation failure, error codes to indicate
13066 failure and freeing up memory if a failure occurs.
13068 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
13070 * Add new -passin argument to dgst.
13074 * Perform some character comparisons of different types in X509_NAME_cmp:
13075 this is needed for some certificates that re-encode DNs into UTF8Strings
13076 (in violation of RFC3280) and can't or won't issue name rollover
13081 * Make an explicit check during certificate validation to see that
13082 the CA setting in each certificate on the chain is correct. As a
13083 side effect always do the following basic checks on extensions,
13084 not just when there's an associated purpose to the check:
13086 - if there is an unhandled critical extension (unless the user
13087 has chosen to ignore this fault)
13088 - if the path length has been exceeded (if one is set at all)
13089 - that certain extensions fit the associated purpose (if one has
13094 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
13096 * Avoid a race condition when CRLs are checked in a multi threaded
13097 environment. This would happen due to the reordering of the revoked
13098 entries during signature checking and serial number lookup. Now the
13099 encoding is cached and the serial number sort performed under a lock.
13100 Add new STACK function sk_is_sorted().
13104 * Add Delta CRL to the extension code.
13108 * Various fixes to s3_pkt.c so alerts are sent properly.
13110 *David Holmes <d.holmes@f5.com>*
13112 * Reduce the chances of duplicate issuer name and serial numbers (in
13113 violation of RFC3280) using the OpenSSL certificate creation utilities.
13114 This is done by creating a random 64 bit value for the initial serial
13115 number when a serial number file is created or when a self signed
13116 certificate is created using 'openssl req -x509'. The initial serial
13117 number file is created using 'openssl x509 -next_serial' in CA.pl
13118 rather than being initialized to 1.
13122 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
13124 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13125 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13127 *Joe Orton, Steve Henson*
13129 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
13132 *Joe Orton, Steve Henson*
13134 * Make it possible to have multiple active certificates with the same
13135 subject in the CA index file. This is done only if the keyword
13136 'unique_subject' is set to 'no' in the main CA section (default
13137 if 'CA_default') of the configuration file. The value is saved
13138 with the database itself in a separate index attribute file,
13139 named like the index file with '.attr' appended to the name.
13143 * X509 verify fixes. Disable broken certificate workarounds when
13144 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
13145 keyUsage extension present. Don't accept CRLs with unhandled critical
13146 extensions: since verify currently doesn't process CRL extensions this
13147 rejects a CRL with *any* critical extensions. Add new verify error codes
13152 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
13153 A clarification of RFC2560 will require the use of OCTET STRINGs and
13154 some implementations cannot handle the current raw format. Since OpenSSL
13155 copies and compares OCSP nonces as opaque blobs without any attempt at
13156 parsing them this should not create any compatibility issues.
13160 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
13161 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
13162 this HMAC (and other) operations are several times slower than OpenSSL
13167 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
13169 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13171 * Use the correct content when signing type "other".
13175 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
13177 * Fix various bugs revealed by running the NISCC test suite:
13179 Stop out of bounds reads in the ASN1 code when presented with
13180 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13182 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
13184 If verify callback ignores invalid public key errors don't try to check
13185 certificate signature with the NULL public key.
13189 * New -ignore_err option in ocsp application to stop the server
13190 exiting on the first error in a request.
13194 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13195 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13200 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13201 extra data after the compression methods not only for TLS 1.0
13202 but also for SSL 3.0 (as required by the specification).
13204 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13206 * Change X509_certificate_type() to mark the key as exported/exportable
13207 when it's 512 *bits* long, not 512 bytes.
13211 * Change AES_cbc_encrypt() so it outputs exact multiple of
13212 blocks during encryption.
13216 * Various fixes to base64 BIO and non blocking I/O. On write
13217 flushes were not handled properly if the BIO retried. On read
13218 data was not being buffered properly and had various logic bugs.
13219 This also affects blocking I/O when the data being decoded is a
13224 * Various S/MIME bugfixes and compatibility changes:
13225 output correct application/pkcs7 MIME type if
13226 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
13227 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
13228 of files as .eml work). Correctly handle very long lines in MIME
13233 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
13235 * Countermeasure against the Klima-Pokorny-Rosa extension of
13236 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13237 a protocol version number mismatch like a decryption error
13238 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13242 * Turn on RSA blinding by default in the default implementation
13243 to avoid a timing attack. Applications that don't want it can call
13244 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13245 They would be ill-advised to do so in most cases.
13247 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13249 * Change RSA blinding code so that it works when the PRNG is not
13250 seeded (in this case, the secret RSA exponent is abused as
13251 an unpredictable seed -- if it is not unpredictable, there
13252 is no point in blinding anyway). Make RSA blinding thread-safe
13253 by remembering the creator's thread ID in rsa->blinding and
13254 having all other threads use local one-time blinding factors
13255 (this requires more computation than sharing rsa->blinding, but
13256 avoids excessive locking; and if an RSA object is not shared
13257 between threads, blinding will still be very fast).
13261 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13262 ENGINE as defaults for all supported algorithms irrespective of
13263 the 'flags' parameter. 'flags' is now honoured, so applications
13264 should make sure they are passing it correctly.
13268 * Target "mingw" now allows native Windows code to be generated in
13269 the Cygwin environment as well as with the MinGW compiler.
13273 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13275 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13276 via timing by performing a MAC computation even if incorrect
13277 block cipher padding has been found. This is a countermeasure
13278 against active attacks where the attacker has to distinguish
13279 between bad padding and a MAC verification error. ([CVE-2003-0078])
13281 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13282 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13283 Martin Vuagnoux (EPFL, Ilion)*
13285 * Make the no-err option work as intended. The intention with no-err
13286 is not to have the whole error stack handling routines removed from
13287 libcrypto, it's only intended to remove all the function name and
13288 reason texts, thereby removing some of the footprint that may not
13289 be interesting if those errors aren't displayed anyway.
13291 NOTE: it's still possible for any application or module to have its
13292 own set of error texts inserted. The routines are there, just not
13293 used by default when no-err is given.
13297 * Add support for FreeBSD on IA64.
13299 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13301 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13302 Kerberos function mit_des_cbc_cksum(). Before this change,
13303 the value returned by DES_cbc_cksum() was like the one from
13304 mit_des_cbc_cksum(), except the bytes were swapped.
13306 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13308 * Allow an application to disable the automatic SSL chain building.
13309 Before this a rather primitive chain build was always performed in
13310 ssl3_output_cert_chain(): an application had no way to send the
13311 correct chain if the automatic operation produced an incorrect result.
13313 Now the chain builder is disabled if either:
13315 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13317 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13319 The reasoning behind this is that an application would not want the
13320 auto chain building to take place if extra chain certificates are
13321 present and it might also want a means of sending no additional
13322 certificates (for example the chain has two certificates and the
13327 * Add the possibility to build without the ENGINE framework.
13329 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13331 * Under Win32 gmtime() can return NULL: check return value in
13332 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13336 * DSA routines: under certain error conditions uninitialized BN objects
13337 could be freed. Solution: make sure initialization is performed early
13338 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13339 Nils Larsch <nla@trustcenter.de> via PR#459)
13343 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13344 checked on reconnect on the client side, therefore session resumption
13345 could still fail with a "ssl session id is different" error. This
13346 behaviour is masked when SSL_OP_ALL is used due to
13347 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13348 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13349 followup to PR #377.
13353 * IA-32 assembler support enhancements: unified ELF targets, support
13354 for SCO/Caldera platforms, fix for Cygwin shared build.
13358 * Add support for FreeBSD on sparc64. As a consequence, support for
13359 FreeBSD on non-x86 processors is separate from x86 processors on
13360 the config script, much like the NetBSD support.
13362 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13364 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13366 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13369 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13370 code (06) was taken as the first octet of the session ID and the last
13371 octet was ignored consequently. As a result SSLv2 client side session
13372 caching could not have worked due to the session ID mismatch between
13374 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13379 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13380 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13385 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13386 seems that in spite of existing for more than a year, many application
13387 author have done nothing to provide the necessary callbacks, which
13388 means that this particular engine will not work properly anywhere.
13389 This is a very unfortunate situation which forces us, in the name
13390 of usability, to give the hw_ncipher.c a static lock, which is part
13392 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13393 appear in 0.9.8 or later. We EXPECT application authors to have
13394 dealt properly with this when 0.9.8 is released (unless we actually
13395 make such changes in the libcrypto locking code that changes will
13396 have to be made anyway).
13400 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13401 octets have been read, EOF or an error occurs. Without this change
13402 some truncated ASN1 structures will not produce an error.
13406 * Disable Heimdal support, since it hasn't been fully implemented.
13407 Still give the possibility to force the use of Heimdal, but with
13408 warnings and a request that patches get sent to openssl-dev.
13412 * Add the VC-CE target, introduce the WINCE sysname, and add
13413 INSTALL.WCE and appropriate conditionals to make it build.
13415 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13417 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13418 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13419 edit numbers of the version.
13421 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13423 * Introduce safe string copy and catenation functions
13424 (BUF_strlcpy() and BUF_strlcat()).
13426 *Ben Laurie (CHATS) and Richard Levitte*
13428 * Avoid using fixed-size buffers for one-line DNs.
13430 *Ben Laurie (CHATS)*
13432 * Add BUF_MEM_grow_clean() to avoid information leakage when
13433 resizing buffers containing secrets, and use where appropriate.
13435 *Ben Laurie (CHATS)*
13437 * Avoid using fixed size buffers for configuration file location.
13439 *Ben Laurie (CHATS)*
13441 * Avoid filename truncation for various CA files.
13443 *Ben Laurie (CHATS)*
13445 * Use sizeof in preference to magic numbers.
13447 *Ben Laurie (CHATS)*
13449 * Avoid filename truncation in cert requests.
13451 *Ben Laurie (CHATS)*
13453 * Add assertions to check for (supposedly impossible) buffer
13456 *Ben Laurie (CHATS)*
13458 * Don't cache truncated DNS entries in the local cache (this could
13459 potentially lead to a spoofing attack).
13461 *Ben Laurie (CHATS)*
13463 * Fix various buffers to be large enough for hex/decimal
13464 representations in a platform independent manner.
13466 *Ben Laurie (CHATS)*
13468 * Add CRYPTO_realloc_clean() to avoid information leakage when
13469 resizing buffers containing secrets, and use where appropriate.
13471 *Ben Laurie (CHATS)*
13473 * Add BIO_indent() to avoid much slightly worrying code to do
13476 *Ben Laurie (CHATS)*
13478 * Convert sprintf()/BIO_puts() to BIO_printf().
13480 *Ben Laurie (CHATS)*
13482 * buffer_gets() could terminate with the buffer only half
13485 *Ben Laurie (CHATS)*
13487 * Add assertions to prevent user-supplied crypto functions from
13488 overflowing internal buffers by having large block sizes, etc.
13490 *Ben Laurie (CHATS)*
13492 * New OPENSSL_assert() macro (similar to assert(), but enabled
13495 *Ben Laurie (CHATS)*
13497 * Eliminate unused copy of key in RC4.
13499 *Ben Laurie (CHATS)*
13501 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13503 *Ben Laurie (CHATS)*
13505 * Fix off-by-one error in EGD path.
13507 *Ben Laurie (CHATS)*
13509 * If RANDFILE path is too long, ignore instead of truncating.
13511 *Ben Laurie (CHATS)*
13513 * Eliminate unused and incorrectly sized X.509 structure
13516 *Ben Laurie (CHATS)*
13518 * Eliminate unused and dangerous function knumber().
13520 *Ben Laurie (CHATS)*
13522 * Eliminate unused and dangerous structure, KSSL_ERR.
13524 *Ben Laurie (CHATS)*
13526 * Protect against overlong session ID context length in an encoded
13527 session object. Since these are local, this does not appear to be
13530 *Ben Laurie (CHATS)*
13532 * Change from security patch (see 0.9.6e below) that did not affect
13533 the 0.9.6 release series:
13535 Remote buffer overflow in SSL3 protocol - an attacker could
13536 supply an oversized master key in Kerberos-enabled versions.
13539 *Ben Laurie (CHATS)*
13541 * Change the SSL kerb5 codes to match RFC 2712.
13545 * Make -nameopt work fully for req and add -reqopt switch.
13547 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13549 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13551 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13553 * Make sure tests can be performed even if the corresponding algorithms
13554 have been removed entirely. This was also the last step to make
13555 OpenSSL compilable with DJGPP under all reasonable conditions.
13557 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13559 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13560 to allow version independent disabling of normally unselected ciphers,
13561 which may be activated as a side-effect of selecting a single cipher.
13563 (E.g., cipher list string "RSA" enables ciphersuites that are left
13564 out of "ALL" because they do not provide symmetric encryption.
13565 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13567 *Lutz Jaenicke, Bodo Moeller*
13569 * Add appropriate support for separate platform-dependent build
13570 directories. The recommended way to make a platform-dependent
13571 build directory is the following (tested on Linux), maybe with
13574 # Place yourself outside of the OpenSSL source tree. In
13575 # this example, the environment variable OPENSSL_SOURCE
13576 # is assumed to contain the absolute OpenSSL source directory.
13577 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13578 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13579 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13580 mkdir -p `dirname $F`
13581 ln -s $OPENSSL_SOURCE/$F $F
13584 To be absolutely sure not to disturb the source tree, a "make clean"
13585 is a good thing. If it isn't successful, don't worry about it,
13586 it probably means the source directory is very clean.
13590 * Make sure any ENGINE control commands make local copies of string
13591 pointers passed to them whenever necessary. Otherwise it is possible
13592 the caller may have overwritten (or deallocated) the original string
13593 data when a later ENGINE operation tries to use the stored values.
13595 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13597 * Improve diagnostics in file reading and command-line digests.
13599 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13601 * Add AES modes CFB and OFB to the object database. Correct an
13602 error in AES-CFB decryption.
13606 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13607 allows existing EVP_CIPHER_CTX structures to be reused after
13608 calling `EVP_*Final()`. This behaviour is used by encryption
13609 BIOs and some applications. This has the side effect that
13610 applications must explicitly clean up cipher contexts with
13611 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13615 * Check the values of dna and dnb in bn_mul_recursive before calling
13616 bn_mul_comba (a non zero value means the a or b arrays do not contain
13617 n2 elements) and fallback to bn_mul_normal if either is not zero.
13621 * Fix escaping of non-ASCII characters when using the -subj option
13622 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13626 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13627 form for "surname", serialNumber has no short form.
13628 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13629 therefore remove "mail" short name for "internet 7".
13630 The OID for unique identifiers in X509 certificates is
13631 x500UniqueIdentifier, not uniqueIdentifier.
13632 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13636 * Add an "init" command to the ENGINE config module and auto initialize
13637 ENGINEs. Without any "init" command the ENGINE will be initialized
13638 after all ctrl commands have been executed on it. If init=1 the
13639 ENGINE is initialized at that point (ctrls before that point are run
13640 on the uninitialized ENGINE and after on the initialized one). If
13641 init=0 then the ENGINE will not be initialized at all.
13645 * Fix the 'app_verify_callback' interface so that the user-defined
13646 argument is actually passed to the callback: In the
13647 SSL_CTX_set_cert_verify_callback() prototype, the callback
13648 declaration has been changed from
13651 int (*cb)(X509_STORE_CTX *,void *);
13652 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13653 i=s->ctx->app_verify_callback(&ctx)
13654 has been changed into
13655 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13657 To update applications using SSL_CTX_set_cert_verify_callback(),
13658 a dummy argument can be added to their callback functions.
13660 *D. K. Smetters <smetters@parc.xerox.com>*
13662 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13664 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13666 * Add and OPENSSL_LOAD_CONF define which will cause
13667 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13668 This allows older applications to transparently support certain
13669 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13670 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13671 load the config file and OPENSSL_add_all_algorithms_conf() which will
13672 always load it have also been added.
13676 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13677 Adjust NIDs and EVP layer.
13679 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13681 * Config modules support in openssl utility.
13683 Most commands now load modules from the config file,
13684 though in a few (such as version) this isn't done
13685 because it couldn't be used for anything.
13687 In the case of ca and req the config file used is
13688 the same as the utility itself: that is the -config
13689 command line option can be used to specify an
13694 * Move default behaviour from OPENSSL_config(). If appname is NULL
13695 use "openssl_conf" if filename is NULL use default openssl config file.
13699 * Add an argument to OPENSSL_config() to allow the use of an alternative
13700 config section name. Add a new flag to tolerate a missing config file
13701 and move code to CONF_modules_load_file().
13705 * Support for crypto accelerator cards from Accelerated Encryption
13706 Processing, www.aep.ie. (Use engine 'aep')
13707 The support was copied from 0.9.6c [engine] and adapted/corrected
13708 to work with the new engine framework.
13710 *AEP Inc. and Richard Levitte*
13712 * Support for SureWare crypto accelerator cards from Baltimore
13713 Technologies. (Use engine 'sureware')
13714 The support was copied from 0.9.6c [engine] and adapted
13715 to work with the new engine framework.
13719 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13720 make the newer ENGINE framework commands for the CHIL engine work.
13722 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13724 * Make it possible to produce shared libraries on ReliantUNIX.
13726 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13728 * Add the configuration target debug-linux-ppro.
13729 Make 'openssl rsa' use the general key loading routines
13730 implemented in `apps.c`, and make those routines able to
13731 handle the key format FORMAT_NETSCAPE and the variant
13734 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13736 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13738 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13740 * Add -keyform to rsautl, and document -engine.
13742 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13744 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13745 BIO_R_NO_SUCH_FILE error code rather than the generic
13746 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13750 * Add new functions
13751 ERR_peek_last_error
13752 ERR_peek_last_error_line
13753 ERR_peek_last_error_line_data.
13754 These are similar to
13756 ERR_peek_error_line
13757 ERR_peek_error_line_data,
13758 but report on the latest error recorded rather than the first one
13759 still in the error queue.
13761 *Ben Laurie, Bodo Moeller*
13763 * default_algorithms option in ENGINE config module. This allows things
13765 default_algorithms = ALL
13766 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13770 * Preliminary ENGINE config module.
13774 * New experimental application configuration code.
13778 * Change the AES code to follow the same name structure as all other
13779 symmetric ciphers, and behave the same way. Move everything to
13780 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13782 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13784 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13786 *Ben Laurie and Theo de Raadt*
13788 * Add option to output public keys in req command.
13790 *Massimiliano Pala madwolf@openca.org*
13792 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13793 (up to about 10% better than before for P-192 and P-224).
13797 * New functions/macros
13799 SSL_CTX_set_msg_callback(ctx, cb)
13800 SSL_CTX_set_msg_callback_arg(ctx, arg)
13801 SSL_set_msg_callback(ssl, cb)
13802 SSL_set_msg_callback_arg(ssl, arg)
13804 to request calling a callback function
13806 void cb(int write_p, int version, int content_type,
13807 const void *buf, size_t len, SSL *ssl, void *arg)
13809 whenever a protocol message has been completely received
13810 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13811 protocol version according to which the SSL library interprets
13812 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13813 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13814 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13815 specification (change_cipher_spec(20), alert(21), handshake(22)).
13816 'buf' and 'len' point to the actual message, 'ssl' to the
13817 SSL object, and 'arg' is the application-defined value set by
13818 SSL[_CTX]_set_msg_callback_arg().
13820 'openssl s_client' and 'openssl s_server' have new '-msg' options
13821 to enable a callback that displays all protocol messages.
13825 * Change the shared library support so shared libraries are built as
13826 soon as the corresponding static library is finished, and thereby get
13827 openssl and the test programs linked against the shared library.
13828 This still only happens when the keyword "shard" has been given to
13829 the configuration scripts.
13831 NOTE: shared library support is still an experimental thing, and
13832 backward binary compatibility is still not guaranteed.
13834 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13836 * Add support for Subject Information Access extension.
13838 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13840 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13841 additional bytes when new memory had to be allocated, not just
13842 when reusing an existing buffer.
13846 * New command line and configuration option 'utf8' for the req command.
13847 This allows field values to be specified as UTF8 strings.
13851 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13852 runs for the former and machine-readable output for the latter.
13856 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13857 of the e-mail address in the DN (i.e., it will go into a certificate
13858 extension only). The new configuration file option 'email_in_dn = no'
13859 has the same effect.
13861 *Massimiliano Pala madwolf@openca.org*
13863 * Change all functions with names starting with `des_` to be starting
13864 with `DES_` instead. Add wrappers that are compatible with libdes,
13865 but are named `_ossl_old_des_*`. Finally, add macros that map the
13866 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13867 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13868 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13871 Since we provide two compatibility mappings, the user needs to
13872 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13873 compatibility is desired. The default (i.e., when that macro
13874 isn't defined) is OpenSSL 0.9.6c compatibility.
13876 There are also macros that enable and disable the support of old
13877 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13878 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13879 are defined, the default will apply: to support the old des routines.
13881 In either case, one must include openssl/des.h to get the correct
13882 definitions. Do not try to just include openssl/des_old.h, that
13885 NOTE: This is a major break of an old API into a new one. Software
13886 authors are encouraged to switch to the `DES_` style functions. Some
13887 time in the future, des_old.h and the libdes compatibility functions
13888 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13889 default), and then completely removed.
13893 * Test for certificates which contain unsupported critical extensions.
13894 If such a certificate is found during a verify operation it is
13895 rejected by default: this behaviour can be overridden by either
13896 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13897 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13898 X509_supported_extension() has also been added which returns 1 if a
13899 particular extension is supported.
13903 * Modify the behaviour of EVP cipher functions in similar way to digests
13904 to retain compatibility with existing code.
13908 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13909 compatibility with existing code. In particular the 'ctx' parameter does
13910 not have to be to be initialized before the call to EVP_DigestInit() and
13911 it is tidied up after a call to EVP_DigestFinal(). New function
13912 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13913 EVP_MD_CTX_copy() changed to not require the destination to be
13914 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13915 requires the destination to be valid.
13917 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13918 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13922 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13923 so that complete 'Handshake' protocol structures are kept in memory
13924 instead of overwriting 'msg_type' and 'length' with 'body' data.
13928 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13930 *Massimo Santin via Richard Levitte*
13932 * Major restructuring to the underlying ENGINE code. This includes
13933 reduction of linker bloat, separation of pure "ENGINE" manipulation
13934 (initialisation, etc) from functionality dealing with implementations
13935 of specific crypto interfaces. This change also introduces integrated
13936 support for symmetric ciphers and digest implementations - so ENGINEs
13937 can now accelerate these by providing EVP_CIPHER and EVP_MD
13938 implementations of their own. This is detailed in
13939 [crypto/engine/README.md](crypto/engine/README.md)
13940 as it couldn't be adequately described here. However, there are a few
13941 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13942 were changed in the original introduction of ENGINE code have now
13943 reverted back - the hooking from this code to ENGINE is now a good
13944 deal more passive and at run-time, operations deal directly with
13945 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13946 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13947 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13948 they were not being used by the framework as there is no concept of a
13949 BIGNUM_METHOD and they could not be generalised to the new
13950 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13951 ENGINE_cpy() has been removed as it cannot be consistently defined in
13956 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13960 * Change mkdef.pl to sort symbols that get the same entry number,
13961 and make sure the automatically generated functions `ERR_load_*`
13962 become part of libeay.num as well.
13966 * New function SSL_renegotiate_pending(). This returns true once
13967 renegotiation has been requested (either SSL_renegotiate() call
13968 or HelloRequest/ClientHello received from the peer) and becomes
13969 false once a handshake has been completed.
13970 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13971 sends a HelloRequest, but does not ensure that a handshake takes
13972 place. SSL_renegotiate_pending() is useful for checking if the
13973 client has followed the request.)
13977 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13978 By default, clients may request session resumption even during
13979 renegotiation (if session ID contexts permit); with this option,
13980 session resumption is possible only in the first handshake.
13982 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13983 more bits available for options that should not be part of
13984 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13988 * Add some demos for certificate and certificate request creation.
13992 * Make maximum certificate chain size accepted from the peer application
13993 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13994 "Douglas E. Engert" <deengert@anl.gov>.
13998 * Add support for shared libraries for Unixware-7
13999 (Boyd Lynn Gerber <gerberb@zenez.com>).
14003 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
14004 be done prior to destruction. Use this to unload error strings from
14005 ENGINEs that load their own error strings. NB: This adds two new API
14006 functions to "get" and "set" this destroy handler in an ENGINE.
14010 * Alter all existing ENGINE implementations (except "openssl" and
14011 "openbsd") to dynamically instantiate their own error strings. This
14012 makes them more flexible to be built both as statically-linked ENGINEs
14013 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
14014 Also, add stub code to each that makes building them as self-contained
14015 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
14019 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
14020 implementations into applications that are completely implemented in
14021 self-contained shared-libraries. The "dynamic" ENGINE exposes control
14022 commands that can be used to configure what shared-library to load and
14023 to control aspects of the way it is handled. Also, made an update to
14024 the [README-Engine.md](README-Engine.md) file
14025 that brings its information up-to-date and
14026 provides some information and instructions on the "dynamic" ENGINE
14027 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
14031 * Make it possible to unload ranges of ERR strings with a new
14032 "ERR_unload_strings" function.
14036 * Add a copy() function to EVP_MD.
14040 * Make EVP_MD routines take a context pointer instead of just the
14041 md_data void pointer.
14045 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
14046 that the digest can only process a single chunk of data
14047 (typically because it is provided by a piece of
14048 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
14049 is only going to provide a single chunk of data, and hence the
14050 framework needn't accumulate the data for oneshot drivers.
14054 * As with "ERR", make it possible to replace the underlying "ex_data"
14055 functions. This change also alters the storage and management of global
14056 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
14057 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
14058 index counters. The API functions that use this state have been changed
14059 to take a "class_index" rather than pointers to the class's local STACK
14060 and counter, and there is now an API function to dynamically create new
14061 classes. This centralisation allows us to (a) plug a lot of the
14062 thread-safety problems that existed, and (b) makes it possible to clean
14063 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
14064 such data would previously have always leaked in application code and
14065 workarounds were in place to make the memory debugging turn a blind eye
14066 to it. Application code that doesn't use this new function will still
14067 leak as before, but their memory debugging output will announce it now
14068 rather than letting it slide.
14070 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
14071 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
14072 has a return value to indicate success or failure.
14076 * Make it possible to replace the underlying "ERR" functions such that the
14077 global state (2 LHASH tables and 2 locks) is only used by the "default"
14078 implementation. This change also adds two functions to "get" and "set"
14079 the implementation prior to it being automatically set the first time
14080 any other ERR function takes place. Ie. an application can call "get",
14081 pass the return value to a module it has just loaded, and that module
14082 can call its own "set" function using that value. This means the
14083 module's "ERR" operations will use (and modify) the error state in the
14084 application and not in its own statically linked copy of OpenSSL code.
14088 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
14089 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
14090 the operation, and provides a more encapsulated way for external code
14091 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
14092 to use these functions rather than manually incrementing the counts.
14094 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
14098 * Add EVP test program.
14102 * Add symmetric cipher support to ENGINE. Expect the API to change!
14106 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
14107 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
14108 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
14109 These allow a CRL to be built without having to access X509_CRL fields
14110 directly. Modify 'ca' application to use new functions.
14114 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
14115 bug workarounds. Rollback attack detection is a security feature.
14116 The problem will only arise on OpenSSL servers when TLSv1 is not
14117 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
14118 Software authors not wanting to support TLSv1 will have special reasons
14119 for their choice and can explicitly enable this option.
14121 *Bodo Moeller, Lutz Jaenicke*
14123 * Rationalise EVP so it can be extended: don't include a union of
14124 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
14125 (similar to those existing for EVP_CIPHER_CTX).
14130 EVP_MD_CTX_init(&md); /* new function call */
14131 EVP_DigestInit(&md, EVP_sha1());
14132 EVP_DigestUpdate(&md, in, len);
14133 EVP_DigestFinal(&md, out, NULL);
14134 EVP_MD_CTX_cleanup(&md); /* new function call */
14138 * Make DES key schedule conform to the usual scheme, as well as
14139 correcting its structure. This means that calls to DES functions
14140 now have to pass a pointer to a des_key_schedule instead of a
14141 plain des_key_schedule (which was actually always a pointer
14144 des_key_schedule ks;
14146 des_set_key_checked(..., &ks);
14147 des_ncbc_encrypt(..., &ks, ...);
14149 (Note that a later change renames 'des_...' into 'DES_...'.)
14153 * Initial reduction of linker bloat: the use of some functions, such as
14154 PEM causes large amounts of unused functions to be linked in due to
14155 poor organisation. For example pem_all.c contains every PEM function
14156 which has a knock on effect of linking in large amounts of (unused)
14157 ASN1 code. Grouping together similar functions and splitting unrelated
14158 functions prevents this.
14162 * Cleanup of EVP macros.
14166 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
14167 correct `_ecb suffix`.
14171 * Add initial OCSP responder support to ocsp application. The
14172 revocation information is handled using the text based index
14173 use by the ca application. The responder can either handle
14174 requests generated internally, supplied in files (for example
14175 via a CGI script) or using an internal minimal server.
14179 * Add configuration choices to get zlib compression for TLS.
14183 * Changes to Kerberos SSL for RFC 2712 compliance:
14184 1. Implemented real KerberosWrapper, instead of just using
14185 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
14186 2. Implemented optional authenticator field of KerberosWrapper.
14188 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
14189 and authenticator structs; see crypto/krb5/.
14191 Generalized Kerberos calls to support multiple Kerberos libraries.
14192 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
14193 via Richard Levitte*
14195 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
14196 already does with RSA. testdsa.h now has 'priv_key/pub_key'
14197 values for each of the key sizes rather than having just
14198 parameters (and 'speed' generating keys each time).
14202 * Speed up EVP routines.
14205 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
14206 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
14207 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
14208 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
14210 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
14211 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
14212 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
14215 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
14217 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
14221 * Added the OS2-EMX target.
14223 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
14225 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
14226 New functions to support `NCONF` routines in extension code.
14227 New function `CONF_set_nconf()`
14228 to allow functions which take an `NCONF` to also handle the old `LHASH`
14229 structure: this means that the old `CONF` compatible routines can be
14230 retained (in particular w.rt. extensions) without having to duplicate the
14231 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
14235 * Enhance the general user interface with mechanisms for inner control
14236 and with possibilities to have yes/no kind of prompts.
14240 * Change all calls to low-level digest routines in the library and
14241 applications to use EVP. Add missing calls to HMAC_cleanup() and
14242 don't assume HMAC_CTX can be copied using memcpy().
14244 *Verdon Walker <VWalker@novell.com>, Steve Henson*
14246 * Add the possibility to control engines through control names but with
14247 arbitrary arguments instead of just a string.
14248 Change the key loaders to take a UI_METHOD instead of a callback
14249 function pointer. NOTE: this breaks binary compatibility with earlier
14250 versions of OpenSSL [engine].
14251 Adapt the nCipher code for these new conditions and add a card insertion
14256 * Enhance the general user interface with mechanisms to better support
14257 dialog box interfaces, application-defined prompts, the possibility
14258 to use defaults (for example default passwords from somewhere else)
14259 and interrupts/cancellations.
14263 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14264 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14268 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14269 tidy up some unnecessarily weird code in 'sk_new()').
14271 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14273 * Change the key loading routines for ENGINEs to use the same kind
14274 callback (pem_password_cb) as all other routines that need this
14279 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14280 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14281 than this minimum value is recommended.
14285 * New random seeder for OpenVMS, using the system process statistics
14286 that are easily reachable.
14290 * Windows apparently can't transparently handle global
14291 variables defined in DLLs. Initialisations such as:
14293 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14295 won't compile. This is used by the any applications that need to
14296 declare their own ASN1 modules. This was fixed by adding the option
14297 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14298 needed for static libraries under Win32.
14302 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14303 setting of purpose and trust fields. New X509_STORE trust and
14304 purpose functions and tidy up setting in other SSL functions.
14308 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14309 structure. These are inherited by X509_STORE_CTX when it is
14310 initialised. This allows various defaults to be set in the
14311 X509_STORE structure (such as flags for CRL checking and custom
14312 purpose or trust settings) for functions which only use X509_STORE_CTX
14313 internally such as S/MIME.
14315 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14316 trust settings if they are not set in X509_STORE. This allows X509_STORE
14317 purposes and trust (in S/MIME for example) to override any set by default.
14319 Add command line options for CRL checking to smime, s_client and s_server
14324 * Initial CRL based revocation checking. If the CRL checking flag(s)
14325 are set then the CRL is looked up in the X509_STORE structure and
14326 its validity and signature checked, then if the certificate is found
14327 in the CRL the verify fails with a revoked error.
14329 Various new CRL related callbacks added to X509_STORE_CTX structure.
14331 Command line options added to 'verify' application to support this.
14333 This needs some additional work, such as being able to handle multiple
14334 CRLs with different times, extension based lookup (rather than just
14335 by subject name) and ultimately more complete V2 CRL extension
14340 * Add a general user interface API (crypto/ui/). This is designed
14341 to replace things like des_read_password and friends (backward
14342 compatibility functions using this new API are provided).
14343 The purpose is to remove prompting functions from the DES code
14344 section as well as provide for prompting through dialog boxes in
14345 a window system and the like.
14349 * Add "ex_data" support to ENGINE so implementations can add state at a
14350 per-structure level rather than having to store it globally.
14354 * Make it possible for ENGINE structures to be copied when retrieved by
14355 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14356 This causes the "original" ENGINE structure to act like a template,
14357 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14358 operational state can be localised to each ENGINE structure, despite the
14359 fact they all share the same "methods". New ENGINE structures returned in
14360 this case have no functional references and the return value is the single
14361 structural reference. This matches the single structural reference returned
14362 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14367 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14368 needs to match any other type at all we need to manually clear the
14373 * Changes to the "openssl engine" utility to include;
14374 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14375 about an ENGINE's available control commands.
14376 - executing control commands from command line arguments using the
14377 '-pre' and '-post' switches. '-post' is only used if '-t' is
14378 specified and the ENGINE is successfully initialised. The syntax for
14379 the individual commands are colon-separated, for example;
14380 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14384 * New dynamic control command support for ENGINEs. ENGINEs can now
14385 declare their own commands (numbers), names (strings), descriptions,
14386 and input types for run-time discovery by calling applications. A
14387 subset of these commands are implicitly classed as "executable"
14388 depending on their input type, and only these can be invoked through
14389 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14390 can be based on user input, config files, etc). The distinction is
14391 that "executable" commands cannot return anything other than a boolean
14392 result and can only support numeric or string input, whereas some
14393 discoverable commands may only be for direct use through
14394 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14395 pointers, or other custom uses. The "executable" commands are to
14396 support parameterisations of ENGINE behaviour that can be
14397 unambiguously defined by ENGINEs and used consistently across any
14398 OpenSSL-based application. Commands have been added to all the
14399 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14400 control over shared-library paths without source code alterations.
14404 * Changed all ENGINE implementations to dynamically allocate their
14405 ENGINEs rather than declaring them statically. Apart from this being
14406 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14407 this also allows the implementations to compile without using the
14408 internal engine_int.h header.
14412 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14413 'const' value. Any code that should be able to modify a RAND_METHOD
14414 should already have non-const pointers to it (ie. they should only
14415 modify their own ones).
14419 * Made a variety of little tweaks to the ENGINE code.
14420 - "atalla" and "ubsec" string definitions were moved from header files
14421 to C code. "nuron" string definitions were placed in variables
14422 rather than hard-coded - allowing parameterisation of these values
14423 later on via ctrl() commands.
14424 - Removed unused "#if 0"'d code.
14425 - Fixed engine list iteration code so it uses ENGINE_free() to release
14426 structural references.
14427 - Constified the RAND_METHOD element of ENGINE structures.
14428 - Constified various get/set functions as appropriate and added
14429 missing functions (including a catch-all ENGINE_cpy that duplicates
14430 all ENGINE values onto a new ENGINE except reference counts/state).
14431 - Removed NULL parameter checks in get/set functions. Setting a method
14432 or function to NULL is a way of cancelling out a previously set
14433 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14434 and doesn't justify the extra error symbols and code.
14435 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14436 flags from engine_int.h to engine.h.
14437 - Changed prototypes for ENGINE handler functions (init(), finish(),
14438 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14442 * Implement binary inversion algorithm for BN_mod_inverse in addition
14443 to the algorithm using long division. The binary algorithm can be
14444 used only if the modulus is odd. On 32-bit systems, it is faster
14445 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14446 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14447 up to 450 bits. In 64-bit environments, the binary algorithm
14448 appears to be advantageous for much longer moduli; here we use it
14449 for moduli up to 2048 bits.
14453 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14454 could not support the combine flag in choice fields.
14458 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14459 extensions from a certificate request to the certificate.
14463 * Allow multiple 'certopt' and 'nameopt' options to be separated
14464 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14465 file: this allows the display of the certificate about to be
14466 signed to be customised, to allow certain fields to be included
14467 or excluded and extension details. The old system didn't display
14468 multicharacter strings properly, omitted fields not in the policy
14469 and couldn't display additional details such as extensions.
14473 * Function EC_POINTs_mul for multiple scalar multiplication
14474 of an arbitrary number of elliptic curve points
14475 \sum scalars[i]*points[i],
14476 optionally including the generator defined for the EC_GROUP:
14477 scalar*generator + \sum scalars[i]*points[i].
14479 EC_POINT_mul is a simple wrapper function for the typical case
14480 that the point list has just one item (besides the optional
14485 * First EC_METHODs for curves over GF(p):
14487 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14488 operations and provides various method functions that can also
14489 operate with faster implementations of modular arithmetic.
14491 EC_GFp_mont_method() reuses most functions that are part of
14492 EC_GFp_simple_method, but uses Montgomery arithmetic.
14494 *Bodo Moeller; point addition and point doubling
14495 implementation directly derived from source code provided by
14496 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14498 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14499 crypto/ec/ec_lib.c):
14501 Curves are EC_GROUP objects (with an optional group generator)
14502 based on EC_METHODs that are built into the library.
14504 Points are EC_POINT objects based on EC_GROUP objects.
14506 Most of the framework would be able to handle curves over arbitrary
14507 finite fields, but as there are no obvious types for fields other
14508 than GF(p), some functions are limited to that for now.
14512 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14513 that the file contains a complete HTTP response.
14517 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14518 change the def and num file printf format specifier from "%-40sXXX"
14519 to "%-39s XXX". The latter will always guarantee a space after the
14520 field while the former will cause them to run together if the field
14521 is 40 of more characters long.
14525 * Constify the cipher and digest 'method' functions and structures
14526 and modify related functions to take constant EVP_MD and EVP_CIPHER
14531 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14532 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14536 * Modify `EVP_Digest*()` routines so they now return values. Although the
14537 internal software routines can never fail additional hardware versions
14542 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14544 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14545 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14548 ERR_R_NESTED_ASN1_ERROR
14550 ERR_R_MISSING_ASN1_EOS
14551 were 4 .. 9, conflicting with
14552 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14554 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14555 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14557 Add new error code 'ERR_R_INTERNAL_ERROR'.
14561 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14566 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14567 sets the subject name for a new request or supersedes the
14568 subject name in a given request. Formats that can be parsed are
14569 'CN=Some Name, OU=myOU, C=IT'
14571 'CN=Some Name/OU=myOU/C=IT'.
14573 Add options '-batch' and '-verbose' to 'openssl req'.
14575 *Massimiliano Pala <madwolf@hackmasters.net>*
14577 * Introduce the possibility to access global variables through
14578 functions on platform were that's the best way to handle exporting
14579 global variables in shared libraries. To enable this functionality,
14580 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14581 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14582 is normally done by Configure or something similar).
14584 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14585 in the source file (foo.c) like this:
14587 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14588 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14590 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14591 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14593 OPENSSL_DECLARE_GLOBAL(int,foo);
14594 #define foo OPENSSL_GLOBAL_REF(foo)
14595 OPENSSL_DECLARE_GLOBAL(double,bar);
14596 #define bar OPENSSL_GLOBAL_REF(bar)
14598 The #defines are very important, and therefore so is including the
14599 header file everywhere where the defined globals are used.
14601 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14602 of ASN.1 items, but that structure is a bit different.
14604 The largest change is in util/mkdef.pl which has been enhanced with
14605 better and easier to understand logic to choose which symbols should
14606 go into the Windows .def files as well as a number of fixes and code
14607 cleanup (among others, algorithm keywords are now sorted
14608 lexicographically to avoid constant rewrites).
14612 * In BN_div() keep a copy of the sign of 'num' before writing the
14613 result to 'rm' because if rm==num the value will be overwritten
14614 and produce the wrong result if 'num' is negative: this caused
14615 problems with BN_mod() and BN_nnmod().
14619 * Function OCSP_request_verify(). This checks the signature on an
14620 OCSP request and verifies the signer certificate. The signer
14621 certificate is just checked for a generic purpose and OCSP request
14626 * Add OCSP_check_validity() function to check the validity of OCSP
14627 responses. OCSP responses are prepared in real time and may only
14628 be a few seconds old. Simply checking that the current time lies
14629 between thisUpdate and nextUpdate max reject otherwise valid responses
14630 caused by either OCSP responder or client clock inaccuracy. Instead
14631 we allow thisUpdate and nextUpdate to fall within a certain period of
14632 the current time. The age of the response can also optionally be
14633 checked. Two new options -validity_period and -status_age added to
14638 * If signature or public key algorithm is unrecognized print out its
14639 OID rather that just UNKNOWN.
14643 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14644 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14645 ID to be generated from the issuer certificate alone which can then be
14646 passed to OCSP_id_issuer_cmp().
14650 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14651 ASN1 modules to export functions returning ASN1_ITEM pointers
14652 instead of the ASN1_ITEM structures themselves. This adds several
14653 new macros which allow the underlying ASN1 function/structure to
14654 be accessed transparently. As a result code should not use ASN1_ITEM
14655 references directly (such as &X509_it) but instead use the relevant
14656 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14657 use of the new ASN1 code on platforms where exporting structures
14658 is problematical (for example in shared libraries) but exporting
14659 functions returning pointers to structures is not.
14663 * Add support for overriding the generation of SSL/TLS session IDs.
14664 These callbacks can be registered either in an SSL_CTX or per SSL.
14665 The purpose of this is to allow applications to control, if they wish,
14666 the arbitrary values chosen for use as session IDs, particularly as it
14667 can be useful for session caching in multiple-server environments. A
14668 command-line switch for testing this (and any client code that wishes
14669 to use such a feature) has been added to "s_server".
14671 *Geoff Thorpe, Lutz Jaenicke*
14673 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14674 of the form `#if defined(...) || defined(...) || ...` and
14675 `#if !defined(...) && !defined(...) && ...`. This also avoids
14676 the growing number of special cases it was previously handling.
14680 * Make all configuration macros available for application by making
14681 sure they are available in opensslconf.h, by giving them names starting
14682 with `OPENSSL_` to avoid conflicts with other packages and by making
14683 sure e_os2.h will cover all platform-specific cases together with
14685 Additionally, it is now possible to define configuration/platform-
14686 specific names (called "system identities"). In the C code, these
14687 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14688 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14689 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14694 * New option -set_serial to 'req' and 'x509' this allows the serial
14695 number to use to be specified on the command line. Previously self
14696 signed certificates were hard coded with serial number 0 and the
14697 CA options of 'x509' had to use a serial number in a file which was
14702 * New options to 'ca' utility to support V2 CRL entry extensions.
14703 Currently CRL reason, invalidity date and hold instruction are
14704 supported. Add new CRL extensions to V3 code and some new objects.
14708 * New function EVP_CIPHER_CTX_set_padding() this is used to
14709 disable standard block padding (aka PKCS#5 padding) in the EVP
14710 API, which was previously mandatory. This means that the data is
14711 not padded in any way and so the total length much be a multiple
14712 of the block size, otherwise an error occurs.
14716 * Initial (incomplete) OCSP SSL support.
14720 * New function OCSP_parse_url(). This splits up a URL into its host,
14721 port and path components: primarily to parse OCSP URLs. New -url
14722 option to ocsp utility.
14726 * New nonce behavior. The return value of OCSP_check_nonce() now
14727 reflects the various checks performed. Applications can decide
14728 whether to tolerate certain situations such as an absent nonce
14729 in a response when one was present in a request: the ocsp application
14730 just prints out a warning. New function OCSP_add1_basic_nonce()
14731 this is to allow responders to include a nonce in a response even if
14732 the request is nonce-less.
14736 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14737 skipped when using openssl x509 multiple times on a single input file,
14738 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14742 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14743 set string type: to handle setting ASN1_TIME structures. Fix ca
14744 utility to correctly initialize revocation date of CRLs.
14748 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14749 the clients preferred ciphersuites and rather use its own preferences.
14750 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14751 Internet Explorer by ensuring unchanged hash method during stepup.
14752 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14756 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14757 to aes and add a new 'exist' option to print out symbols that don't
14762 * Additional options to ocsp utility to allow flags to be set and
14763 additional certificates supplied.
14767 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14768 OCSP client a number of certificate to only verify the response
14773 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14774 handle the new API. Currently only ECB, CBC modes supported. Add new
14777 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14778 Encryption Standard (AES) Ciphersuites for Transport Layer
14779 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14780 not enabled by default and were not part of the "ALL" ciphersuite
14781 alias because they were not yet official; they could be
14782 explicitly requested by specifying the "AESdraft" ciphersuite
14783 group alias. In the final release of OpenSSL 0.9.7, the group
14784 alias is called "AES" and is part of "ALL".)
14786 *Ben Laurie, Steve Henson, Bodo Moeller*
14788 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14789 request to response.
14793 * Functions for OCSP responders. OCSP_request_onereq_count(),
14794 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14795 extract information from a certificate request. OCSP_response_create()
14796 creates a response and optionally adds a basic response structure.
14797 OCSP_basic_add1_status() adds a complete single response to a basic
14798 response and returns the OCSP_SINGLERESP structure just added (to allow
14799 extensions to be included for example). OCSP_basic_add1_cert() adds a
14800 certificate to a basic response and OCSP_basic_sign() signs a basic
14801 response with various flags. New helper functions ASN1_TIME_check()
14802 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14803 (converts ASN1_TIME to GeneralizedTime).
14807 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14808 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14809 structure from a certificate. X509_pubkey_digest() digests the public_key
14810 contents: this is used in various key identifiers.
14814 * Make sk_sort() tolerate a NULL argument.
14816 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14818 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14819 passed by the function are trusted implicitly. If any of them signed the
14820 response then it is assumed to be valid and is not verified.
14824 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14825 to data. This was previously part of the PKCS7 ASN1 code. This
14826 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14827 *Steve Henson, reported by Kenneth R. Robinette
14828 <support@securenetterm.com>*
14830 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14831 routines: without these tracing memory leaks is very painful.
14832 Fix leaks in PKCS12 and PKCS7 routines.
14836 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14837 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14838 effectively meant GeneralizedTime would never be used. Now it
14839 is initialised to -1 but X509_time_adj() now has to check the value
14840 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14841 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14842 *Steve Henson, reported by Kenneth R. Robinette
14843 <support@securenetterm.com>*
14845 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14846 result in a zero length in the ASN1_INTEGER structure which was
14847 not consistent with the structure when d2i_ASN1_INTEGER() was used
14848 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14849 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14850 where it did not print out a minus for negative ASN1_INTEGER.
14854 * Add summary printout to ocsp utility. The various functions which
14855 convert status values to strings have been renamed to:
14856 OCSP_response_status_str(), OCSP_cert_status_str() and
14857 OCSP_crl_reason_str() and are no longer static. New options
14858 to verify nonce values and to disable verification. OCSP response
14859 printout format cleaned up.
14863 * Add additional OCSP certificate checks. These are those specified
14864 in RFC2560. This consists of two separate checks: the CA of the
14865 certificate being checked must either be the OCSP signer certificate
14866 or the issuer of the OCSP signer certificate. In the latter case the
14867 OCSP signer certificate must contain the OCSP signing extended key
14868 usage. This check is performed by attempting to match the OCSP
14869 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14870 in the OCSP_CERTID structures of the response.
14874 * Initial OCSP certificate verification added to OCSP_basic_verify()
14875 and related routines. This uses the standard OpenSSL certificate
14876 verify routines to perform initial checks (just CA validity) and
14877 to obtain the certificate chain. Then additional checks will be
14878 performed on the chain. Currently the root CA is checked to see
14879 if it is explicitly trusted for OCSP signing. This is used to set
14880 a root CA as a global signing root: that is any certificate that
14881 chains to that CA is an acceptable OCSP signing certificate.
14885 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14886 extensions from a separate configuration file.
14887 As when reading extensions from the main configuration file,
14888 the '-extensions ...' option may be used for specifying the
14891 *Massimiliano Pala <madwolf@comune.modena.it>*
14893 * New OCSP utility. Allows OCSP requests to be generated or
14894 read. The request can be sent to a responder and the output
14895 parsed, outputted or printed in text form. Not complete yet:
14896 still needs to check the OCSP response validity.
14900 * New subcommands for 'openssl ca':
14901 `openssl ca -status <serial>` prints the status of the cert with
14902 the given serial number (according to the index file).
14903 `openssl ca -updatedb` updates the expiry status of certificates
14906 *Massimiliano Pala <madwolf@comune.modena.it>*
14908 * New '-newreq-nodes' command option to CA.pl. This is like
14909 '-newreq', but calls 'openssl req' with the '-nodes' option
14910 so that the resulting key is not encrypted.
14912 *Damien Miller <djm@mindrot.org>*
14914 * New configuration for the GNU Hurd.
14916 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14918 * Initial code to implement OCSP basic response verify. This
14919 is currently incomplete. Currently just finds the signer's
14920 certificate and verifies the signature on the response.
14924 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14925 value of OPENSSLDIR. This is available via the new '-d' option
14926 to 'openssl version', and is also included in 'openssl version -a'.
14930 * Allowing defining memory allocation callbacks that will be given
14931 file name and line number information in additional arguments
14932 (a `const char*` and an int). The basic functionality remains, as
14933 well as the original possibility to just replace malloc(),
14934 realloc() and free() by functions that do not know about these
14935 additional arguments. To register and find out the current
14936 settings for extended allocation functions, the following
14937 functions are provided:
14939 CRYPTO_set_mem_ex_functions
14940 CRYPTO_set_locked_mem_ex_functions
14941 CRYPTO_get_mem_ex_functions
14942 CRYPTO_get_locked_mem_ex_functions
14944 These work the same way as CRYPTO_set_mem_functions and friends.
14945 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14946 extended allocation function is enabled.
14947 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14948 a conventional allocation function is enabled.
14950 *Richard Levitte, Bodo Moeller*
14952 * Finish off removing the remaining LHASH function pointer casts.
14953 There should no longer be any prototype-casting required when using
14954 the LHASH abstraction, and any casts that remain are "bugs". See
14955 the callback types and macros at the head of lhash.h for details
14956 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14960 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14961 If /dev/[u]random devices are not available or do not return enough
14962 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14964 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14965 /etc/entropy will be queried once each in this sequence, querying stops
14966 when enough entropy was collected without querying more sockets.
14970 * Change the Unix RAND_poll() variant to be able to poll several
14971 random devices, as specified by DEVRANDOM, until a sufficient amount
14972 of data has been collected. We spend at most 10 ms on each file
14973 (select timeout) and read in non-blocking mode. DEVRANDOM now
14974 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14975 (previously it was just the string "/dev/urandom"), so on typical
14976 platforms the 10 ms delay will never occur.
14977 Also separate out the Unix variant to its own file, rand_unix.c.
14978 For VMS, there's a currently-empty rand_vms.c.
14982 * Move OCSP client related routines to ocsp_cl.c. These
14983 provide utility functions which an application needing
14984 to issue a request to an OCSP responder and analyse the
14985 response will typically need: as opposed to those which an
14986 OCSP responder itself would need which will be added later.
14988 OCSP_request_sign() signs an OCSP request with an API similar
14989 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14990 response. OCSP_response_get1_basic() extracts basic response
14991 from response. OCSP_resp_find_status(): finds and extracts status
14992 information from an OCSP_CERTID structure (which will be created
14993 when the request structure is built). These are built from lower
14994 level functions which work on OCSP_SINGLERESP structures but
14995 won't normally be used unless the application wishes to examine
14996 extensions in the OCSP response for example.
14998 Replace nonce routines with a pair of functions.
14999 OCSP_request_add1_nonce() adds a nonce value and optionally
15000 generates a random value. OCSP_check_nonce() checks the
15001 validity of the nonce in an OCSP response.
15005 * Change function OCSP_request_add() to OCSP_request_add0_id().
15006 This doesn't copy the supplied OCSP_CERTID and avoids the
15007 need to free up the newly created id. Change return type
15008 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
15009 This can then be used to add extensions to the request.
15010 Deleted OCSP_request_new(), since most of its functionality
15011 is now in OCSP_REQUEST_new() (and the case insensitive name
15012 clash) apart from the ability to set the request name which
15013 will be added elsewhere.
15017 * Update OCSP API. Remove obsolete extensions argument from
15018 various functions. Extensions are now handled using the new
15019 OCSP extension code. New simple OCSP HTTP function which
15020 can be used to send requests and parse the response.
15024 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
15025 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
15026 uses the special reorder version of SET OF to sort the attributes
15027 and reorder them to match the encoded order. This resolves a long
15028 standing problem: a verify on a PKCS7 structure just after signing
15029 it used to fail because the attribute order did not match the
15030 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
15031 it uses the received order. This is necessary to tolerate some broken
15032 software that does not order SET OF. This is handled by encoding
15033 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
15034 to produce the required SET OF.
15038 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
15039 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
15040 files to get correct declarations of the ASN.1 item variables.
15044 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
15045 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
15046 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
15047 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
15048 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
15049 ASN1_ITEM and no wrapper functions.
15053 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
15054 replace the old function pointer based I/O routines. Change most of
15055 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
15059 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
15060 lines, recognize more "algorithms" that can be deselected, and make
15061 it complain about algorithm deselection that isn't recognised.
15065 * New ASN1 functions to handle dup, sign, verify, digest, pack and
15066 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
15067 to use new functions. Add NO_ASN1_OLD which can be set to remove
15068 some old style ASN1 functions: this can be used to determine if old
15069 code will still work when these eventually go away.
15073 * New extension functions for OCSP structures, these follow the
15074 same conventions as certificates and CRLs.
15078 * New function X509V3_add1_i2d(). This automatically encodes and
15079 adds an extension. Its behaviour can be customised with various
15080 flags to append, replace or delete. Various wrappers added for
15081 certificates and CRLs.
15085 * Fix to avoid calling the underlying ASN1 print routine when
15086 an extension cannot be parsed. Correct a typo in the
15087 OCSP_SERVICELOC extension. Tidy up print OCSP format.
15091 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
15092 entries for variables.
15096 * Add functionality to `apps/openssl.c` for detecting locking
15097 problems: As the program is single-threaded, all we have
15098 to do is register a locking callback using an array for
15099 storing which locks are currently held by the program.
15103 * Use a lock around the call to CRYPTO_get_ex_new_index() in
15104 SSL_get_ex_data_X509_STORE_idx(), which is used in
15105 ssl_verify_cert_chain() and thus can be called at any time
15106 during TLS/SSL handshakes so that thread-safety is essential.
15107 Unfortunately, the ex_data design is not at all suited
15108 for multi-threaded use, so it probably should be abolished.
15112 * Added Broadcom "ubsec" ENGINE to OpenSSL.
15114 *Broadcom, tweaked and integrated by Geoff Thorpe*
15116 * Move common extension printing code to new function
15117 X509V3_print_extensions(). Reorganise OCSP print routines and
15118 implement some needed OCSP ASN1 functions. Add OCSP extensions.
15122 * New function X509_signature_print() to remove duplication in some
15127 * Add a special meaning when SET OF and SEQUENCE OF flags are both
15128 set (this was treated exactly the same as SET OF previously). This
15129 is used to reorder the STACK representing the structure to match the
15130 encoding. This will be used to get round a problem where a PKCS7
15131 structure which was signed could not be verified because the STACK
15132 order did not reflect the encoded order.
15136 * Reimplement the OCSP ASN1 module using the new code.
15140 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
15141 for its ASN1 operations. The old style function pointers still exist
15142 for now but they will eventually go away.
15146 * Merge in replacement ASN1 code from the ASN1 branch. This almost
15147 completely replaces the old ASN1 functionality with a table driven
15148 encoder and decoder which interprets an ASN1_ITEM structure describing
15149 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
15150 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
15151 has also been converted to the new form.
15155 * Change BN_mod_exp_recp so that negative moduli are tolerated
15156 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
15157 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
15158 for negative moduli.
15162 * Fix BN_uadd and BN_usub: Always return non-negative results instead
15163 of not touching the result's sign bit.
15167 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
15172 * Changed the LHASH code to use prototypes for callbacks, and created
15173 macros to declare and implement thin (optionally static) functions
15174 that provide type-safety and avoid function pointer casting for the
15175 type-specific callbacks.
15179 * Added Kerberos Cipher Suites to be used with TLS, as written in
15181 *Veers Staats <staatsvr@asc.hpc.mil>,
15182 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
15184 * Reformat the FAQ so the different questions and answers can be divided
15185 in sections depending on the subject.
15189 * Have the zlib compression code load ZLIB.DLL dynamically under
15194 * New function BN_mod_sqrt for computing square roots modulo a prime
15195 (using the probabilistic Tonelli-Shanks algorithm unless
15196 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
15197 be handled deterministically).
15199 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15201 * Make BN_mod_inverse faster by explicitly handling small quotients
15202 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
15203 512 bits], about 30% for larger ones [1024 or 2048 bits].)
15207 * New function BN_kronecker.
15211 * Fix BN_gcd so that it works on negative inputs; the result is
15212 positive unless both parameters are zero.
15213 Previously something reasonably close to an infinite loop was
15214 possible because numbers could be growing instead of shrinking
15215 in the implementation of Euclid's algorithm.
15219 * Fix BN_is_word() and BN_is_one() macros to take into account the
15220 sign of the number in question.
15222 Fix BN_is_word(a,w) to work correctly for w == 0.
15224 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
15225 because its test if the absolute value of 'a' equals 'w'.
15226 Note that BN_abs_is_word does *not* handle w == 0 reliably;
15227 it exists mostly for use in the implementations of BN_is_zero(),
15228 BN_is_one(), and BN_is_word().
15232 * New function BN_swap.
15236 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
15237 the exponentiation functions are more likely to produce reasonable
15238 results on negative inputs.
15242 * Change BN_mod_mul so that the result is always non-negative.
15243 Previously, it could be negative if one of the factors was negative;
15244 I don't think anyone really wanted that behaviour.
15248 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
15249 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
15250 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
15251 and add new functions:
15260 BN_mod_lshift1_quick
15262 BN_mod_lshift_quick
15264 These functions always generate non-negative results.
15266 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15267 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15269 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15270 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15271 be reduced modulo `m`.
15273 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15276 The following entry accidentally appeared in the CHANGES file
15277 distributed with OpenSSL 0.9.7. The modifications described in
15278 it do *not* apply to OpenSSL 0.9.7.
15280 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15281 was actually never needed) and in BN_mul(). The removal in BN_mul()
15282 required a small change in bn_mul_part_recursive() and the addition
15283 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15284 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15285 bn_sub_words() and bn_add_words() except they take arrays with
15291 * In 'openssl passwd', verify passwords read from the terminal
15292 unless the '-salt' option is used (which usually means that
15293 verification would just waste user's time since the resulting
15294 hash is going to be compared with some given password hash)
15295 or the new '-noverify' option is used.
15297 This is an incompatible change, but it does not affect
15298 non-interactive use of 'openssl passwd' (passwords on the command
15299 line, '-stdin' option, '-in ...' option) and thus should not
15300 cause any problems.
15304 * Remove all references to RSAref, since there's no more need for it.
15308 * Make DSO load along a path given through an environment variable
15309 (SHLIB_PATH) with shl_load().
15313 * Constify the ENGINE code as a result of BIGNUM constification.
15314 Also constify the RSA code and most things related to it. In a
15315 few places, most notable in the depth of the ASN.1 code, ugly
15316 casts back to non-const were required (to be solved at a later
15321 * Make it so the openssl application has all engines loaded by default.
15325 * Constify the BIGNUM routines a little more.
15329 * Add the following functions:
15331 ENGINE_load_cswift()
15333 ENGINE_load_atalla()
15334 ENGINE_load_nuron()
15335 ENGINE_load_builtin_engines()
15337 That way, an application can itself choose if external engines that
15338 are built-in in OpenSSL shall ever be used or not. The benefit is
15339 that applications won't have to be linked with libdl or other dso
15340 libraries unless it's really needed.
15342 Changed 'openssl engine' to load all engines on demand.
15343 Changed the engine header files to avoid the duplication of some
15344 declarations (they differed!).
15348 * 'openssl engine' can now list capabilities.
15352 * Better error reporting in 'openssl engine'.
15356 * Never call load_dh_param(NULL) in s_server.
15360 * Add engine application. It can currently list engines by name and
15361 identity, and test if they are actually available.
15365 * Improve RPM specification file by forcing symbolic linking and making
15366 sure the installed documentation is also owned by root.root.
15368 *Damien Miller <djm@mindrot.org>*
15370 * Give the OpenSSL applications more possibilities to make use of
15371 keys (public as well as private) handled by engines.
15375 * Add OCSP code that comes from CertCo.
15379 * Add VMS support for the Rijndael code.
15383 * Added untested support for Nuron crypto accelerator.
15387 * Add support for external cryptographic devices. This code was
15388 previously distributed separately as the "engine" branch.
15390 *Geoff Thorpe, Richard Levitte*
15392 * Rework the filename-translation in the DSO code. It is now possible to
15393 have far greater control over how a "name" is turned into a filename
15394 depending on the operating environment and any oddities about the
15395 different shared library filenames on each system.
15399 * Support threads on FreeBSD-elf in Configure.
15403 * Fix for SHA1 assembly problem with MASM: it produces
15404 warnings about corrupt line number information when assembling
15405 with debugging information. This is caused by the overlapping
15408 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15411 NCONF_get_number() has no error checking at all. As a replacement,
15412 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15413 promoted strongly. The old NCONF_get_number is kept around for
15414 binary backward compatibility.
15415 Make it possible for methods to load from something other than a BIO,
15416 by providing a function pointer that is given a name instead of a BIO.
15417 For example, this could be used to load configuration data from an
15422 * Fix for non blocking accept BIOs. Added new I/O special reason
15423 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15424 with non blocking I/O was not possible because no retry code was
15425 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15430 * Added the beginnings of Rijndael support.
15434 * Fix for bug in DirectoryString mask setting. Add support for
15435 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15436 to allow certificate printing to more controllable, additional
15437 'certopt' option to 'x509' to allow new printing options to be
15442 * Clean old EAY MD5 hack from e_os.h.
15446 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15448 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15449 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15451 *Joe Orton, Steve Henson*
15453 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15455 * Fix additional bug revealed by the NISCC test suite:
15457 Stop bug triggering large recursion when presented with
15458 certain ASN.1 tags ([CVE-2003-0851])
15462 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15464 * Fix various bugs revealed by running the NISCC test suite:
15466 Stop out of bounds reads in the ASN1 code when presented with
15467 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15469 If verify callback ignores invalid public key errors don't try to check
15470 certificate signature with the NULL public key.
15474 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15475 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15480 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15481 extra data after the compression methods not only for TLS 1.0
15482 but also for SSL 3.0 (as required by the specification).
15484 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15486 * Change X509_certificate_type() to mark the key as exported/exportable
15487 when it's 512 *bits* long, not 512 bytes.
15491 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15493 * Countermeasure against the Klima-Pokorny-Rosa extension of
15494 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15495 a protocol version number mismatch like a decryption error
15496 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15500 * Turn on RSA blinding by default in the default implementation
15501 to avoid a timing attack. Applications that don't want it can call
15502 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15503 They would be ill-advised to do so in most cases.
15505 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15507 * Change RSA blinding code so that it works when the PRNG is not
15508 seeded (in this case, the secret RSA exponent is abused as
15509 an unpredictable seed -- if it is not unpredictable, there
15510 is no point in blinding anyway). Make RSA blinding thread-safe
15511 by remembering the creator's thread ID in rsa->blinding and
15512 having all other threads use local one-time blinding factors
15513 (this requires more computation than sharing rsa->blinding, but
15514 avoids excessive locking; and if an RSA object is not shared
15515 between threads, blinding will still be very fast).
15519 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15521 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15522 via timing by performing a MAC computation even if incorrect
15523 block cipher padding has been found. This is a countermeasure
15524 against active attacks where the attacker has to distinguish
15525 between bad padding and a MAC verification error. ([CVE-2003-0078])
15527 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15528 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15529 Martin Vuagnoux (EPFL, Ilion)*
15531 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15533 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15534 memory from its contents. This is done with a counter that will
15535 place alternating values in each byte. This can be used to solve
15536 two issues: 1) the removal of calls to memset() by highly optimizing
15537 compilers, and 2) cleansing with other values than 0, since those can
15538 be read through on certain media, for example a swap space on disk.
15542 * Bugfix: client side session caching did not work with external caching,
15543 because the session->cipher setting was not restored when reloading
15544 from the external cache. This problem was masked, when
15545 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15546 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15550 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15551 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15553 *Zeev Lieber <zeev-l@yahoo.com>*
15555 * Undo an undocumented change introduced in 0.9.6e which caused
15556 repeated calls to OpenSSL_add_all_ciphers() and
15557 OpenSSL_add_all_digests() to be ignored, even after calling
15562 * Change the default configuration reader to deal with last line not
15563 being properly terminated.
15567 * Change X509_NAME_cmp() so it applies the special rules on handling
15568 DN values that are of type PrintableString, as well as RDNs of type
15569 emailAddress where the value has the type ia5String.
15571 *stefank@valicert.com via Richard Levitte*
15573 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15574 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15575 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15576 the bitwise-OR of the two for use by the majority of applications
15577 wanting this behaviour, and update the docs. The documented
15578 behaviour and actual behaviour were inconsistent and had been
15579 changing anyway, so this is more a bug-fix than a behavioural
15582 *Geoff Thorpe, diagnosed by Nadav Har'El*
15584 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15585 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15589 * Fix initialization code race conditions in
15590 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15591 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15592 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15593 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15594 ssl2_get_cipher_by_char(),
15595 ssl3_get_cipher_by_char().
15597 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15599 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15600 the cached sessions are flushed, as the remove_cb() might use ex_data
15601 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15602 (see [openssl.org #212]).
15604 *Geoff Thorpe, Lutz Jaenicke*
15606 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15607 length, instead of the encoding length to d2i_ASN1_OBJECT.
15611 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15613 * [In 0.9.6g-engine release:]
15614 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15616 *Lynn Gazis <lgazis@rainbow.com>*
15618 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15620 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15621 and get fix the header length calculation.
15622 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15623 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15625 * Use proper error handling instead of 'assertions' in buffer
15626 overflow checks added in 0.9.6e. This prevents DoS (the
15627 assertions could call abort()).
15629 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15631 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15633 * Add various sanity checks to asn1_get_length() to reject
15634 the ASN1 length bytes if they exceed sizeof(long), will appear
15635 negative or the content length exceeds the length of the
15638 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15640 * Fix cipher selection routines: ciphers without encryption had no flags
15641 for the cipher strength set and where therefore not handled correctly
15642 by the selection routines (PR #130).
15646 * Fix EVP_dsa_sha macro.
15651 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15652 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15653 that was added in OpenSSL 0.9.6d.
15655 As the countermeasure turned out to be incompatible with some
15656 broken SSL implementations, the new option is part of SSL_OP_ALL.
15657 SSL_OP_ALL is usually employed when compatibility with weird SSL
15658 implementations is desired (e.g. '-bugs' option to 's_client' and
15659 's_server'), so the new option is automatically set in many
15664 * Changes in security patch:
15666 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15667 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15668 Air Force Materiel Command, USAF, under agreement number
15671 * Add various sanity checks to asn1_get_length() to reject
15672 the ASN1 length bytes if they exceed sizeof(long), will appear
15673 negative or the content length exceeds the length of the
15674 supplied buffer. ([CVE-2002-0659])
15676 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15678 * Assertions for various potential buffer overflows, not known to
15679 happen in practice.
15681 *Ben Laurie (CHATS)*
15683 * Various temporary buffers to hold ASCII versions of integers were
15684 too small for 64 bit platforms. ([CVE-2002-0655])
15685 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15687 * Remote buffer overflow in SSL3 protocol - an attacker could
15688 supply an oversized session ID to a client. ([CVE-2002-0656])
15690 *Ben Laurie (CHATS)*
15692 * Remote buffer overflow in SSL2 protocol - an attacker could
15693 supply an oversized client master key. ([CVE-2002-0656])
15695 *Ben Laurie (CHATS)*
15697 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15699 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15700 encoded as NULL) with id-dsa-with-sha1.
15702 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15704 * Check various `X509_...()` return values in `apps/req.c`.
15706 *Nils Larsch <nla@trustcenter.de>*
15708 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15709 an end-of-file condition would erroneously be flagged, when the CRLF
15710 was just at the end of a processed block. The bug was discovered when
15711 processing data through a buffering memory BIO handing the data to a
15712 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15713 <ptsekov@syntrex.com> and Nedelcho Stanev.
15717 * Implement a countermeasure against a vulnerability recently found
15718 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15719 before application data chunks to avoid the use of known IVs
15720 with data potentially chosen by the attacker.
15724 * Fix length checks in ssl3_get_client_hello().
15728 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15729 to prevent ssl3_read_internal() from incorrectly assuming that
15730 ssl3_read_bytes() found application data while handshake
15731 processing was enabled when in fact s->s3->in_read_app_data was
15732 merely automatically cleared during the initial handshake.
15734 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15736 * Fix object definitions for Private and Enterprise: they were not
15737 recognized in their shortname (=lowercase) representation. Extend
15738 obj_dat.pl to issue an error when using undefined keywords instead
15739 of silently ignoring the problem (Svenning Sorensen
15740 <sss@sss.dnsalias.net>).
15744 * Fix DH_generate_parameters() so that it works for 'non-standard'
15745 generators, i.e. generators other than 2 and 5. (Previously, the
15746 code did not properly initialise the 'add' and 'rem' values to
15747 BN_generate_prime().)
15749 In the new general case, we do not insist that 'generator' is
15750 actually a primitive root: This requirement is rather pointless;
15751 a generator of the order-q subgroup is just as good, if not
15756 * Map new X509 verification errors to alerts. Discovered and submitted by
15757 Tom Wu <tom@arcot.com>.
15761 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15762 returning non-zero before the data has been completely received
15763 when using non-blocking I/O.
15765 *Bodo Moeller; problem pointed out by John Hughes*
15767 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15769 *Ben Laurie, Lutz Jaenicke*
15771 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15772 Yoram Zahavi <YoramZ@gilian.com>).
15776 * Add information about CygWin 1.3 and on, and preserve proper
15777 configuration for the versions before that.
15779 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15781 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15782 check whether we deal with a copy of a session and do not delete from
15783 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15784 <izhar@checkpoint.com>.
15788 * Do not store session data into the internal session cache, if it
15789 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15790 flag is set). Proposed by Aslam <aslam@funk.com>.
15794 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15799 * [In 0.9.6d-engine release:]
15800 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15802 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15804 * Add the configuration target linux-s390x.
15806 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15808 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15809 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15810 variable as an indication that a ClientHello message has been
15811 received. As the flag value will be lost between multiple
15812 invocations of ssl3_accept when using non-blocking I/O, the
15813 function may not be aware that a handshake has actually taken
15814 place, thus preventing a new session from being added to the
15817 To avoid this problem, we now set s->new_session to 2 instead of
15818 using a local variable.
15820 *Lutz Jaenicke, Bodo Moeller*
15822 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15823 if the SSL_R_LENGTH_MISMATCH error is detected.
15825 *Geoff Thorpe, Bodo Moeller*
15827 * New 'shared_ldflag' column in Configure platform table.
15831 * Fix EVP_CIPHER_mode macro.
15833 *"Dan S. Camper" <dan@bti.net>*
15835 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15836 type, we must throw them away by setting rr->length to 0.
15838 *D P Chang <dpc@qualys.com>*
15840 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15842 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15843 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15844 worked incorrectly for those cases where range = `10..._2` and
15845 `3*range` is two bits longer than range.)
15849 * Only add signing time to PKCS7 structures if it is not already
15854 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15855 OBJ_ld_ce should be OBJ_id_ce.
15856 Also some ip-pda OIDs in crypto/objects/objects.txt were
15857 incorrect (cf. RFC 3039).
15859 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15861 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15862 returns early because it has nothing to do.
15864 *Andy Schneider <andy.schneider@bjss.co.uk>*
15866 * [In 0.9.6c-engine release:]
15867 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15869 *Andy Schneider <andy.schneider@bjss.co.uk>*
15871 * [In 0.9.6c-engine release:]
15872 Add support for Cryptographic Appliance's keyserver technology.
15873 (Use engine 'keyclient')
15875 *Cryptographic Appliances and Geoff Thorpe*
15877 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15878 is called via tools/c89.sh because arguments have to be
15879 rearranged (all '-L' options must appear before the first object
15882 *Richard Shapiro <rshapiro@abinitio.com>*
15884 * [In 0.9.6c-engine release:]
15885 Add support for Broadcom crypto accelerator cards, backported
15888 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15890 * [In 0.9.6c-engine release:]
15891 Add support for SureWare crypto accelerator cards from
15892 Baltimore Technologies. (Use engine 'sureware')
15894 *Baltimore Technologies and Mark Cox*
15896 * [In 0.9.6c-engine release:]
15897 Add support for crypto accelerator cards from Accelerated
15898 Encryption Processing, www.aep.ie. (Use engine 'aep')
15900 *AEP Inc. and Mark Cox*
15902 * Add a configuration entry for gcc on UnixWare.
15904 *Gary Benson <gbenson@redhat.com>*
15906 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15907 messages are stored in a single piece (fixed-length part and
15908 variable-length part combined) and fix various bugs found on the way.
15912 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15913 instead. BIO_gethostbyname() does not know what timeouts are
15914 appropriate, so entries would stay in cache even when they have
15916 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15918 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15919 faced with a pathologically small ClientHello fragment that does
15920 not contain client_version: Instead of aborting with an error,
15921 simply choose the highest available protocol version (i.e.,
15922 TLS 1.0 unless it is disabled). In practice, ClientHello
15923 messages are never sent like this, but this change gives us
15924 strictly correct behaviour at least for TLS.
15928 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15929 never resets s->method to s->ctx->method when called from within
15930 one of the SSL handshake functions.
15932 *Bodo Moeller; problem pointed out by Niko Baric*
15934 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15935 (sent using the client's version number) if client_version is
15936 smaller than the protocol version in use. Also change
15937 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15938 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15939 the client will at least see that alert.
15943 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15948 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15949 client receives HelloRequest while in a handshake.
15951 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15953 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15954 should end in 'break', not 'goto end' which circumvents various
15955 cleanups done in state SSL_ST_OK. But session related stuff
15956 must be disabled for SSL_ST_OK in the case that we just sent a
15959 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15960 before just sending a HelloRequest.
15962 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15964 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15965 reveal whether illegal block cipher padding was found or a MAC
15966 verification error occurred. (Neither SSLerr() codes nor alerts
15967 are directly visible to potential attackers, but the information
15968 may leak via logfiles.)
15970 Similar changes are not required for the SSL 2.0 implementation
15971 because the number of padding bytes is sent in clear for SSL 2.0,
15972 and the extra bytes are just ignored. However ssl/s2_pkt.c
15973 failed to verify that the purported number of padding bytes is in
15978 * Add OpenUNIX-8 support including shared libraries
15979 (Boyd Lynn Gerber <gerberb@zenez.com>).
15983 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15984 'wristwatch attack' using huge encoding parameters (cf.
15985 James H. Manger's CRYPTO 2001 paper). Note that the
15986 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15987 encoding parameters and hence was not vulnerable.
15991 * BN_sqr() bug fix.
15993 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15995 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15996 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15997 followed by modular reduction.
15999 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
16001 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
16002 equivalent based on BN_pseudo_rand() instead of BN_rand().
16006 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
16007 This function was broken, as the check for a new client hello message
16008 to handle SGC did not allow these large messages.
16009 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
16013 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
16017 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
16018 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
16022 * Rework the configuration and shared library support for Tru64 Unix.
16023 The configuration part makes use of modern compiler features and
16024 still retains old compiler behavior for those that run older versions
16025 of the OS. The shared library support part includes a variant that
16026 uses the RPATH feature, and is available through the special
16027 configuration target "alpha-cc-rpath", which will never be selected
16030 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
16032 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
16033 with the same message size as in ssl3_get_certificate_request().
16034 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
16035 messages might inadvertently be reject as too long.
16037 *Petr Lampa <lampa@fee.vutbr.cz>*
16039 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
16043 * Modified SSL library such that the verify_callback that has been set
16044 specifically for an SSL object with SSL_set_verify() is actually being
16045 used. Before the change, a verify_callback set with this function was
16046 ignored and the verify_callback() set in the SSL_CTX at the time of
16047 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
16048 to allow the necessary settings.
16052 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
16053 explicitly to NULL, as at least on Solaris 8 this seems not always to be
16054 done automatically (in contradiction to the requirements of the C
16055 standard). This made problems when used from OpenSSH.
16059 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
16060 dh->length and always used
16062 BN_rand_range(priv_key, dh->p).
16064 BN_rand_range() is not necessary for Diffie-Hellman, and this
16065 specific range makes Diffie-Hellman unnecessarily inefficient if
16066 dh->length (recommended exponent length) is much smaller than the
16067 length of dh->p. We could use BN_rand_range() if the order of
16068 the subgroup was stored in the DH structure, but we only have
16073 BN_rand(priv_key, l, ...)
16075 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
16082 RSA_eay_public_encrypt
16083 RSA_eay_private_decrypt
16084 RSA_eay_private_encrypt (signing)
16085 RSA_eay_public_decrypt (signature verification)
16087 (default implementations for RSA_public_encrypt,
16088 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
16089 always reject numbers >= n.
16093 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
16094 to synchronize access to 'locking_thread'. This is necessary on
16095 systems where access to 'locking_thread' (an 'unsigned long'
16096 variable) is not atomic.
16100 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
16101 *before* setting the 'crypto_lock_rand' flag. The previous code had
16102 a race condition if 0 is a valid thread ID.
16104 *Travis Vitek <vitek@roguewave.com>*
16106 * Add support for shared libraries under Irix.
16108 *Albert Chin-A-Young <china@thewrittenword.com>*
16110 * Add configuration option to build on Linux on both big-endian and
16111 little-endian MIPS.
16113 *Ralf Baechle <ralf@uni-koblenz.de>*
16115 * Add the possibility to create shared libraries on HP-UX.
16119 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
16121 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
16122 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
16123 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
16124 PRNG state recovery was possible based on the output of
16125 one PRNG request appropriately sized to gain knowledge on
16126 'md' followed by enough consecutive 1-byte PRNG requests
16127 to traverse all of 'state'.
16129 1. When updating 'md_local' (the current thread's copy of 'md')
16130 during PRNG output generation, hash all of the previous
16131 'md_local' value, not just the half used for PRNG output.
16133 2. Make the number of bytes from 'state' included into the hash
16134 independent from the number of PRNG bytes requested.
16136 The first measure alone would be sufficient to avoid
16137 Markku-Juhani's attack. (Actually it had never occurred
16138 to me that the half of 'md_local' used for chaining was the
16139 half from which PRNG output bytes were taken -- I had always
16140 assumed that the secret half would be used.) The second
16141 measure makes sure that additional data from 'state' is never
16142 mixed into 'md_local' in small portions; this heuristically
16143 further strengthens the PRNG.
16147 * Fix crypto/bn/asm/mips3.s.
16151 * When only the key is given to "enc", the IV is undefined. Print out
16152 an error message in this case.
16156 * Handle special case when X509_NAME is empty in X509 printing routines.
16160 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
16161 positive and less than q.
16165 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
16166 used: it isn't thread safe and the add_lock_callback should handle
16169 *Paul Rose <Paul.Rose@bridge.com>*
16171 * Verify that incoming data obeys the block size in
16172 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
16178 *Ulf Möller, Bodo Möller*
16180 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
16181 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
16182 when fixing the server behaviour for backwards-compatible 'client
16183 hello' messages. (Note that the attack is impractical against
16184 SSL 3.0 and TLS 1.0 anyway because length and version checking
16185 means that the probability of guessing a valid ciphertext is
16186 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
16189 Before 0.9.5, the countermeasure (hide the error by generating a
16190 random 'decryption result') did not work properly because
16191 ERR_clear_error() was missing, meaning that SSL_get_error() would
16192 detect the supposedly ignored error.
16194 Both problems are now fixed.
16198 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
16199 (previously it was 1024).
16203 * Fix for compatibility mode trust settings: ignore trust settings
16204 unless some valid trust or reject settings are present.
16208 * Fix for blowfish EVP: its a variable length cipher.
16212 * Fix various bugs related to DSA S/MIME verification. Handle missing
16213 parameters in DSA public key structures and return an error in the
16214 DSA routines if parameters are absent.
16218 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
16219 in the current directory if neither $RANDFILE nor $HOME was set.
16220 RAND_file_name() in 0.9.6a returned NULL in this case. This has
16221 caused some confusion to Windows users who haven't defined $HOME.
16222 Thus RAND_file_name() is changed again: e_os.h can define a
16223 DEFAULT_HOME, which will be used if $HOME is not set.
16224 For Windows, we use "C:"; on other platforms, we still require
16225 environment variables.
16227 * Move 'if (!initialized) RAND_poll()' into regions protected by
16228 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
16229 having multiple threads call RAND_poll() concurrently.
16233 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
16234 combination of a flag and a thread ID variable.
16235 Otherwise while one thread is in ssleay_rand_bytes (which sets the
16236 flag), *other* threads can enter ssleay_add_bytes without obeying
16237 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
16238 that they do not hold after the first thread unsets add_do_not_lock).
16242 * Change bctest again: '-x' expressions are not available in all
16243 versions of 'test'.
16247 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
16249 * Fix a couple of memory leaks in PKCS7_dataDecode()
16251 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
16253 * Change Configure and Makefiles to provide EXE_EXT, which will contain
16254 the default extension for executables, if any. Also, make the perl
16255 scripts that use symlink() to test if it really exists and use "cp"
16256 if it doesn't. All this made OpenSSL compilable and installable in
16261 * Fix for asn1_GetSequence() for indefinite length constructed data.
16262 If SEQUENCE is length is indefinite just set c->slen to the total
16263 amount of data available.
16265 *Steve Henson, reported by shige@FreeBSD.org*
16267 *This change does not apply to 0.9.7.*
16269 * Change bctest to avoid here-documents inside command substitution
16270 (workaround for FreeBSD /bin/sh bug).
16271 For compatibility with Ultrix, avoid shell functions (introduced
16272 in the bctest version that searches along $PATH).
16276 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16277 with des_encrypt() defined on some operating systems, like Solaris
16282 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16283 On the Importance of Eliminating Errors in Cryptographic
16284 Computations, J. Cryptology 14 (2001) 2, 101-119,
16285 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16289 * MIPS assembler BIGNUM division bug fix.
16293 * Disabled incorrect Alpha assembler code.
16297 * Fix PKCS#7 decode routines so they correctly update the length
16298 after reading an EOC for the EXPLICIT tag.
16302 *This change does not apply to 0.9.7.*
16304 * Fix bug in PKCS#12 key generation routines. This was triggered
16305 if a 3DES key was generated with a 0 initial byte. Include
16306 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16307 (but broken) behaviour.
16311 * Enhance bctest to search for a working bc along $PATH and print
16314 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16316 * Fix memory leaks in err.c: free err_data string if necessary;
16317 don't write to the wrong index in ERR_set_error_data.
16321 * Implement ssl23_peek (analogous to ssl23_read), which previously
16326 * Replace rdtsc with `_emit` statements for VC++ version 5.
16328 *Jeremy Cooper <jeremy@baymoo.org>*
16330 * Make it possible to reuse SSLv2 sessions.
16334 * In copy_email() check for >= 0 as a return value for
16335 X509_NAME_get_index_by_NID() since 0 is a valid index.
16337 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16339 * Avoid coredump with unsupported or invalid public keys by checking if
16340 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16341 PKCS7_verify() fails with non detached data.
16345 * Don't use getenv in library functions when run as setuid/setgid.
16346 New function OPENSSL_issetugid().
16350 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16351 due to incorrect handling of multi-threading:
16353 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16355 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16357 3. Count how many times MemCheck_off() has been called so that
16358 nested use can be treated correctly. This also avoids
16359 inband-signalling in the previous code (which relied on the
16360 assumption that thread ID 0 is impossible).
16364 * Add "-rand" option also to s_client and s_server.
16368 * Fix CPU detection on Irix 6.x.
16369 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16370 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16372 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16377 *This change does not apply to 0.9.7.*
16379 * Use the cached encoding of an X509_NAME structure rather than
16380 copying it. This is apparently the reason for the libsafe "errors"
16381 but the code is actually correct.
16385 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16386 Bleichenbacher's DSA attack.
16387 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16388 to be set and top=0 forces the highest bit to be set; top=-1 is new
16389 and leaves the highest bit random.
16391 *Ulf Moeller, Bodo Moeller*
16393 * In the `NCONF_...`-based implementations for `CONF_...` queries
16394 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16395 a temporary CONF structure with the data component set to NULL
16396 (which gives segmentation faults in lh_retrieve).
16397 Instead, use NULL for the CONF pointer in CONF_get_string and
16398 CONF_get_number (which may use environment variables) and directly
16399 return NULL from CONF_get_section.
16403 * Fix potential buffer overrun for EBCDIC.
16407 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16408 keyUsage if basicConstraints absent for a CA.
16412 * Make SMIME_write_PKCS7() write mail header values with a format that
16413 is more generally accepted (no spaces before the semicolon), since
16414 some programs can't parse those values properly otherwise. Also make
16415 sure BIO's that break lines after each write do not create invalid
16420 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16421 macros previously used would not encode an empty SEQUENCE OF
16422 and break the signature.
16426 *This change does not apply to 0.9.7.*
16428 * Zero the premaster secret after deriving the master secret in
16433 * Add some EVP_add_digest_alias registrations (as found in
16434 OpenSSL_add_all_digests()) to SSL_library_init()
16435 aka OpenSSL_add_ssl_algorithms(). This provides improved
16436 compatibility with peers using X.509 certificates
16437 with unconventional AlgorithmIdentifier OIDs.
16441 * Fix for Irix with NO_ASM.
16443 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16445 * ./config script fixes.
16447 *Ulf Moeller, Richard Levitte*
16449 * Fix 'openssl passwd -1'.
16453 * Change PKCS12_key_gen_asc() so it can cope with non null
16454 terminated strings whose length is passed in the passlen
16455 parameter, for example from PEM callbacks. This was done
16456 by adding an extra length parameter to asc2uni().
16458 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16460 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16461 call failed, free the DSA structure.
16465 * Fix to uni2asc() to cope with zero length Unicode strings.
16466 These are present in some PKCS#12 files.
16470 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16471 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16472 when writing a 32767 byte record.
16474 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16476 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16477 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16479 (RSA objects have a reference count access to which is protected
16480 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16481 so they are meant to be shared between threads.)
16482 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16483 "Reddie, Steven" <Steven.Reddie@ca.com>*
16485 * Fix a deadlock in CRYPTO_mem_leaks().
16489 * Use better test patterns in bntest.
16493 * rand_win.c fix for Borland C.
16497 * BN_rshift bugfix for n == 0.
16501 * Add a 'bctest' script that checks for some known 'bc' bugs
16502 so that 'make test' does not abort just because 'bc' is broken.
16506 * Store verify_result within SSL_SESSION also for client side to
16507 avoid potential security hole. (Re-used sessions on the client side
16508 always resulted in verify_result==X509_V_OK, not using the original
16509 result of the server certificate verification.)
16513 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16514 SSL3_RT_APPLICATION_DATA, return 0.
16515 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16520 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16521 releases, have been re-implemented by renaming the previous
16522 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16523 and ssl3_read_internal, respectively, and adding 'peek' parameters
16524 to them. The new ssl[23]_{read,peek} functions are calls to
16525 ssl[23]_read_internal with the 'peek' flag set appropriately.
16526 A 'peek' parameter has also been added to ssl3_read_bytes, which
16527 does the actual work for ssl3_read_internal.
16531 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16532 the method-specific "init()" handler. Also clean up ex_data after
16533 calling the method-specific "finish()" handler. Previously, this was
16534 happening the other way round.
16538 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16539 The previous value, 12, was not always sufficient for BN_mod_exp().
16543 * Make sure that shared libraries get the internal name engine with
16544 the full version number and not just 0. This should mark the
16545 shared libraries as not backward compatible. Of course, this should
16546 be changed again when we can guarantee backward binary compatibility.
16550 * Fix typo in get_cert_by_subject() in by_dir.c
16552 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16554 * Rework the system to generate shared libraries:
16556 - Make note of the expected extension for the shared libraries and
16557 if there is a need for symbolic links from for example libcrypto.so.0
16558 to libcrypto.so.0.9.7. There is extended info in Configure for
16561 - Make as few rebuilds of the shared libraries as possible.
16563 - Still avoid linking the OpenSSL programs with the shared libraries.
16565 - When installing, install the shared libraries separately from the
16570 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16572 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16573 and not in SSL_clear because the latter is also used by the
16574 accept/connect functions; previously, the settings made by
16575 SSL_set_read_ahead would be lost during the handshake.
16577 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16579 * Correct util/mkdef.pl to be selective about disabled algorithms.
16580 Previously, it would create entries for disabled algorithms no
16585 * Added several new manual pages for SSL_* function.
16589 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16591 * In ssl23_get_client_hello, generate an error message when faced
16592 with an initial SSL 3.0/TLS record that is too small to contain the
16593 first two bytes of the ClientHello message, i.e. client_version.
16594 (Note that this is a pathologic case that probably has never happened
16595 in real life.) The previous approach was to use the version number
16596 from the record header as a substitute; but our protocol choice
16597 should not depend on that one because it is not authenticated
16598 by the Finished messages.
16602 * More robust randomness gathering functions for Windows.
16604 *Jeffrey Altman <jaltman@columbia.edu>*
16606 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16607 not set then we don't setup the error code for issuer check errors
16608 to avoid possibly overwriting other errors which the callback does
16609 handle. If an application does set the flag then we assume it knows
16610 what it is doing and can handle the new informational codes
16615 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16616 a general "ANY" type, as such it should be able to decode anything
16617 including tagged types. However it didn't check the class so it would
16618 wrongly interpret tagged types in the same way as their universal
16619 counterpart and unknown types were just rejected. Changed so that the
16620 tagged and unknown types are handled in the same way as a SEQUENCE:
16621 that is the encoding is stored intact. There is also a new type
16622 "V_ASN1_OTHER" which is used when the class is not universal, in this
16623 case we have no idea what the actual type is so we just lump them all
16628 * On VMS, stdout may very well lead to a file that is written to
16629 in a record-oriented fashion. That means that every write() will
16630 write a separate record, which will be read separately by the
16631 programs trying to read from it. This can be very confusing.
16633 The solution is to put a BIO filter in the way that will buffer
16634 text until a linefeed is reached, and then write everything a
16635 line at a time, so every record written will be an actual line,
16636 not chunks of lines and not (usually doesn't happen, but I've
16637 seen it once) several lines in one record. BIO_f_linebuffer() is
16640 Currently, it's a VMS-only method, because that's where it has
16641 been tested well enough.
16645 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16646 it can return incorrect results.
16647 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16648 but it was in 0.9.6-beta[12].)
16652 * Disable the check for content being present when verifying detached
16653 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16654 include zero length content when signing messages.
16658 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16659 BIO_ctrl (for BIO pairs).
16663 * Add DSO method for VMS.
16667 * Bug fix: Montgomery multiplication could produce results with the
16672 * Add RPM specification openssl.spec and modify it to build three
16673 packages. The default package contains applications, application
16674 documentation and run-time libraries. The devel package contains
16675 include files, static libraries and function documentation. The
16676 doc package contains the contents of the doc directory. The original
16677 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16681 * Add a large number of documentation files for many SSL routines.
16683 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16685 * Add a configuration entry for Sony News 4.
16687 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16689 * Don't set the two most significant bits to one when generating a
16690 random number < q in the DSA library.
16694 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16695 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16696 the underlying transport is blocking) if a handshake took place.
16697 (The default behaviour is needed by applications such as s_client
16698 and s_server that use select() to determine when to use SSL_read;
16699 but for applications that know in advance when to expect data, it
16700 just makes things more complicated.)
16704 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16709 * Add a few more EBCDIC conditionals that make `req` and `x509`
16710 work better on such systems.
16712 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16714 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16715 Update PKCS12_parse() so it copies the friendlyName and the
16716 keyid to the certificates aux info.
16720 * Fix bug in PKCS7_verify() which caused an infinite loop
16721 if there was more than one signature.
16723 *Sven Uszpelkat <su@celocom.de>*
16725 * Major change in util/mkdef.pl to include extra information
16726 about each symbol, as well as presenting variables as well
16727 as functions. This change means that there's n more need
16728 to rebuild the .num files when some algorithms are excluded.
16732 * Allow the verify time to be set by an application,
16733 rather than always using the current time.
16737 * Phase 2 verify code reorganisation. The certificate
16738 verify code now looks up an issuer certificate by a
16739 number of criteria: subject name, authority key id
16740 and key usage. It also verifies self signed certificates
16741 by the same criteria. The main comparison function is
16742 X509_check_issued() which performs these checks.
16744 Lot of changes were necessary in order to support this
16745 without completely rewriting the lookup code.
16747 Authority and subject key identifier are now cached.
16749 The LHASH 'certs' is X509_STORE has now been replaced
16750 by a STACK_OF(X509_OBJECT). This is mainly because an
16751 LHASH can't store or retrieve multiple objects with
16752 the same hash value.
16754 As a result various functions (which were all internal
16755 use only) have changed to handle the new X509_STORE
16756 structure. This will break anything that messed round
16757 with X509_STORE internally.
16759 The functions X509_STORE_add_cert() now checks for an
16760 exact match, rather than just subject name.
16762 The X509_STORE API doesn't directly support the retrieval
16763 of multiple certificates matching a given criteria, however
16764 this can be worked round by performing a lookup first
16765 (which will fill the cache with candidate certificates)
16766 and then examining the cache for matches. This is probably
16767 the best we can do without throwing out X509_LOOKUP
16768 entirely (maybe later...).
16770 The X509_VERIFY_CTX structure has been enhanced considerably.
16772 All certificate lookup operations now go via a get_issuer()
16773 callback. Although this currently uses an X509_STORE it
16774 can be replaced by custom lookups. This is a simple way
16775 to bypass the X509_STORE hackery necessary to make this
16776 work and makes it possible to use more efficient techniques
16777 in future. A very simple version which uses a simple
16778 STACK for its trusted certificate store is also provided
16779 using X509_STORE_CTX_trusted_stack().
16781 The verify_cb() and verify() callbacks now have equivalents
16782 in the X509_STORE_CTX structure.
16784 X509_STORE_CTX also has a 'flags' field which can be used
16785 to customise the verify behaviour.
16789 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16790 excludes S/MIME capabilities.
16794 * When a certificate request is read in keep a copy of the
16795 original encoding of the signed data and use it when outputting
16796 again. Signatures then use the original encoding rather than
16797 a decoded, encoded version which may cause problems if the
16798 request is improperly encoded.
16802 * For consistency with other BIO_puts implementations, call
16803 buffer_write(b, ...) directly in buffer_puts instead of calling
16806 In BIO_puts, increment b->num_write as in BIO_write.
16808 *Peter.Sylvester@EdelWeb.fr*
16810 * Fix BN_mul_word for the case where the word is 0. (We have to use
16811 BN_zero, we may not return a BIGNUM with an array consisting of
16812 words set to zero.)
16816 * Avoid calling abort() from within the library when problems are
16817 detected, except if preprocessor symbols have been defined
16818 (such as REF_CHECK, BN_DEBUG etc.).
16822 * New openssl application 'rsautl'. This utility can be
16823 used for low-level RSA operations. DER public key
16824 BIO/fp routines also added.
16828 * New Configure entry and patches for compiling on QNX 4.
16830 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16832 * A demo state-machine implementation was sponsored by
16833 Nuron (<http://www.nuron.com/>) and is now available in
16834 demos/state_machine.
16838 * New options added to the 'dgst' utility for signature
16839 generation and verification.
16843 * Unrecognized PKCS#7 content types are now handled via a
16844 catch all ASN1_TYPE structure. This allows unsupported
16845 types to be stored as a "blob" and an application can
16846 encode and decode it manually.
16850 * Fix various signed/unsigned issues to make a_strex.c
16851 compile under VC++.
16853 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16855 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16856 length if passed a buffer. ASN1_INTEGER_to_BN failed
16857 if passed a NULL BN and its argument was negative.
16859 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16861 * Modification to PKCS#7 encoding routines to output definite
16862 length encoding. Since currently the whole structures are in
16863 memory there's not real point in using indefinite length
16864 constructed encoding. However if OpenSSL is compiled with
16865 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16869 * Added BIO_vprintf() and BIO_vsnprintf().
16873 * Added more prefixes to parse for in the strings written
16874 through a logging bio, to cover all the levels that are available
16875 through syslog. The prefixes are now:
16877 PANIC, EMERG, EMR => LOG_EMERG
16878 ALERT, ALR => LOG_ALERT
16879 CRIT, CRI => LOG_CRIT
16880 ERROR, ERR => LOG_ERR
16881 WARNING, WARN, WAR => LOG_WARNING
16882 NOTICE, NOTE, NOT => LOG_NOTICE
16883 INFO, INF => LOG_INFO
16884 DEBUG, DBG => LOG_DEBUG
16886 and as before, if none of those prefixes are present at the
16887 beginning of the string, LOG_ERR is chosen.
16889 On Win32, the `LOG_*` levels are mapped according to this:
16891 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16892 LOG_WARNING => EVENTLOG_WARNING_TYPE
16893 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16897 * Made it possible to reconfigure with just the configuration
16898 argument "reconf" or "reconfigure". The command line arguments
16899 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16900 and are retrieved from there when reconfiguring.
16906 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16908 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16912 * The obj_dat.pl script was messing up the sorting of object
16913 names. The reason was that it compared the quoted version
16914 of strings as a result "OCSP" > "OCSP Signing" because
16915 " > SPACE. Changed script to store unquoted versions of
16916 names and add quotes on output. It was also omitting some
16917 names from the lookup table if they were given a default
16918 value (that is if SN is missing it is given the same
16919 value as LN and vice versa), these are now added on the
16920 grounds that if an object has a name we should be able to
16921 look it up. Finally added warning output when duplicate
16922 short or long names are found.
16926 * Changes needed for Tandem NSK.
16928 *Scott Uroff <scott@xypro.com>*
16930 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16931 RSA_padding_check_SSLv23(), special padding was never detected
16932 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16933 version rollback attacks was not effective.
16935 In s23_clnt.c, don't use special rollback-attack detection padding
16936 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16937 client; similarly, in s23_srvr.c, don't do the rollback check if
16938 SSL 2.0 is the only protocol enabled in the server.
16942 * Make it possible to get hexdumps of unprintable data with 'openssl
16943 asn1parse'. By implication, the functions ASN1_parse_dump() and
16944 BIO_dump_indent() are added.
16948 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16949 these print out strings and name structures based on various
16950 flags including RFC2253 support and proper handling of
16951 multibyte characters. Added options to the 'x509' utility
16952 to allow the various flags to be set.
16956 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16957 Also change the functions X509_cmp_current_time() and
16958 X509_gmtime_adj() work with an ASN1_TIME structure,
16959 this will enable certificates using GeneralizedTime in validity
16960 dates to be checked.
16964 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16965 negative public key encodings) on by default,
16966 NO_NEG_PUBKEY_BUG can be set to disable it.
16970 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16971 content octets. An i2c_ASN1_OBJECT is unnecessary because
16972 the encoding can be trivially obtained from the structure.
16976 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16977 not read locks (`CRYPTO_r_[un]lock`).
16981 * A first attempt at creating official support for shared
16982 libraries through configuration. I've kept it so the
16983 default is static libraries only, and the OpenSSL programs
16984 are always statically linked for now, but there are
16985 preparations for dynamic linking in place.
16986 This has been tested on Linux and Tru64.
16990 * Randomness polling function for Win9x, as described in:
16991 Peter Gutmann, Software Generation of Practically Strong
16996 * Fix so PRNG is seeded in req if using an already existing
17001 * New options to smime application. -inform and -outform
17002 allow alternative formats for the S/MIME message including
17003 PEM and DER. The -content option allows the content to be
17004 specified separately. This should allow things like Netscape
17005 form signing output easier to verify.
17009 * Fix the ASN1 encoding of tags using the 'long form'.
17013 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
17014 STRING types. These convert content octets to and from the
17015 underlying type. The actual tag and length octets are
17016 already assumed to have been read in and checked. These
17017 are needed because all other string types have virtually
17018 identical handling apart from the tag. By having versions
17019 of the ASN1 functions that just operate on content octets
17020 IMPLICIT tagging can be handled properly. It also allows
17021 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
17022 and ASN1_INTEGER are identical apart from the tag.
17026 * Change the handling of OID objects as follows:
17028 - New object identifiers are inserted in objects.txt, following
17029 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
17030 - objects.pl is used to process obj_mac.num and create a new
17032 - obj_dat.pl is used to create a new obj_dat.h, using the data in
17035 This is currently kind of a hack, and the perl code in objects.pl
17036 isn't very elegant, but it works as I intended. The simplest way
17037 to check that it worked correctly is to look in obj_dat.h and
17038 check the array nid_objs and make sure the objects haven't moved
17039 around (this is important!). Additions are OK, as well as
17040 consistent name changes.
17044 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
17048 * Addition of the command line parameter '-rand file' to 'openssl req'.
17049 The given file adds to whatever has already been seeded into the
17050 random pool through the RANDFILE configuration file option or
17051 environment variable, or the default random state file.
17055 * mkstack.pl now sorts each macro group into lexical order.
17056 Previously the output order depended on the order the files
17057 appeared in the directory, resulting in needless rewriting
17062 * Patches to make OpenSSL compile under Win32 again. Mostly
17063 work arounds for the VC++ problem that it treats func() as
17064 func(void). Also stripped out the parts of mkdef.pl that
17065 added extra typesafe functions: these no longer exist.
17069 * Reorganisation of the stack code. The macros are now all
17070 collected in safestack.h . Each macro is defined in terms of
17071 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
17072 DEBUG_SAFESTACK is now handled in terms of function casts,
17073 this has the advantage of retaining type safety without the
17074 use of additional functions. If DEBUG_SAFESTACK is not defined
17075 then the non typesafe macros are used instead. Also modified the
17076 mkstack.pl script to handle the new form. Needs testing to see
17077 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
17078 the default if no major problems. Similar behaviour for ASN1_SET_OF
17079 and PKCS12_STACK_OF.
17083 * When some versions of IIS use the 'NET' form of private key the
17084 key derivation algorithm is different. Normally MD5(password) is
17085 used as a 128 bit RC4 key. In the modified case
17086 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
17087 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
17088 as the old Netscape_RSA functions except they have an additional
17089 'sgckey' parameter which uses the modified algorithm. Also added
17090 an -sgckey command line option to the rsa utility. Thanks to
17091 Adrian Peck <bertie@ncipher.com> for posting details of the modified
17092 algorithm to openssl-dev.
17096 * The evp_local.h macros were using 'c.##kname' which resulted in
17097 invalid expansion on some systems (SCO 5.0.5 for example).
17098 Corrected to 'c.kname'.
17100 *Phillip Porch <root@theporch.com>*
17102 * New X509_get1_email() and X509_REQ_get1_email() functions that return
17103 a STACK of email addresses from a certificate or request, these look
17104 in the subject name and the subject alternative name extensions and
17105 omit any duplicate addresses.
17109 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
17110 This makes DSA verification about 2 % faster.
17114 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
17115 (meaning that now 2^5 values will be precomputed, which is only 4 KB
17116 plus overhead for 1024 bit moduli).
17117 This makes exponentiations about 0.5 % faster for 1024 bit
17118 exponents (as measured by "openssl speed rsa2048").
17122 * Rename memory handling macros to avoid conflicts with other
17124 Malloc => OPENSSL_malloc
17125 Malloc_locked => OPENSSL_malloc_locked
17126 Realloc => OPENSSL_realloc
17127 Free => OPENSSL_free
17131 * New function BN_mod_exp_mont_word for small bases (roughly 15%
17132 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
17136 * CygWin32 support.
17138 *John Jarvie <jjarvie@newsguy.com>*
17140 * The type-safe stack code has been rejigged. It is now only compiled
17141 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
17142 by default all type-specific stack functions are "#define"d back to
17143 standard stack functions. This results in more streamlined output
17144 but retains the type-safety checking possibilities of the original
17149 * The STACK code has been cleaned up, and certain type declarations
17150 that didn't make a lot of sense have been brought in line. This has
17151 also involved a cleanup of sorts in safestack.h to more correctly
17152 map type-safe stack functions onto their plain stack counterparts.
17153 This work has also resulted in a variety of "const"ifications of
17154 lots of the code, especially `_cmp` operations which should normally
17155 be prototyped with "const" parameters anyway.
17159 * When generating bytes for the first time in md_rand.c, 'stir the pool'
17160 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
17161 (The PRNG state consists of two parts, the large pool 'state' and 'md',
17162 where all of 'md' is used each time the PRNG is used, but 'state'
17163 is used only indexed by a cyclic counter. As entropy may not be
17164 well distributed from the beginning, 'md' is important as a
17165 chaining variable. However, the output function chains only half
17166 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
17167 all of 'md', and seeding with STATE_SIZE dummy bytes will result
17168 in all of 'state' being rewritten, with the new values depending
17169 on virtually all of 'md'. This overcomes the 80 bit limitation.)
17173 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
17174 the handshake is continued after ssl_verify_cert_chain();
17175 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
17176 can lead to 'unexplainable' connection aborts later.
17178 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
17180 * Major EVP API cipher revision.
17181 Add hooks for extra EVP features. This allows various cipher
17182 parameters to be set in the EVP interface. Support added for variable
17183 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
17184 setting of RC2 and RC5 parameters.
17186 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
17189 Remove lots of duplicated code from the EVP library. For example *every*
17190 cipher init() function handles the 'iv' in the same way according to the
17191 cipher mode. They also all do nothing if the 'key' parameter is NULL and
17192 for CFB and OFB modes they zero ctx->num.
17194 New functionality allows removal of S/MIME code RC2 hack.
17196 Most of the routines have the same form and so can be declared in terms
17199 By shifting this to the top level EVP_CipherInit() it can be removed from
17200 all individual ciphers. If the cipher wants to handle IVs or keys
17201 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
17204 Change lots of functions like EVP_EncryptUpdate() to now return a
17205 value: although software versions of the algorithms cannot fail
17206 any installed hardware versions can.
17210 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
17211 this option is set, tolerate broken clients that send the negotiated
17212 protocol version number instead of the requested protocol version
17217 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
17218 i.e. non-zero for export ciphersuites, zero otherwise.
17219 Previous versions had this flag inverted, inconsistent with
17220 rsa_tmp_cb (..._TMP_RSA_CB).
17222 *Bodo Moeller; problem reported by Amit Chopra*
17224 * Add missing DSA library text string. Work around for some IIS
17225 key files with invalid SEQUENCE encoding.
17229 * Add a document (doc/standards.txt) that list all kinds of standards
17230 and so on that are implemented in OpenSSL.
17234 * Enhance c_rehash script. Old version would mishandle certificates
17235 with the same subject name hash and wouldn't handle CRLs at all.
17236 Added -fingerprint option to crl utility, to support new c_rehash
17241 * Eliminate non-ANSI declarations in crypto.h and stack.h.
17245 * Fix for SSL server purpose checking. Server checking was
17246 rejecting certificates which had extended key usage present
17247 but no ssl client purpose.
17249 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
17251 * Make PKCS#12 code work with no password. The PKCS#12 spec
17252 is a little unclear about how a blank password is handled.
17253 Since the password in encoded as a BMPString with terminating
17254 double NULL a zero length password would end up as just the
17255 double NULL. However no password at all is different and is
17256 handled differently in the PKCS#12 key generation code. NS
17257 treats a blank password as zero length. MSIE treats it as no
17258 password on export: but it will try both on import. We now do
17259 the same: PKCS12_parse() tries zero length and no password if
17260 the password is set to "" or NULL (NULL is now a valid password:
17261 it wasn't before) as does the pkcs12 application.
17265 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17266 perror when PEM_read_bio_X509_REQ fails, the error message must
17267 be obtained from the error queue.
17271 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17272 it in ERR_remove_state if appropriate, and change ERR_get_state
17273 accordingly to avoid race conditions (this is necessary because
17274 thread_hash is no longer constant once set).
17278 * Bugfix for linux-elf makefile.one.
17282 * RSA_get_default_method() will now cause a default
17283 RSA_METHOD to be chosen if one doesn't exist already.
17284 Previously this was only set during a call to RSA_new()
17285 or RSA_new_method(NULL) meaning it was possible for
17286 RSA_get_default_method() to return NULL.
17290 * Added native name translation to the existing DSO code
17291 that will convert (if the flag to do so is set) filenames
17292 that are sufficiently small and have no path information
17293 into a canonical native form. Eg. "blah" converted to
17294 "libblah.so" or "blah.dll" etc.
17298 * New function ERR_error_string_n(e, buf, len) which is like
17299 ERR_error_string(e, buf), but writes at most 'len' bytes
17300 including the 0 terminator. For ERR_error_string_n, 'buf'
17303 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17305 * CONF library reworked to become more general. A new CONF
17306 configuration file reader "class" is implemented as well as a
17307 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17308 old `CONF_*` functions are still there, but are reimplemented to
17309 work in terms of the new functions. Also, a set of functions
17310 to handle the internal storage of the configuration data is
17311 provided to make it easier to write new configuration file
17312 reader "classes" (I can definitely see something reading a
17313 configuration file in XML format, for example), called `_CONF_*`,
17314 or "the configuration storage API"...
17316 The new configuration file reading functions are:
17318 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17319 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17321 NCONF_default, NCONF_WIN32
17323 NCONF_dump_fp, NCONF_dump_bio
17325 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17326 NCONF_new creates a new CONF object. This works in the same way
17327 as other interfaces in OpenSSL, like the BIO interface.
17328 `NCONF_dump_*` dump the internal storage of the configuration file,
17329 which is useful for debugging. All other functions take the same
17330 arguments as the old `CONF_*` functions with the exception of the
17331 first that must be a `CONF *` instead of a `LHASH *`.
17333 To make it easier to use the new classes with the old `CONF_*` functions,
17334 the function CONF_set_default_method is provided.
17338 * Add '-tls1' option to 'openssl ciphers', which was already
17339 mentioned in the documentation but had not been implemented.
17340 (This option is not yet really useful because even the additional
17341 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17345 * Initial DSO code added into libcrypto for letting OpenSSL (and
17346 OpenSSL-based applications) load shared libraries and bind to
17347 them in a portable way.
17349 *Geoff Thorpe, with contributions from Richard Levitte*
17351 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17353 * Make sure _lrotl and _lrotr are only used with MSVC.
17355 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17356 (the default implementation of RAND_status).
17358 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17359 to '-clrext' (= clear extensions), as intended and documented.
17360 *Bodo Moeller; inconsistency pointed out by Michael Attili
17361 <attili@amaxo.com>*
17363 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17364 was larger than the MD block size.
17366 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17368 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17369 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17370 using the passed key: if the passed key was a private key the result
17371 of X509_print(), for example, would be to print out all the private key
17376 * des_quad_cksum() byte order bug fix.
17377 *Ulf Möller, using the problem description in krb4-0.9.7, where
17378 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17380 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17383 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17385 * For easily testing in shell scripts whether some command
17386 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17387 returns with exit code 0 iff no command of the given name is available.
17388 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17389 the output goes to stdout and nothing is printed to stderr.
17390 Additional arguments are always ignored.
17392 Since for each cipher there is a command of the same name,
17393 the 'no-cipher' compilation switches can be tested this way.
17395 ('openssl no-XXX' is not able to detect pseudo-commands such
17396 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17400 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17404 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17405 is set; it will be thrown away anyway because each handshake creates
17407 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17408 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17409 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17410 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17414 * New s_client option -ign_eof: EOF at stdin is ignored, and
17415 'Q' and 'R' lose their special meanings (quit/renegotiate).
17416 This is part of what -quiet does; unlike -quiet, -ign_eof
17417 does not suppress any output.
17421 * Add compatibility options to the purpose and trust code. The
17422 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17423 accepts a certificate or CA, this was the previous behaviour,
17424 with all the associated security issues.
17426 X509_TRUST_COMPAT is the old trust behaviour: only and
17427 automatically trust self signed roots in certificate store. A
17428 new trust setting X509_TRUST_DEFAULT is used to specify that
17429 a purpose has no associated trust setting and it should instead
17430 use the value in the default purpose.
17434 * Fix the PKCS#8 DSA private key code so it decodes keys again
17435 and fix a memory leak.
17439 * In util/mkerr.pl (which implements 'make errors'), preserve
17440 reason strings from the previous version of the .c file, as
17441 the default to have only downcase letters (and digits) in
17442 automatically generated reasons codes is not always appropriate.
17446 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17447 using strerror. Previously, ERR_reason_error_string() returned
17448 library names as reason strings for SYSerr; but SYSerr is a special
17449 case where small numbers are errno values, not library numbers.
17453 * Add '-dsaparam' option to 'openssl dhparam' application. This
17454 converts DSA parameters into DH parameters. (When creating parameters,
17455 DSA_generate_parameters is used.)
17459 * Include 'length' (recommended exponent length) in C code generated
17460 by 'openssl dhparam -C'.
17464 * The second argument to set_label in perlasm was already being used
17465 so couldn't be used as a "file scope" flag. Moved to third argument
17470 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17471 instead of RAND_bytes for encryption IVs and salts.
17475 * Include RAND_status() into RAND_METHOD instead of implementing
17476 it only for md_rand.c Otherwise replacing the PRNG by calling
17477 RAND_set_rand_method would be impossible.
17481 * Don't let DSA_generate_key() enter an infinite loop if the random
17482 number generation fails.
17486 * New 'rand' application for creating pseudo-random output.
17490 * Added configuration support for Linux/IA64
17492 *Rolf Haberrecker <rolf@suse.de>*
17494 * Assembler module support for Mingw32.
17498 * Shared library support for HPUX (in shlib/).
17500 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17502 * Shared library support for Solaris gcc.
17504 *Lutz Behnke <behnke@trustcenter.de>*
17506 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17508 * PKCS7_encrypt() was adding text MIME headers twice because they
17509 were added manually and by SMIME_crlf_copy().
17513 * In bntest.c don't call BN_rand with zero bits argument.
17515 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17517 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17518 case was implemented. This caused BN_div_recp() to fail occasionally.
17522 * Add an optional second argument to the set_label() in the perl
17523 assembly language builder. If this argument exists and is set
17524 to 1 it signals that the assembler should use a symbol whose
17525 scope is the entire file, not just the current function. This
17526 is needed with MASM which uses the format label:: for this scope.
17528 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17530 * Change the ASN1 types so they are typedefs by default. Before
17531 almost all types were #define'd to ASN1_STRING which was causing
17532 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17537 * Change names of new functions to the new get1/get0 naming
17538 convention: After 'get1', the caller owns a reference count
17539 and has to call `..._free`; 'get0' returns a pointer to some
17540 data structure without incrementing reference counters.
17541 (Some of the existing 'get' functions increment a reference
17542 counter, some don't.)
17543 Similarly, 'set1' and 'add1' functions increase reference
17544 counters or duplicate objects.
17548 * Allow for the possibility of temp RSA key generation failure:
17549 the code used to assume it always worked and crashed on failure.
17553 * Fix potential buffer overrun problem in BIO_printf().
17554 *Ulf Möller, using public domain code by Patrick Powell; problem
17555 pointed out by David Sacerdote <das33@cornell.edu>*
17557 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17558 RAND_egd() and RAND_status(). In the command line application,
17559 the EGD socket can be specified like a seed file using RANDFILE
17564 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17565 Some CAs (e.g. Verisign) distribute certificates in this form.
17569 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17570 list to exclude them. This means that no special compilation option
17571 is needed to use anonymous DH: it just needs to be included in the
17576 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17577 EVP_MD_type. The old functionality is available in a new macro called
17578 EVP_MD_md(). Change code that uses it and update docs.
17582 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17583 where the `void *` argument is replaced by a function pointer argument.
17584 Previously `void *` was abused to point to functions, which works on
17585 many platforms, but is not correct. As these functions are usually
17586 called by macros defined in OpenSSL header files, most source code
17587 should work without changes.
17591 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17592 sections with information on -D... compiler switches used for
17593 compiling the library so that applications can see them. To enable
17594 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17595 must be defined. E.g.,
17596 #define OPENSSL_ALGORITHM_DEFINES
17597 #include <openssl/opensslconf.h>
17598 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17600 *Richard Levitte, Ulf and Bodo Möller*
17602 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17607 * Change the 'other' type in certificate aux info to a STACK_OF
17608 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17609 the required ASN1 format: arbitrary types determined by an OID.
17613 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17614 argument to 'req'. This is not because the function is newer or
17615 better than others it just uses the work 'NEW' in the certificate
17616 request header lines. Some software needs this.
17620 * Reorganise password command line arguments: now passwords can be
17621 obtained from various sources. Delete the PEM_cb function and make
17622 it the default behaviour: i.e. if the callback is NULL and the
17623 usrdata argument is not NULL interpret it as a null terminated pass
17624 phrase. If usrdata and the callback are NULL then the pass phrase
17625 is prompted for as usual.
17629 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17630 the support is automatically enabled. The resulting binaries will
17631 autodetect the card and use it if present.
17633 *Ben Laurie and Compaq Inc.*
17635 * Work around for Netscape hang bug. This sends certificate request
17636 and server done in one record. Since this is perfectly legal in the
17637 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17638 the bugs/SSLv3 entry for more info.
17642 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17646 * Add -rand argument to smime and pkcs12 applications and read/write
17651 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17655 * Add command line password options to the remaining applications.
17659 * Bug fix for BN_div_recp() for numerators with an even number of
17664 * More tests in bntest.c, and changed test_bn output.
17668 * ./config recognizes MacOS X now.
17672 * Bug fix for BN_div() when the first words of num and divisor are
17673 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17677 * Add support for various broken PKCS#8 formats, and command line
17678 options to produce them.
17682 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17683 get temporary BIGNUMs from a BN_CTX.
17687 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17692 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17693 include a #define from the old name to the new. The original intent
17694 was that statically linked binaries could for example just call
17695 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17696 link with digests. This never worked because SSLeay_add_all_digests()
17697 and SSLeay_add_all_ciphers() were in the same source file so calling
17698 one would link with the other. They are now in separate source files.
17702 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17706 * Use a less unusual form of the Miller-Rabin primality test (it used
17707 a binary algorithm for exponentiation integrated into the Miller-Rabin
17708 loop, our standard modexp algorithms are faster).
17712 * Support for the EBCDIC character set completed.
17714 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17716 * Source code cleanups: use const where appropriate, eliminate casts,
17717 use `void *` instead of `char *` in lhash.
17721 * Bugfix: ssl3_send_server_key_exchange was not restartable
17722 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17723 this the server could overwrite ephemeral keys that the client
17728 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17729 using 50 iterations of the Rabin-Miller test.
17731 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17732 iterations of the Rabin-Miller test as required by the appendix
17733 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17734 As BN_is_prime_fasttest includes trial division, DSA parameter
17735 generation becomes much faster.
17737 This implies a change for the callback functions in DSA_is_prime
17738 and DSA_generate_parameters: The callback function is called once
17739 for each positive witness in the Rabin-Miller test, not just
17740 occasionally in the inner loop; and the parameters to the
17741 callback function now provide an iteration count for the outer
17742 loop rather than for the current invocation of the inner loop.
17743 DSA_generate_parameters additionally can call the callback
17744 function with an 'iteration count' of -1, meaning that a
17745 candidate has passed the trial division test (when q is generated
17746 from an application-provided seed, trial division is skipped).
17750 * New function BN_is_prime_fasttest that optionally does trial
17751 division before starting the Rabin-Miller test and has
17752 an additional BN_CTX * argument (whereas BN_is_prime always
17753 has to allocate at least one BN_CTX).
17754 'callback(1, -1, cb_arg)' is called when a number has passed the
17755 trial division stage.
17759 * Fix for bug in CRL encoding. The validity dates weren't being handled
17764 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17768 * New function BN_pseudo_rand().
17772 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17773 bignum version of BN_from_montgomery() with the working code from
17774 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17779 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17780 made it impossible to use the same SSL_SESSION data structure in
17781 SSL2 clients in multiple threads.
17785 * The return value of RAND_load_file() no longer counts bytes obtained
17786 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17787 to seed the PRNG (previously an explicit byte count was required).
17789 *Ulf Möller, Bodo Möller*
17791 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17792 used `char *` instead of `void *` and had casts all over the place.
17796 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17800 * Retain source code compatibility for BN_prime_checks macro:
17801 BN_is_prime(..., BN_prime_checks, ...) now uses
17802 BN_prime_checks_for_size to determine the appropriate number of
17803 Rabin-Miller iterations.
17807 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17808 DH_CHECK_P_NOT_SAFE_PRIME.
17809 (Check if this is true? OpenPGP calls them "strong".)
17813 * Merge the functionality of "dh" and "gendh" programs into a new program
17814 "dhparam". The old programs are retained for now but will handle DH keys
17815 (instead of parameters) in future.
17819 * Make the ciphers, s_server and s_client programs check the return values
17820 when a new cipher list is set.
17824 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17825 ciphers. Before when the 56bit ciphers were enabled the sorting was
17828 The syntax for the cipher sorting has been extended to support sorting by
17829 cipher-strength (using the strength_bits hard coded in the tables).
17830 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17832 Fix a bug in the cipher-command parser: when supplying a cipher command
17833 string with an "undefined" symbol (neither command nor alphanumeric
17834 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17835 an error is flagged.
17837 Due to the strength-sorting extension, the code of the
17838 ssl_create_cipher_list() function was completely rearranged. I hope that
17839 the readability was also increased :-)
17841 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17843 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17844 for the first serial number and places 2 in the serial number file. This
17845 avoids problems when the root CA is created with serial number zero and
17846 the first user certificate has the same issuer name and serial number
17851 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17852 the new code. Add documentation for this stuff.
17856 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17857 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17858 structures and behave in an analogous way to the X509v3 functions:
17859 they shouldn't be called directly but wrapper functions should be used
17862 So we also now have some wrapper functions that call the X509at functions
17863 when passed certificate requests. (TO DO: similar things can be done with
17864 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17865 things. Some of these need some d2i or i2d and print functionality
17866 because they handle more complex structures.)
17870 * Add missing #ifndefs that caused missing symbols when building libssl
17871 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17872 NO_RSA in `ssl/s2*.c`.
17874 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17876 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17877 has a return value which indicates the quality of the random data
17878 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17879 error queue. New function RAND_pseudo_bytes() generates output that is
17880 guaranteed to be unique but not unpredictable. RAND_add is like
17881 RAND_seed, but takes an extra argument for an entropy estimate
17882 (RAND_seed always assumes full entropy).
17886 * Do more iterations of Rabin-Miller probable prime test (specifically,
17887 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17888 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17889 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17890 false-positive rate of at most 2^-80 for random input.
17894 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17898 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17899 in the 0.9.5 release), this returns the chain
17900 from an X509_CTX structure with a dup of the stack and all
17901 the X509 reference counts upped: so the stack will exist
17902 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17905 Also make SSL_SESSION_print() print out the verify return
17910 * Add manpage for the pkcs12 command. Also change the default
17911 behaviour so MAC iteration counts are used unless the new
17912 -nomaciter option is used. This improves file security and
17913 only older versions of MSIE (4.0 for example) need it.
17917 * Honor the no-xxx Configure options when creating .DEF files.
17921 * Add PKCS#10 attributes to field table: challengePassword,
17922 unstructuredName and unstructuredAddress. These are taken from
17923 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17924 international characters are used.
17926 More changes to X509_ATTRIBUTE code: allow the setting of types
17927 based on strings. Remove the 'loc' parameter when adding
17928 attributes because these will be a SET OF encoding which is sorted
17933 * Initial changes to the 'req' utility to allow request generation
17934 automation. This will allow an application to just generate a template
17935 file containing all the field values and have req construct the
17938 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17939 used all over the place including certificate requests and PKCS#7
17940 structures. They are currently handled manually where necessary with
17941 some primitive wrappers for PKCS#7. The new functions behave in a
17942 manner analogous to the X509 extension functions: they allow
17943 attributes to be looked up by NID and added.
17945 Later something similar to the X509V3 code would be desirable to
17946 automatically handle the encoding, decoding and printing of the
17947 more complex types. The string types like challengePassword can
17948 be handled by the string table functions.
17950 Also modified the multi byte string table handling. Now there is
17951 a 'global mask' which masks out certain types. The table itself
17952 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17953 is useful when for example there is only one permissible type
17954 (as in countryName) and using the mask might result in no valid
17959 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17960 SSL_get_peer_finished to allow applications to obtain the latest
17961 Finished messages sent to the peer or expected from the peer,
17962 respectively. (SSL_get_peer_finished is usually the Finished message
17963 actually received from the peer, otherwise the protocol will be aborted.)
17965 As the Finished message are message digests of the complete handshake
17966 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17967 be used for external authentication procedures when the authentication
17968 provided by SSL/TLS is not desired or is not enough.
17972 * Enhanced support for Alpha Linux is added. Now ./config checks if
17973 the host supports BWX extension and if Compaq C is present on the
17974 $PATH. Just exploiting of the BWX extension results in 20-30%
17975 performance kick for some algorithms, e.g. DES and RC4 to mention
17976 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17981 * Add support for MS "fast SGC". This is arguably a violation of the
17982 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17983 weak crypto and after checking the certificate is SGC a second one
17984 with strong crypto. MS SGC stops the first handshake after receiving
17985 the server certificate message and sends a second client hello. Since
17986 a server will typically do all the time consuming operations before
17987 expecting any further messages from the client (server key exchange
17988 is the most expensive) there is little difference between the two.
17990 To get OpenSSL to support MS SGC we have to permit a second client
17991 hello message after we have sent server done. In addition we have to
17992 reset the MAC if we do get this second client hello.
17996 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17997 if a DER encoded private key is RSA or DSA traditional format. Changed
17998 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17999 format DER encoded private key. Newer code should use PKCS#8 format which
18000 has the key type encoded in the ASN1 structure. Added DER private key
18001 support to pkcs8 application.
18005 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
18006 ciphersuites has been selected (as required by the SSL 3/TLS 1
18007 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
18008 is set, we interpret this as a request to violate the specification
18009 (the worst that can happen is a handshake failure, and 'correct'
18010 behaviour would result in a handshake failure anyway).
18014 * In SSL_CTX_add_session, take into account that there might be multiple
18015 SSL_SESSION structures with the same session ID (e.g. when two threads
18016 concurrently obtain them from an external cache).
18017 The internal cache can handle only one SSL_SESSION with a given ID,
18018 so if there's a conflict, we now throw out the old one to achieve
18023 * Add OIDs for idea and blowfish in CBC mode. This will allow both
18024 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
18025 some routines that use cipher OIDs: some ciphers do not have OIDs
18026 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
18031 * Simplify the trust setting structure and code. Now we just have
18032 two sequences of OIDs for trusted and rejected settings. These will
18033 typically have values the same as the extended key usage extension
18034 and any application specific purposes.
18036 The trust checking code now has a default behaviour: it will just
18037 check for an object with the same NID as the passed id. Functions can
18038 be provided to override either the default behaviour or the behaviour
18039 for a given id. SSL client, server and email already have functions
18040 in place for compatibility: they check the NID and also return "trusted"
18041 if the certificate is self signed.
18045 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
18046 traditional format into an EVP_PKEY structure.
18050 * Add a password callback function PEM_cb() which either prompts for
18051 a password if usr_data is NULL or otherwise assumes it is a null
18052 terminated password. Allow passwords to be passed on command line
18053 environment or config files in a few more utilities.
18057 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
18058 keys. Add some short names for PKCS#8 PBE algorithms and allow them
18059 to be specified on the command line for the pkcs8 and pkcs12 utilities.
18060 Update documentation.
18064 * Support for ASN1 "NULL" type. This could be handled before by using
18065 ASN1_TYPE but there wasn't any function that would try to read a NULL
18066 and produce an error if it couldn't. For compatibility we also have
18067 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
18068 don't allocate anything because they don't need to.
18072 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
18075 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
18077 * Rebuild of the memory allocation routines used by OpenSSL code and
18078 possibly others as well. The purpose is to make an interface that
18079 provide hooks so anyone can build a separate set of allocation and
18080 deallocation routines to be used by OpenSSL, for example memory
18081 pool implementations, or something else, which was previously hard
18082 since Malloc(), Realloc() and Free() were defined as macros having
18083 the values malloc, realloc and free, respectively (except for Win32
18084 compilations). The same is provided for memory debugging code.
18085 OpenSSL already comes with functionality to find memory leaks, but
18086 this gives people a chance to debug other memory problems.
18088 With these changes, a new set of functions and macros have appeared:
18090 CRYPTO_set_mem_debug_functions() [F]
18091 CRYPTO_get_mem_debug_functions() [F]
18092 CRYPTO_dbg_set_options() [F]
18093 CRYPTO_dbg_get_options() [F]
18094 CRYPTO_malloc_debug_init() [M]
18096 The memory debug functions are NULL by default, unless the library
18097 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
18098 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
18099 gives the standard debugging functions that come with OpenSSL) or
18100 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
18101 provided by the library user) must be used. When the standard
18102 debugging functions are used, CRYPTO_dbg_set_options can be used to
18103 request additional information:
18104 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
18105 the CRYPTO_MDEBUG_xxx macro when compiling the library.
18107 Also, things like CRYPTO_set_mem_functions will always give the
18108 expected result (the new set of functions is used for allocation
18109 and deallocation) at all times, regardless of platform and compiler
18112 To finish it up, some functions that were never use in any other
18113 way than through macros have a new API and new semantic:
18115 CRYPTO_dbg_malloc()
18116 CRYPTO_dbg_realloc()
18119 All macros of value have retained their old syntax.
18121 *Richard Levitte and Bodo Moeller*
18123 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
18124 ordering of SMIMECapabilities wasn't in "strength order" and there
18125 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
18130 * Some ASN1 types with illegal zero length encoding (INTEGER,
18131 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
18133 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
18135 * Merge in my S/MIME library for OpenSSL. This provides a simple
18136 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
18137 functionality to handle multipart/signed properly) and a utility
18138 called 'smime' to call all this stuff. This is based on code I
18139 originally wrote for Celo who have kindly allowed it to be
18140 included in OpenSSL.
18144 * Add variants des_set_key_checked and des_set_key_unchecked of
18145 des_set_key (aka des_key_sched). Global variable des_check_key
18146 decides which of these is called by des_set_key; this way
18147 des_check_key behaves as it always did, but applications and
18148 the library itself, which was buggy for des_check_key == 1,
18149 have a cleaner way to pick the version they need.
18153 * New function PKCS12_newpass() which changes the password of a
18158 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
18159 dynamic mix. In both cases the ids can be used as an index into the
18160 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
18161 functions so they accept a list of the field values and the
18162 application doesn't need to directly manipulate the X509_TRUST
18167 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
18172 * Modify the way the V3 extension code looks up extensions. This now
18173 works in a similar way to the object code: we have some "standard"
18174 extensions in a static table which is searched with OBJ_bsearch()
18175 and the application can add dynamic ones if needed. The file
18176 crypto/x509v3/ext_dat.h now has the info: this file needs to be
18177 updated whenever a new extension is added to the core code and kept
18178 in ext_nid order. There is a simple program 'tabtest.c' which checks
18179 this. New extensions are not added too often so this file can readily
18180 be maintained manually.
18182 There are two big advantages in doing things this way. The extensions
18183 can be looked up immediately and no longer need to be "added" using
18184 X509V3_add_standard_extensions(): this function now does nothing.
18185 Side note: I get *lots* of email saying the extension code doesn't
18186 work because people forget to call this function.
18187 Also no dynamic allocation is done unless new extensions are added:
18188 so if we don't add custom extensions there is no need to call
18189 X509V3_EXT_cleanup().
18193 * Modify enc utility's salting as follows: make salting the default. Add a
18194 magic header, so unsalted files fail gracefully instead of just decrypting
18195 to garbage. This is because not salting is a big security hole, so people
18196 should be discouraged from doing it.
18200 * Fixes and enhancements to the 'x509' utility. It allowed a message
18201 digest to be passed on the command line but it only used this
18202 parameter when signing a certificate. Modified so all relevant
18203 operations are affected by the digest parameter including the
18204 -fingerprint and -x509toreq options. Also -x509toreq choked if a
18205 DSA key was used because it didn't fix the digest.
18209 * Initial certificate chain verify code. Currently tests the untrusted
18210 certificates for consistency with the verify purpose (which is set
18211 when the X509_STORE_CTX structure is set up) and checks the pathlength.
18213 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
18214 this is because it will reject chains with invalid extensions whereas
18215 every previous version of OpenSSL and SSLeay made no checks at all.
18217 Trust code: checks the root CA for the relevant trust settings. Trust
18218 settings have an initial value consistent with the verify purpose: e.g.
18219 if the verify purpose is for SSL client use it expects the CA to be
18220 trusted for SSL client use. However the default value can be changed to
18221 permit custom trust settings: one example of this would be to only trust
18222 certificates from a specific "secure" set of CAs.
18224 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
18225 which should be used for version portability: especially since the
18226 verify structure is likely to change more often now.
18228 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
18229 to set them. If not set then assume SSL clients will verify SSL servers
18232 Two new options to the verify program: -untrusted allows a set of
18233 untrusted certificates to be passed in and -purpose which sets the
18234 intended purpose of the certificate. If a purpose is set then the
18235 new chain verify code is used to check extension consistency.
18239 * Support for the authority information access extension.
18243 * Modify RSA and DSA PEM read routines to transparently handle
18244 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
18245 public keys in a format compatible with certificate
18246 SubjectPublicKeyInfo structures. Unfortunately there were already
18247 functions called *_PublicKey_* which used various odd formats so
18248 these are retained for compatibility: however the DSA variants were
18249 never in a public release so they have been deleted. Changed dsa/rsa
18250 utilities to handle the new format: note no releases ever handled public
18251 keys so we should be OK.
18253 The primary motivation for this change is to avoid the same fiasco
18254 that dogs private keys: there are several incompatible private key
18255 formats some of which are standard and some OpenSSL specific and
18256 require various evil hacks to allow partial transparent handling and
18257 even then it doesn't work with DER formats. Given the option anything
18258 other than PKCS#8 should be dumped: but the other formats have to
18259 stay in the name of compatibility.
18261 With public keys and the benefit of hindsight one standard format
18262 is used which works with EVP_PKEY, RSA or DSA structures: though
18263 it clearly returns an error if you try to read the wrong kind of key.
18265 Added a -pubkey option to the 'x509' utility to output the public key.
18266 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18267 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18268 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18269 that do the same as the `EVP_PKEY_assign_*()` except they up the
18270 reference count of the added key (they don't "swallow" the
18275 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18276 CRLs would fail if the file contained no certificates or no CRLs:
18277 added a new function to read in both types and return the number
18278 read: this means that if none are read it will be an error. The
18279 DER versions of the certificate and CRL reader would always fail
18280 because it isn't possible to mix certificates and CRLs in DER format
18281 without choking one or the other routine. Changed this to just read
18282 a certificate: this is the best we can do. Also modified the code
18283 in `apps/verify.c` to take notice of return codes: it was previously
18284 attempting to read in certificates from NULL pointers and ignoring
18285 any errors: this is one reason why the cert and CRL reader seemed
18286 to work. It doesn't check return codes from the default certificate
18287 routines: these may well fail if the certificates aren't installed.
18291 * Code to support otherName option in GeneralName.
18295 * First update to verify code. Change the verify utility
18296 so it warns if it is passed a self signed certificate:
18297 for consistency with the normal behaviour. X509_verify
18298 has been modified to it will now verify a self signed
18299 certificate if *exactly* the same certificate appears
18300 in the store: it was previously impossible to trust a
18301 single self signed certificate. This means that:
18302 openssl verify ss.pem
18303 now gives a warning about a self signed certificate but
18304 openssl verify -CAfile ss.pem ss.pem
18309 * For servers, store verify_result in SSL_SESSION data structure
18310 (and add it to external session representation).
18311 This is needed when client certificate verifications fails,
18312 but an application-provided verification callback (set by
18313 SSL_CTX_set_cert_verify_callback) allows accepting the session
18314 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18315 but returns 1): When the session is reused, we have to set
18316 ssl->verify_result to the appropriate error code to avoid
18319 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18321 * Fix a bug in the new PKCS#7 code: it didn't consider the
18322 case in PKCS7_dataInit() where the signed PKCS7 structure
18323 didn't contain any existing data because it was being created.
18325 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18327 * Add a salt to the key derivation routines in enc.c. This
18328 forms the first 8 bytes of the encrypted file. Also add a
18329 -S option to allow a salt to be input on the command line.
18333 * New function X509_cmp(). Oddly enough there wasn't a function
18334 to compare two certificates. We do this by working out the SHA1
18335 hash and comparing that. X509_cmp() will be needed by the trust
18340 * SSL_get1_session() is like SSL_get_session(), but increments
18341 the reference count in the SSL_SESSION returned.
18343 *Geoff Thorpe <geoff@eu.c2.net>*
18345 * Fix for 'req': it was adding a null to request attributes.
18346 Also change the X509_LOOKUP and X509_INFO code to handle
18347 certificate auxiliary information.
18351 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18356 * Add the possibility to add extra information to the memory leak
18357 detecting output, to form tracebacks, showing from where each
18358 allocation was originated: CRYPTO_push_info("constant string") adds
18359 the string plus current file name and line number to a per-thread
18360 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18361 is like calling CYRPTO_pop_info() until the stack is empty.
18362 Also updated memory leak detection code to be multi-thread-safe.
18366 * Add options -text and -noout to pkcs7 utility and delete the
18367 encryption options which never did anything. Update docs.
18371 * Add options to some of the utilities to allow the pass phrase
18372 to be included on either the command line (not recommended on
18373 OSes like Unix) or read from the environment. Update the
18374 manpages and fix a few bugs.
18378 * Add a few manpages for some of the openssl commands.
18382 * Fix the -revoke option in ca. It was freeing up memory twice,
18383 leaking and not finding already revoked certificates.
18387 * Extensive changes to support certificate auxiliary information.
18388 This involves the use of X509_CERT_AUX structure and X509_AUX
18389 functions. An X509_AUX function such as PEM_read_X509_AUX()
18390 can still read in a certificate file in the usual way but it
18391 will also read in any additional "auxiliary information". By
18392 doing things this way a fair degree of compatibility can be
18393 retained: existing certificates can have this information added
18394 using the new 'x509' options.
18396 Current auxiliary information includes an "alias" and some trust
18397 settings. The trust settings will ultimately be used in enhanced
18398 certificate chain verification routines: currently a certificate
18399 can only be trusted if it is self signed and then it is trusted
18404 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18405 The problem was that one of the replacement routines had not been working
18406 since SSLeay releases. For now the offending routine has been replaced
18407 with non-optimised assembler. Even so, this now gives around 95%
18408 performance improvement for 1024 bit RSA signs.
18412 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18413 handling. Most clients have the effective key size in bits equal to
18414 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18415 A few however don't do this and instead use the size of the decrypted key
18416 to determine the RC2 key length and the AlgorithmIdentifier to determine
18417 the effective key length. In this case the effective key length can still
18418 be 40 bits but the key length can be 168 bits for example. This is fixed
18419 by manually forcing an RC2 key into the EVP_PKEY structure because the
18420 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18421 the key length and effective key length are equal.
18425 * Add a bunch of functions that should simplify the creation of
18426 X509_NAME structures. Now you should be able to do:
18427 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18428 and have it automatically work out the correct field type and fill in
18429 the structures. The more adventurous can try:
18430 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18431 and it will (hopefully) work out the correct multibyte encoding.
18435 * Change the 'req' utility to use the new field handling and multibyte
18436 copy routines. Before the DN field creation was handled in an ad hoc
18437 way in req, ca, and x509 which was rather broken and didn't support
18438 BMPStrings or UTF8Strings. Since some software doesn't implement
18439 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18440 using the dirstring_type option. See the new comment in the default
18441 openssl.cnf for more info.
18445 * Make crypto/rand/md_rand.c more robust:
18446 - Assure unique random numbers after fork().
18447 - Make sure that concurrent threads access the global counter and
18448 md serializably so that we never lose entropy in them
18449 or use exactly the same state in multiple threads.
18450 Access to the large state is not always serializable because
18451 the additional locking could be a performance killer, and
18452 md should be large enough anyway.
18456 * New file `apps/app_rand.c` with commonly needed functionality
18457 for handling the random seed file.
18459 Use the random seed file in some applications that previously did not:
18461 dsaparam -genkey (which also ignored its '-rand' option),
18464 x509 (when signing).
18465 Except on systems with /dev/urandom, it is crucial to have a random
18466 seed file at least for key creation, DSA signing, and for DH exchanges;
18467 for RSA signatures we could do without one.
18469 gendh and gendsa (unlike genrsa) used to read only the first byte
18470 of each file listed in the '-rand' option. The function as previously
18471 found in genrsa is now in app_rand.c and is used by all programs
18472 that support '-rand'.
18476 * In RAND_write_file, use mode 0600 for creating files;
18477 don't just chmod when it may be too late.
18481 * Report an error from X509_STORE_load_locations
18482 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18486 * New function ASN1_mbstring_copy() this copies a string in either
18487 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18488 into an ASN1_STRING type. A mask of permissible types is passed
18489 and it chooses the "minimal" type to use or an error if not type
18494 * Add function equivalents to the various macros in asn1.h. The old
18495 macros are retained with an `M_` prefix. Code inside the library can
18496 use the `M_` macros. External code (including the openssl utility)
18497 should *NOT* in order to be "shared library friendly".
18501 * Add various functions that can check a certificate's extensions
18502 to see if it usable for various purposes such as SSL client,
18503 server or S/MIME and CAs of these types. This is currently
18504 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18505 verification. Also added a -purpose flag to x509 utility to
18506 print out all the purposes.
18510 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18515 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18516 for, obtain and decode and extension and obtain its critical flag.
18517 This allows all the necessary extension code to be handled in a
18518 single function call.
18522 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18523 platforms. See crypto/rc4/rc4_enc.c for further details.
18527 * New -noout option to asn1parse. This causes no output to be produced
18528 its main use is when combined with -strparse and -out to extract data
18529 from a file (which may not be in ASN.1 format).
18533 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18534 when producing the local key id.
18536 *Richard Levitte <levitte@stacken.kth.se>*
18538 * New option -dhparam in s_server. This allows a DH parameter file to be
18539 stated explicitly. If it is not stated then it tries the first server
18540 certificate file. The previous behaviour hard coded the filename
18545 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18546 a public key to be input or output. For example:
18547 openssl rsa -in key.pem -pubout -out pubkey.pem
18548 Also added necessary DSA public key functions to handle this.
18552 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18553 in the message. This was handled by allowing
18554 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18556 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18558 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18559 to the end of the strings whereas this didn't. This would cause problems
18560 if strings read with d2i_ASN1_bytes() were later modified.
18562 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18564 * Fix for base64 decode bug. When a base64 bio reads only one line of
18565 data and it contains EOF it will end up returning an error. This is
18566 caused by input 46 bytes long. The cause is due to the way base64
18567 BIOs find the start of base64 encoded data. They do this by trying a
18568 trial decode on each line until they find one that works. When they
18569 do a flag is set and it starts again knowing it can pass all the
18570 data directly through the decoder. Unfortunately it doesn't reset
18571 the context it uses. This means that if EOF is reached an attempt
18572 is made to pass two EOFs through the context and this causes the
18573 resulting error. This can also cause other problems as well. As is
18574 usual with these problems it takes *ages* to find and the fix is
18575 trivial: move one line.
18577 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18579 * Ugly workaround to get s_client and s_server working under Windows. The
18580 old code wouldn't work because it needed to select() on sockets and the
18581 tty (for keypresses and to see if data could be written). Win32 only
18582 supports select() on sockets so we select() with a 1s timeout on the
18583 sockets and then see if any characters are waiting to be read, if none
18584 are present then we retry, we also assume we can always write data to
18585 the tty. This isn't nice because the code then blocks until we've
18586 received a complete line of data and it is effectively polling the
18587 keyboard at 1s intervals: however it's quite a bit better than not
18588 working at all :-) A dedicated Windows application might handle this
18589 with an event loop for example.
18593 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18594 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18595 will be called when RSA_sign() and RSA_verify() are used. This is useful
18596 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18597 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18598 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18599 This necessitated the support of an extra signature type NID_md5_sha1
18600 for SSL signatures and modifications to the SSL library to use it instead
18601 of calling RSA_public_decrypt() and RSA_private_encrypt().
18605 * Add new -verify -CAfile and -CApath options to the crl program, these
18606 will lookup a CRL issuers certificate and verify the signature in a
18607 similar way to the verify program. Tidy up the crl program so it
18608 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18609 less strict. It will now permit CRL extensions even if it is not
18610 a V2 CRL: this will allow it to tolerate some broken CRLs.
18614 * Initialize all non-automatic variables each time one of the openssl
18615 sub-programs is started (this is necessary as they may be started
18616 multiple times from the "OpenSSL>" prompt).
18618 *Lennart Bang, Bodo Moeller*
18620 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18621 removing all other RSA functionality (this is what NO_RSA does). This
18622 is so (for example) those in the US can disable those operations covered
18623 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18628 * Non-copying interface to BIO pairs.
18629 (still largely untested)
18633 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18634 ASCII string. This was handled independently in various places before.
18638 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18639 UTF8 strings a character at a time.
18643 * Use client_version from client hello to select the protocol
18644 (s23_srvr.c) and for RSA client key exchange verification
18645 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18649 * Add various utility functions to handle SPKACs, these were previously
18650 handled by poking round in the structure internals. Added new function
18651 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18652 print, verify and generate SPKACs. Based on an original idea from
18653 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18657 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18661 * Allow the config file extension section to be overwritten on the
18662 command line. Based on an original idea from Massimiliano Pala
18663 <madwolf@comune.modena.it>. The new option is called -extensions
18664 and can be applied to ca, req and x509. Also -reqexts to override
18665 the request extensions in req and -crlexts to override the crl extensions
18670 * Add new feature to the SPKAC handling in ca. Now you can include
18671 the same field multiple times by preceding it by "XXXX." for example:
18674 this is the same syntax as used in the req config file.
18678 * Allow certificate extensions to be added to certificate requests. These
18679 are specified in a 'req_extensions' option of the req section of the
18680 config file. They can be printed out with the -text option to req but
18681 are otherwise ignored at present.
18685 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18686 data read consists of only the final block it would not decrypted because
18687 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18688 A misplaced 'break' also meant the decrypted final block might not be
18689 copied until the next read.
18693 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18694 a few extra parameters to the DH structure: these will be useful if
18695 for example we want the value of 'q' or implement X9.42 DH.
18699 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18700 provides hooks that allow the default DSA functions or functions on a
18701 "per key" basis to be replaced. This allows hardware acceleration and
18702 hardware key storage to be handled without major modification to the
18703 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18704 associated functions.
18708 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18709 as "read only": it can't be written to and the buffer it points to will
18710 not be freed. Reading from a read only BIO is much more efficient than
18711 a normal memory BIO. This was added because there are several times when
18712 an area of memory needs to be read from a BIO. The previous method was
18713 to create a memory BIO and write the data to it, this results in two
18714 copies of the data and an O(n^2) reading algorithm. There is a new
18715 function BIO_new_mem_buf() which creates a read only memory BIO from
18716 an area of memory. Also modified the PKCS#7 routines to use read only
18721 * Bugfix: ssl23_get_client_hello did not work properly when called in
18722 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18723 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18724 but a retry condition occurred while trying to read the rest.
18728 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18729 NID_pkcs7_encrypted by default: this was wrong since this should almost
18730 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18731 the encrypted data type: this is a more sensible place to put it and it
18732 allows the PKCS#12 code to be tidied up that duplicated this
18737 * Changed obj_dat.pl script so it takes its input and output files on
18738 the command line. This should avoid shell escape redirection problems
18743 * Initial support for certificate extension requests, these are included
18744 in things like Xenroll certificate requests. Included functions to allow
18745 extensions to be obtained and added.
18749 * -crlf option to s_client and s_server for sending newlines as
18750 CRLF (as required by many protocols).
18754 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18756 * Install libRSAglue.a when OpenSSL is built with RSAref.
18758 *Ralf S. Engelschall*
18760 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18762 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18764 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18769 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18770 DH parameters/keys (q is lost during that conversion, but the resulting
18771 DH parameters contain its length).
18773 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18774 much faster than DH_generate_parameters (which creates parameters
18775 where `p = 2*q + 1`), and also the smaller q makes DH computations
18776 much more efficient (160-bit exponentiation instead of 1024-bit
18777 exponentiation); so this provides a convenient way to support DHE
18778 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18779 utter importance to use
18780 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18782 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18783 when such DH parameters are used, because otherwise small subgroup
18784 attacks may become possible!
18788 * Avoid memory leak in i2d_DHparams.
18792 * Allow the -k option to be used more than once in the enc program:
18793 this allows the same encrypted message to be read by multiple recipients.
18797 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18798 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18799 it will always use the numerical form of the OID, even if it has a short
18804 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18805 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18806 otherwise bn_mod_exp was called. In the case of hardware keys for example
18807 no private key components need be present and it might store extra data
18808 in the RSA structure, which cannot be accessed from bn_mod_exp.
18809 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18810 private key operations.
18814 * Added support for SPARC Linux.
18818 * pem_password_cb function type incompatibly changed from
18819 typedef int pem_password_cb(char *buf, int size, int rwflag);
18821 ....(char *buf, int size, int rwflag, void *userdata);
18822 so that applications can pass data to their callbacks:
18823 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18824 additional void * argument, which is just handed through whenever
18825 the password callback is called.
18827 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18829 New function SSL_CTX_set_default_passwd_cb_userdata.
18831 Compatibility note: As many C implementations push function arguments
18832 onto the stack in reverse order, the new library version is likely to
18833 interoperate with programs that have been compiled with the old
18834 pem_password_cb definition (PEM_whatever takes some data that
18835 happens to be on the stack as its last argument, and the callback
18836 just ignores this garbage); but there is no guarantee whatsoever that
18839 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18840 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18841 problems not only on Windows, but also on some Unix platforms.
18842 To avoid problematic command lines, these definitions are now in an
18843 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18844 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18848 * MIPS III/IV assembler module is reimplemented.
18852 * More DES library cleanups: remove references to srand/rand and
18853 delete an unused file.
18857 * Add support for the free Netwide assembler (NASM) under Win32,
18858 since not many people have MASM (ml) and it can be hard to obtain.
18859 This is currently experimental but it seems to work OK and pass all
18860 the tests. Check out INSTALL.W32 for info.
18864 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18865 without temporary keys kept an extra copy of the server key,
18866 and connections with temporary keys did not free everything in case
18871 * New function RSA_check_key and new openssl rsa option -check
18872 for verifying the consistency of RSA keys.
18874 *Ulf Moeller, Bodo Moeller*
18876 * Various changes to make Win32 compile work:
18877 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18878 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18879 comparison" warnings.
18880 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18884 * Add a debugging option to PKCS#5 v2 key generation function: when
18885 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18886 derived keys are printed to stderr.
18890 * Copy the flags in ASN1_STRING_dup().
18892 *Roman E. Pavlov <pre@mo.msk.ru>*
18894 * The x509 application mishandled signing requests containing DSA
18895 keys when the signing key was also DSA and the parameters didn't match.
18897 It was supposed to omit the parameters when they matched the signing key:
18898 the verifying software was then supposed to automatically use the CA's
18899 parameters if they were absent from the end user certificate.
18901 Omitting parameters is no longer recommended. The test was also
18902 the wrong way round! This was probably due to unusual behaviour in
18903 EVP_cmp_parameters() which returns 1 if the parameters match.
18904 This meant that parameters were omitted when they *didn't* match and
18905 the certificate was useless. Certificates signed with 'ca' didn't have
18908 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18910 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18911 The interface is as follows:
18912 Applications can use
18913 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18914 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18915 "off" is now the default.
18916 The library internally uses
18917 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18918 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18919 to disable memory-checking temporarily.
18921 Some inconsistent states that previously were possible (and were
18922 even the default) are now avoided.
18924 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18925 with each memory chunk allocated; this is occasionally more helpful
18926 than just having a counter.
18928 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18930 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18935 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18936 which largely parallels "options", but is for changing API behaviour,
18937 whereas "options" are about protocol behaviour.
18938 Initial "mode" flags are:
18940 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18941 a single record has been written.
18942 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18943 retries use the same buffer location.
18944 (But all of the contents must be
18949 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18952 * Fix problems with no-hmac etc.
18954 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18956 * New functions RSA_get_default_method(), RSA_set_method() and
18957 RSA_get_method(). These allows replacement of RSA_METHODs without having
18958 to mess around with the internals of an RSA structure.
18962 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18963 Also really enable memory leak checks in openssl.c and in some
18966 *Chad C. Mulligan, Bodo Moeller*
18968 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18969 up the length of negative integers. This has now been simplified to just
18970 store the length when it is first determined and use it later, rather
18971 than trying to keep track of where data is copied and updating it to
18973 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18975 * Add a new function PKCS7_signatureVerify. This allows the verification
18976 of a PKCS#7 signature but with the signing certificate passed to the
18977 function itself. This contrasts with PKCS7_dataVerify which assumes the
18978 certificate is present in the PKCS#7 structure. This isn't always the
18979 case: certificates can be omitted from a PKCS#7 structure and be
18980 distributed by "out of band" means (such as a certificate database).
18984 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18985 function prototypes in pem.h, also change util/mkdef.pl to add the
18986 necessary function names.
18990 * mk1mf.pl (used by Windows builds) did not properly read the
18991 options set by Configure in the top level Makefile, and Configure
18992 was not even able to write more than one option correctly.
18993 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18997 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18998 file to be loaded from a BIO or FILE pointer. The BIO version will
18999 for example allow memory BIOs to contain config info.
19003 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
19004 Whoever hopes to achieve shared-library compatibility across versions
19005 must use this, not the compile-time macro.
19006 (Exercise 0.9.4: Which is the minimum library version required by
19008 Note: All this applies only to multi-threaded programs, others don't
19013 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
19014 through a BIO pair triggered the default case, i.e.
19015 SSLerr(...,SSL_R_UNKNOWN_STATE).
19019 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
19020 can use the SSL library even if none of the specific BIOs is
19025 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
19026 for the encoded length.
19028 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
19030 * Add initial documentation of the X509V3 functions.
19034 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
19035 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
19036 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
19037 secure PKCS#8 private key format with a high iteration count.
19041 * Fix determination of Perl interpreter: A perl or perl5
19042 *directory* in $PATH was also accepted as the interpreter.
19044 *Ralf S. Engelschall*
19046 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
19047 wrong with it but it was very old and did things like calling
19048 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
19049 unusual formatting.
19053 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
19054 to use the new extension code.
19058 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
19059 with macros. This should make it easier to change their form, add extra
19060 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
19065 * Add to configuration table a new entry that can specify an alternative
19066 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
19067 according to Mark Crispin <MRC@Panda.COM>.
19071 * DES CBC did not update the IV. Weird.
19075 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
19076 Changing the behaviour of the former might break existing programs --
19077 where IV updating is needed, des_ncbc_encrypt can be used.
19080 * When bntest is run from "make test" it drives bc to check its
19081 calculations, as well as internally checking them. If an internal check
19082 fails, it needs to cause bc to give a non-zero result or make test carries
19083 on without noticing the failure. Fixed.
19087 * DES library cleanups.
19091 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
19092 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
19093 ciphers. NOTE: although the key derivation function has been verified
19094 against some published test vectors it has not been extensively tested
19095 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
19100 * Instead of "mkdir -p", which is not fully portable, use new
19101 Perl script "util/mkdir-p.pl".
19105 * Rewrite the way password based encryption (PBE) is handled. It used to
19106 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
19107 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
19108 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
19109 the 'parameter' field of the AlgorithmIdentifier is passed to the
19110 underlying key generation function so it must do its own ASN1 parsing.
19111 This has also changed the EVP_PBE_CipherInit() function which now has a
19112 'parameter' argument instead of literal salt and iteration count values
19113 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
19117 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
19118 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
19119 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
19120 KEY" because this clashed with PKCS#8 unencrypted string. Since this
19121 value was just used as a "magic string" and not used directly its
19122 value doesn't matter.
19126 * Introduce some semblance of const correctness to BN. Shame C doesn't
19131 * "linux-sparc64" configuration (ultrapenguin).
19133 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
19134 "linux-sparc" configuration.
19136 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
19138 * config now generates no-xxx options for missing ciphers.
19142 * Support the EBCDIC character set (work in progress).
19143 File ebcdic.c not yet included because it has a different license.
19145 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19147 * Support BS2000/OSD-POSIX.
19149 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19151 * Make callbacks for key generation use `void *` instead of `char *`.
19155 * Make S/MIME samples compile (not yet tested).
19159 * Additional typesafe stacks.
19163 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
19167 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
19169 * New configuration variant "sco5-gcc".
19171 * Updated some demos.
19173 *Sean O Riordain, Wade Scholine*
19175 * Add missing BIO_free at exit of pkcs12 application.
19179 * Fix memory leak in conf.c.
19183 * Updates for Win32 to assembler version of MD5.
19187 * Set #! path to perl in `apps/der_chop` to where we found it
19188 instead of using a fixed path.
19192 * SHA library changes for irix64-mips4-cc.
19196 * Improvements for VMS support.
19200 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
19202 * Bignum library bug fix. IRIX 6 passes "make test" now!
19203 This also avoids the problems with SC4.2 and unpatched SC5.
19205 *Andy Polyakov <appro@fy.chalmers.se>*
19207 * New functions sk_num, sk_value and sk_set to replace the previous macros.
19208 These are required because of the typesafe stack would otherwise break
19209 existing code. If old code used a structure member which used to be STACK
19210 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
19211 sk_num or sk_value it would produce an error because the num, data members
19212 are not present in STACK_OF. Now it just produces a warning. sk_set
19213 replaces the old method of assigning a value to sk_value
19214 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
19215 that does this will no longer work (and should use sk_set instead) but
19216 this could be regarded as a "questionable" behaviour anyway.
19220 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
19221 correctly handle encrypted S/MIME data.
19225 * Change type of various DES function arguments from des_cblock
19226 (which means, in function argument declarations, pointer to char)
19227 to des_cblock * (meaning pointer to array with 8 char elements),
19228 which allows the compiler to do more typechecking; it was like
19229 that back in SSLeay, but with lots of ugly casts.
19231 Introduce new type const_des_cblock.
19235 * Reorganise the PKCS#7 library and get rid of some of the more obvious
19236 problems: find RecipientInfo structure that matches recipient certificate
19237 and initialise the ASN1 structures properly based on passed cipher.
19241 * Belatedly make the BN tests actually check the results.
19245 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
19246 to and from BNs: it was completely broken. New compilation option
19247 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
19248 key elements as negative integers.
19252 * Reorganize and speed up MD5.
19254 *Andy Polyakov <appro@fy.chalmers.se>*
19258 *Richard Levitte <richard@levitte.org>*
19260 * New option -out to asn1parse to allow the parsed structure to be
19261 output to a file. This is most useful when combined with the -strparse
19262 option to examine the output of things like OCTET STRINGS.
19266 * Make SSL library a little more fool-proof by not requiring any longer
19267 that `SSL_set_{accept,connect}_state` be called before
19268 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19269 in many applications because usually everything *appeared* to work as
19270 intended anyway -- now it really works as intended).
19274 * Move openssl.cnf out of lib/.
19278 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19279 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19280 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19282 *Ralf S. Engelschall*
19284 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19285 handle PKCS#7 enveloped data properly.
19287 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19289 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19290 copying pointers. The cert_st handling is changed by this in
19291 various ways (and thus what used to be known as ctx->default_cert
19292 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19293 any longer when s->cert does not give us what we need).
19294 ssl_cert_instantiate becomes obsolete by this change.
19295 As soon as we've got the new code right (possibly it already is?),
19296 we have solved a couple of bugs of the earlier code where s->cert
19297 was used as if it could not have been shared with other SSL structures.
19299 Note that using the SSL API in certain dirty ways now will result
19300 in different behaviour than observed with earlier library versions:
19301 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19302 does not influence s as it used to.
19304 In order to clean up things more thoroughly, inside SSL_SESSION
19305 we don't use CERT any longer, but a new structure SESS_CERT
19306 that holds per-session data (if available); currently, this is
19307 the peer's certificate chain and, for clients, the server's certificate
19308 and temporary key. CERT holds only those values that can have
19309 meaningful defaults in an SSL_CTX.
19313 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19314 from the internal representation. Various PKCS#7 fixes: remove some
19315 evil casts and set the enc_dig_alg field properly based on the signing
19320 * Allow PKCS#12 password to be set from the command line or the
19321 environment. Let 'ca' get its config file name from the environment
19322 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19327 * Allow certificate policies extension to use an IA5STRING for the
19328 organization field. This is contrary to the PKIX definition but
19329 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19334 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19335 without disallowing inline assembler and the like for non-pedantic builds.
19339 * Support Borland C++ builder.
19341 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19347 * SHA-1 cleanups and performance enhancements.
19349 *Andy Polyakov <appro@fy.chalmers.se>*
19351 * Sparc v8plus assembler for the bignum library.
19353 *Andy Polyakov <appro@fy.chalmers.se>*
19355 * Accept any -xxx and +xxx compiler options in Configure.
19359 * Update HPUX configuration.
19363 * Add missing `sk_<type>_unshift()` function to safestack.h
19365 *Ralf S. Engelschall*
19367 * New function SSL_CTX_use_certificate_chain_file that sets the
19368 "extra_cert"s in addition to the certificate. (This makes sense
19369 only for "PEM" format files, as chains as a whole are not
19374 * Support verify_depth from the SSL API.
19375 x509_vfy.c had what can be considered an off-by-one-error:
19376 Its depth (which was not part of the external interface)
19377 was actually counting the number of certificates in a chain;
19378 now it really counts the depth.
19382 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19383 instead of X509err, which often resulted in confusing error
19384 messages since the error codes are not globally unique
19385 (e.g. an alleged error in ssl3_accept when a certificate
19386 didn't match the private key).
19388 * New function SSL_CTX_set_session_id_context that allows to set a default
19389 value (so that you don't need SSL_set_session_id_context for each
19390 connection using the SSL_CTX).
19394 * OAEP decoding bug fix.
19398 * Support INSTALL_PREFIX for package builders, as proposed by
19403 * New Configure options "threads" and "no-threads". For systems
19404 where the proper compiler options are known (currently Solaris
19405 and Linux), "threads" is the default.
19409 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19413 * Install various scripts to $(OPENSSLDIR)/misc, not to
19414 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19415 such as /usr/local/bin.
19419 * "make linux-shared" to build shared libraries.
19421 *Niels Poppe <niels@netbox.org>*
19423 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19427 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19428 extension adding in x509 utility.
19432 * Remove NOPROTO sections and error code comments.
19436 * Partial rewrite of the DEF file generator to now parse the ANSI
19441 * New Configure options --prefix=DIR and --openssldir=DIR.
19445 * Complete rewrite of the error code script(s). It is all now handled
19446 by one script at the top level which handles error code gathering,
19447 header rewriting and C source file generation. It should be much better
19448 than the old method: it now uses a modified version of Ulf's parser to
19449 read the ANSI prototypes in all header files (thus the old K&R definitions
19450 aren't needed for error creation any more) and do a better job of
19451 translating function codes into names. The old 'ASN1 error code embedded
19452 in a comment' is no longer necessary and it doesn't use .err files which
19453 have now been deleted. Also the error code call doesn't have to appear all
19454 on one line (which resulted in some large lines...).
19458 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19462 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19463 0 (which usually indicates a closed connection), but continue reading.
19467 * Fix some race conditions.
19471 * Add support for CRL distribution points extension. Add Certificate
19472 Policies and CRL distribution points documentation.
19476 * Move the autogenerated header file parts to crypto/opensslconf.h.
19480 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19481 8 of keying material. Merlin has also confirmed interop with this fix
19482 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19484 *Merlin Hughes <merlin@baltimore.ie>*
19486 * Fix lots of warnings.
19488 *Richard Levitte <levitte@stacken.kth.se>*
19490 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19491 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19493 *Richard Levitte <levitte@stacken.kth.se>*
19495 * Fix problems with sizeof(long) == 8.
19497 *Andy Polyakov <appro@fy.chalmers.se>*
19499 * Change functions to ANSI C.
19503 * Fix typos in error codes.
19505 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19507 * Remove defunct assembler files from Configure.
19511 * SPARC v8 assembler BIGNUM implementation.
19513 *Andy Polyakov <appro@fy.chalmers.se>*
19515 * Support for Certificate Policies extension: both print and set.
19516 Various additions to support the r2i method this uses.
19520 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19521 return a const string when you are expecting an allocated buffer.
19525 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19526 types DirectoryString and DisplayText.
19530 * Add code to allow r2i extensions to access the configuration database,
19531 add an LHASH database driver and add several ctx helper functions.
19535 * Fix an evil bug in bn_expand2() which caused various BN functions to
19536 fail when they extended the size of a BIGNUM.
19540 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19541 support typesafe stack.
19545 * Fix typo in SSL_[gs]et_options().
19547 *Nils Frostberg <nils@medcom.se>*
19549 * Delete various functions and files that belonged to the (now obsolete)
19550 old X509V3 handling code.
19554 * New Configure option "rsaref".
19558 * Don't auto-generate pem.h.
19562 * Introduce type-safe ASN.1 SETs.
19566 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19568 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19570 * Introduce type-safe STACKs. This will almost certainly break lots of code
19571 that links with OpenSSL (well at least cause lots of warnings), but fear
19572 not: the conversion is trivial, and it eliminates loads of evil casts. A
19573 few STACKed things have been converted already. Feel free to convert more.
19574 In the fullness of time, I'll do away with the STACK type altogether.
19578 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19579 specified in `<certfile>` by updating the entry in the index.txt file.
19580 This way one no longer has to edit the index.txt file manually for
19581 revoking a certificate. The -revoke option does the gory details now.
19583 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19585 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19586 `-text` option at all and this way the `-noout -text` combination was
19587 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19589 *Ralf S. Engelschall*
19591 * Make sure a corresponding plain text error message exists for the
19592 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19593 verify callback function determined that a certificate was revoked.
19595 *Ralf S. Engelschall*
19597 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19598 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19599 all available ciphers including rc5, which was forgotten until now.
19600 In order to let the testing shell script know which algorithms
19601 are available, a new (up to now undocumented) command
19602 `openssl list-cipher-commands` is used.
19606 * Bugfix: s_client occasionally would sleep in select() when
19607 it should have checked SSL_pending() first.
19611 * New functions DSA_do_sign and DSA_do_verify to provide access to
19612 the raw DSA values prior to ASN.1 encoding.
19616 * Tweaks to Configure
19618 *Niels Poppe <niels@netbox.org>*
19620 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19625 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19629 * New config option to avoid instructions that are illegal on the 80386.
19630 The default code is faster, but requires at least a 486.
19634 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19635 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19636 same as SSL2_VERSION anyway.
19640 * New "-showcerts" option for s_client.
19644 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19645 application. Various cleanups and fixes.
19649 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19650 modify error routines to work internally. Add error codes and PBE init
19651 to library startup routines.
19655 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19656 packing functions to asn1 and evp. Changed function names and error
19657 codes along the way.
19661 * PKCS12 integration: and so it begins... First of several patches to
19662 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19663 objects to objects.h
19667 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19668 and display support for Thawte strong extranet extension.
19672 * Add LinuxPPC support.
19674 *Jeff Dubrule <igor@pobox.org>*
19676 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19677 bn_div_words in alpha.s.
19679 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19681 * Make sure the RSA OAEP test is skipped under -DRSAref because
19682 OAEP isn't supported when OpenSSL is built with RSAref.
19684 *Ulf Moeller <ulf@fitug.de>*
19686 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19687 so they no longer are missing under -DNOPROTO.
19689 *Soren S. Jorvang <soren@t.dk>*
19691 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19693 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19694 doesn't work when the session is reused. Coming soon!
19698 * Fix a security hole, that allows sessions to be reused in the wrong
19699 context thus bypassing client cert protection! All software that uses
19700 client certs and session caches in multiple contexts NEEDS PATCHING to
19701 allow session reuse! A fuller solution is in the works.
19703 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19705 * Some more source tree cleanups (removed obsolete files
19706 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19707 permission on "config" script to be executable) and a fix for the INSTALL
19710 *Ulf Moeller <ulf@fitug.de>*
19712 * Remove some legacy and erroneous uses of malloc, free instead of
19715 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19717 * Make rsa_oaep_test return non-zero on error.
19719 *Ulf Moeller <ulf@fitug.de>*
19721 * Add support for native Solaris shared libraries. Configure
19722 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19723 if someone would make that last step automatic.
19725 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19727 * ctx_size was not built with the right compiler during "make links". Fixed.
19731 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19732 except NULL ciphers". This means the default cipher list will no longer
19733 enable NULL ciphers. They need to be specifically enabled e.g. with
19734 the string "DEFAULT:eNULL".
19738 * Fix to RSA private encryption routines: if p < q then it would
19739 occasionally produce an invalid result. This will only happen with
19740 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19744 * Be less restrictive and allow also `perl util/perlpath.pl
19745 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19746 because this way one can also use an interpreter named `perl5` (which is
19747 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19748 installed as `perl`).
19750 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19752 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19754 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19756 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19757 advapi32.lib to Win32 build and change the pem test comparison
19758 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19759 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19760 and crypto/des/ede_cbcm_enc.c.
19764 * DES quad checksum was broken on big-endian architectures. Fixed.
19768 * Comment out two functions in bio.h that aren't implemented. Fix up the
19769 Win32 test batch file so it (might) work again. The Win32 test batch file
19770 is horrible: I feel ill....
19774 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19775 in e_os.h. Audit of header files to check ANSI and non ANSI
19776 sections: 10 functions were absent from non ANSI section and not exported
19777 from Windows DLLs. Fixed up libeay.num for new functions.
19781 * Make `openssl version` output lines consistent.
19783 *Ralf S. Engelschall*
19785 * Fix Win32 symbol export lists for BIO functions: Added
19786 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19787 to ms/libeay{16,32}.def.
19789 *Ralf S. Engelschall*
19791 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19792 fine under Unix and passes some trivial tests I've now added. But the
19793 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19794 added to make sure no one expects that this stuff really works in the
19795 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19796 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19799 *Ralf S. Engelschall*
19801 * Fix the generation of two part addresses in perl.
19803 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19805 * Add config entry for Linux on MIPS.
19807 *John Tobey <jtobey@channel1.com>*
19809 * Make links whenever Configure is run, unless we are on Windoze.
19813 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19814 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19819 * Add a useful kludge to allow package maintainers to specify compiler and
19820 other platforms details on the command line without having to patch the
19821 Configure script every time: One now can use
19822 `perl Configure <id>:<details>`,
19823 i.e. platform ids are allowed to have details appended
19824 to them (separated by colons). This is treated as there would be a static
19825 pre-configured entry in Configure's %table under key `<id>` with value
19826 `<details>` and `perl Configure <id>` is called. So, when you want to
19827 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19828 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19829 now, which overrides the FreeBSD-elf entry on-the-fly.
19831 *Ralf S. Engelschall*
19833 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19837 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19838 on the `perl Configure ...` command line. This way one can compile
19839 OpenSSL libraries with Position Independent Code (PIC) which is needed
19840 for linking it into DSOs.
19842 *Ralf S. Engelschall*
19844 * Remarkably, export ciphers were totally broken and no-one had noticed!
19849 * Cleaned up the LICENSE document: The official contact for any license
19850 questions now is the OpenSSL core team under openssl-core@openssl.org.
19851 And add a paragraph about the dual-license situation to make sure people
19852 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19853 to the OpenSSL toolkit.
19855 *Ralf S. Engelschall*
19857 * General source tree makefile cleanups: Made `making xxx in yyy...`
19858 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19859 Additionally cleaned up the `make links` target: Remove unnecessary
19860 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19861 to speed processing and no longer clutter the display with confusing
19862 stuff. Instead only the actually done links are displayed.
19864 *Ralf S. Engelschall*
19866 * Permit null encryption ciphersuites, used for authentication only. It used
19867 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19868 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19873 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19874 signed attributes when verifying signatures (this would break them),
19875 the detached data encoding was wrong and public keys obtained using
19876 X509_get_pubkey() weren't freed.
19880 * Add text documentation for the BUFFER functions. Also added a work around
19881 to a Win95 console bug. This was triggered by the password read stuff: the
19882 last character typed gets carried over to the next fread(). If you were
19883 generating a new cert request using 'req' for example then the last
19884 character of the passphrase would be CR which would then enter the first
19889 * Added the new 'Includes OpenSSL Cryptography Software' button as
19890 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19891 button and can be used by applications based on OpenSSL to show the
19892 relationship to the OpenSSL project.
19894 *Ralf S. Engelschall*
19896 * Remove confusing variables in function signatures in files
19897 ssl/ssl_lib.c and ssl/ssl.h.
19899 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19901 * Don't install bss_file.c under PREFIX/include/
19903 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19905 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19906 functions that return function pointers and has support for NT specific
19907 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19908 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19909 unsigned to signed types: this was killing the Win32 compile.
19913 * Add new certificate file to stack functions,
19914 SSL_add_dir_cert_subjects_to_stack() and
19915 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19916 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19917 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19918 This means that Apache-SSL and similar packages don't have to mess around
19919 to add as many CAs as they want to the preferred list.
19923 * Experiment with doxygen documentation. Currently only partially applied to
19925 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19926 openssl.doxy as the configuration file.
19930 * Get rid of remaining C++-style comments which strict C compilers hate.
19932 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19934 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19935 compiled in by default: it has problems with large keys.
19939 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19940 DH private keys and/or callback functions which directly correspond to
19941 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19942 is needed for applications which have to configure certificates on a
19943 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19945 For the RSA certificate situation is makes no difference, but
19946 for the DSA certificate situation this fixes the "no shared cipher"
19947 problem where the OpenSSL cipher selection procedure failed because the
19948 temporary keys were not overtaken from the context and the API provided
19949 no way to reconfigure them.
19950 The new functions now let applications reconfigure the stuff and they
19951 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19952 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19953 non-public-API function ssl_cert_instantiate() is used as a helper
19954 function and also to reduce code redundancy inside ssl_rsa.c.
19956 *Ralf S. Engelschall*
19958 * Move s_server -dcert and -dkey options out of the undocumented feature
19959 area because they are useful for the DSA situation and should be
19960 recognized by the users.
19962 *Ralf S. Engelschall*
19964 * Fix the cipher decision scheme for export ciphers: the export bits are
19965 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19966 SSL_EXP_MASK. So, the original variable has to be used instead of the
19967 already masked variable.
19969 *Richard Levitte <levitte@stacken.kth.se>*
19971 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19973 *Richard Levitte <levitte@stacken.kth.se>*
19975 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19976 from `int` to `unsigned int` because it is a length and initialized by
19977 EVP_DigestFinal() which expects an `unsigned int *`.
19979 *Richard Levitte <levitte@stacken.kth.se>*
19981 * Don't hard-code path to Perl interpreter on shebang line of Configure
19982 script. Instead use the usual Shell->Perl transition trick.
19984 *Ralf S. Engelschall*
19986 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19987 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19988 -noout -modulus` as it's already the case for `openssl rsa -noout
19989 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19990 currently the public key is printed (a decision which was already done by
19991 `openssl dsa -modulus` in the past) which serves a similar purpose.
19992 Additionally the NO_RSA no longer completely removes the whole -modulus
19993 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19996 *Ralf S. Engelschall*
19998 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19999 BIO. See the source (crypto/evp/bio_ok.c) for more info.
20001 *Arne Ansper <arne@ats.cyber.ee>*
20003 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
20004 to be added. Now both 'req' and 'ca' can use new objects defined in the
20009 * Add cool BIO that does syslog (or event log on NT).
20011 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
20013 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
20014 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
20015 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
20016 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
20020 * Add preliminary config info for new extension code.
20024 * Make RSA_NO_PADDING really use no padding.
20026 *Ulf Moeller <ulf@fitug.de>*
20028 * Generate errors when private/public key check is done.
20032 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
20033 for some CRL extensions and new objects added.
20037 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
20038 key usage extension and fuller support for authority key id.
20042 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
20043 padding method for RSA, which is recommended for new applications in PKCS
20044 #1 v2.0 (RFC 2437, October 1998).
20045 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
20046 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
20047 against Bleichbacher's attack on RSA.
20048 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
20051 * Updates to the new SSL compression code
20053 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20055 * Fix so that the version number in the master secret, when passed
20056 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
20057 (because the server will not accept higher), that the version number
20058 is 0x03,0x01, not 0x03,0x00
20060 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20062 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
20063 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
20064 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
20068 * Support for RAW extensions where an arbitrary extension can be
20069 created by including its DER encoding. See `apps/openssl.cnf` for
20074 * Make sure latest Perl versions don't interpret some generated C array
20075 code as Perl array code in the crypto/err/err_genc.pl script.
20077 *Lars Weber <3weber@informatik.uni-hamburg.de>*
20079 * Modify ms/do_ms.bat to not generate assembly language makefiles since
20080 not many people have the assembler. Various Win32 compilation fixes and
20081 update to the INSTALL.W32 file with (hopefully) more accurate Win32
20082 build instructions.
20086 * Modify configure script 'Configure' to automatically create crypto/date.h
20087 file under Win32 and also build pem.h from pem.org. New script
20088 util/mkfiles.pl to create the MINFO file on environments that can't do a
20089 'make files': perl util/mkfiles.pl >MINFO should work.
20093 * Major rework of DES function declarations, in the pursuit of correctness
20094 and purity. As a result, many evil casts evaporated, and some weirdness,
20095 too. You may find this causes warnings in your code. Zapping your evil
20096 casts will probably fix them. Mostly.
20100 * Fix for a typo in asn1.h. Bug fix to object creation script
20101 obj_dat.pl. It considered a zero in an object definition to mean
20102 "end of object": none of the objects in objects.h have any zeros
20103 so it wasn't spotted.
20105 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
20107 * Add support for Triple DES Cipher Block Chaining with Output Feedback
20108 Masking (CBCM). In the absence of test vectors, the best I have been able
20109 to do is check that the decrypt undoes the encrypt, so far. Send me test
20110 vectors if you have them.
20114 * Correct calculation of key length for export ciphers (too much space was
20115 allocated for null ciphers). This has not been tested!
20119 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
20120 message is now correct (it understands "crypto" and "ssl" on its
20121 command line). There is also now an "update" option. This will update
20122 the util/ssleay.num and util/libeay.num files with any new functions.
20124 perl util/mkdef.pl crypto ssl update
20125 it will update them.
20129 * Overhauled the Perl interface:
20130 - ported BN stuff to OpenSSL's different BN library
20131 - made the perl/ source tree CVS-aware
20132 - renamed the package from SSLeay to OpenSSL (the files still contain
20133 their history because I've copied them in the repository)
20134 - removed obsolete files (the test scripts will be replaced
20135 by better Test::Harness variants in the future)
20137 *Ralf S. Engelschall*
20139 * First cut for a very conservative source tree cleanup:
20140 1. merge various obsolete readme texts into doc/ssleay.txt
20141 where we collect the old documents and readme texts.
20142 2. remove the first part of files where I'm already sure that we no
20143 longer need them because of three reasons: either they are just temporary
20144 files which were left by Eric or they are preserved original files where
20145 I've verified that the diff is also available in the CVS via "cvs diff
20146 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
20147 the crypto/md/ stuff).
20149 *Ralf S. Engelschall*
20151 * More extension code. Incomplete support for subject and issuer alt
20152 name, issuer and authority key id. Change the i2v function parameters
20153 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
20154 what that's for :-) Fix to ASN1 macro which messed up
20155 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
20159 * Preliminary support for ENUMERATED type. This is largely copied from the
20164 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
20166 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20168 * Make sure `make rehash` target really finds the `openssl` program.
20170 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
20172 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
20173 like to hear about it if this slows down other processors.
20177 * Add CygWin32 platform information to Configure script.
20179 *Alan Batie <batie@aahz.jf.intel.com>*
20181 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
20183 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
20185 * New program nseq to manipulate netscape certificate sequences
20189 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
20194 * Fixes to BN code. Previously the default was to define BN_RECURSION
20195 but the BN code had some problems that would cause failures when
20196 doing certificate verification and some other functions.
20198 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20200 * Add ASN1 and PEM code to support netscape certificate sequences.
20204 * Add ASN1 and PEM code to support netscape certificate sequences.
20208 * Add several PKIX and private extended key usage OIDs.
20212 * Modify the 'ca' program to handle the new extension code. Modify
20213 openssl.cnf for new extension format, add comments.
20217 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
20218 and add a sample to openssl.cnf so req -x509 now adds appropriate
20223 * Continued X509 V3 changes. Add to other makefiles, integrate with the
20224 error code, add initial support to X509_print() and x509 application.
20228 * Takes a deep breath and start adding X509 V3 extension support code. Add
20229 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
20230 stuff is currently isolated and isn't even compiled yet.
20234 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
20235 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
20236 Removed the versions check from X509 routines when loading extensions:
20237 this allows certain broken certificates that don't set the version
20238 properly to be processed.
20242 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
20243 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
20244 can still be regenerated with "make depend".
20248 * Spelling mistake in C version of CAST-128.
20250 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
20252 * Changes to the error generation code. The perl script err-code.pl
20253 now reads in the old error codes and retains the old numbers, only
20254 adding new ones if necessary. It also only changes the .err files if new
20255 codes are added. The makefiles have been modified to only insert errors
20256 when needed (to avoid needlessly modifying header files). This is done
20257 by only inserting errors if the .err file is newer than the auto generated
20258 C file. To rebuild all the error codes from scratch (the old behaviour)
20259 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
20260 or delete all the .err files.
20264 * CAST-128 was incorrectly implemented for short keys. The C version has
20265 been fixed, but is untested. The assembler versions are also fixed, but
20266 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20267 to regenerate it if needed.
20268 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20269 Hagino <itojun@kame.net>*
20271 * File was opened incorrectly in randfile.c.
20273 *Ulf Möller <ulf@fitug.de>*
20275 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20276 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20277 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20278 al: it's just almost always a UTCTime. Note this patch adds new error
20279 codes so do a "make errors" if there are problems.
20283 * Correct Linux 1 recognition in config.
20285 *Ulf Möller <ulf@fitug.de>*
20287 * Remove pointless MD5 hash when using DSA keys in ca.
20289 *Anonymous <nobody@replay.com>*
20291 * Generate an error if given an empty string as a cert directory. Also
20292 generate an error if handed NULL (previously returned 0 to indicate an
20293 error, but didn't set one).
20295 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20297 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20301 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20302 parameters. This was causing a warning which killed off the Win32 compile.
20306 * Remove C++ style comments from crypto/bn/bn_local.h.
20308 *Neil Costigan <neil.costigan@celocom.com>*
20310 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20311 based on a text string, looking up short and long names and finally
20312 "dot" format. The "dot" format stuff didn't work. Added new function
20313 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20314 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20315 OID is not part of the table.
20319 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20320 X509_LOOKUP_by_alias().
20324 * Sort openssl functions by name.
20328 * Get the `gendsa` command working and add it to the `list` command. Remove
20329 encryption from sample DSA keys (in case anyone is interested the password
20334 * Make *all* `*_free` functions accept a NULL pointer.
20336 *Frans Heymans <fheymans@isaserver.be>*
20338 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20341 *Anonymous <nobody@replay.com>*
20343 * s_server should send the CAfile as acceptable CAs, not its own cert.
20345 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20347 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20349 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20351 * Temp key "for export" tests were wrong in s3_srvr.c.
20353 *Anonymous <nobody@replay.com>*
20355 * Add prototype for temp key callback functions
20356 SSL_CTX_set_tmp_{rsa,dh}_callback().
20360 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20361 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20365 * X509_name_add_entry() freed the wrong thing after an error.
20367 *Arne Ansper <arne@ats.cyber.ee>*
20369 * rsa_eay.c would attempt to free a NULL context.
20371 *Arne Ansper <arne@ats.cyber.ee>*
20373 * BIO_s_socket() had a broken should_retry() on Windoze.
20375 *Arne Ansper <arne@ats.cyber.ee>*
20377 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20379 *Arne Ansper <arne@ats.cyber.ee>*
20381 * Make sure the already existing X509_STORE->depth variable is initialized
20382 in X509_STORE_new(), but document the fact that this variable is still
20383 unused in the certificate verification process.
20385 *Ralf S. Engelschall*
20387 * Fix the various library and `apps/` files to free up pkeys obtained from
20388 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20392 * Fix reference counting in X509_PUBKEY_get(). This makes
20393 demos/maurice/example2.c work, amongst others, probably.
20395 *Steve Henson and Ben Laurie*
20397 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20398 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20399 are no longer created. This way we have a single and consistent command
20400 line interface `openssl <command>`, similar to `cvs <command>`.
20402 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20404 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20405 BIT STRING wrapper always have zero unused bits.
20409 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20413 * Make the top-level INSTALL documentation easier to understand.
20417 * Makefiles updated to exit if an error occurs in a sub-directory
20418 make (including if user presses ^C) [Paul Sutton]
20420 * Make Montgomery context stuff explicit in RSA data structure.
20424 * Fix build order of pem and err to allow for generated pem.h.
20428 * Fix renumbering bug in X509_NAME_delete_entry().
20432 * Enhanced the err-ins.pl script so it makes the error library number
20433 global and can add a library name. This is needed for external ASN1 and
20434 other error libraries.
20438 * Fixed sk_insert which never worked properly.
20442 * Fix ASN1 macros so they can handle indefinite length constructed
20443 EXPLICIT tags. Some non standard certificates use these: they can now
20448 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20449 into a single doc/ssleay.txt bundle. This way the information is still
20450 preserved but no longer messes up this directory. Now it's new room for
20451 the new set of documentation files.
20453 *Ralf S. Engelschall*
20455 * SETs were incorrectly DER encoded. This was a major pain, because they
20456 shared code with SEQUENCEs, which aren't coded the same. This means that
20457 almost everything to do with SETs or SEQUENCEs has either changed name or
20458 number of arguments.
20460 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20462 * Fix test data to work with the above.
20466 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20467 was already fixed by Eric for 0.9.1 it seems.
20469 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20471 * Autodetect FreeBSD3.
20475 * Fix various bugs in Configure. This affects the following platforms:
20479 unixware-2.0-pentium
20484 * Eliminate generated files from CVS. Reorder tests to regenerate files
20485 before they are needed.
20489 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20493 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20495 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20496 changed SSLeay to OpenSSL in version strings.
20498 *Ralf S. Engelschall*
20500 * Some fixups to the top-level documents.
20504 * Fixed the nasty bug where rsaref.h was not found under compile-time
20505 because the symlink to include/ was missing.
20507 *Ralf S. Engelschall*
20509 * Incorporated the popular no-RSA/DSA-only patches
20510 which allow to compile an RSA-free SSLeay.
20512 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20514 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20515 when "ssleay" is still not found.
20517 *Ralf S. Engelschall*
20519 * Added more platforms to Configure: Cray T3E, HPUX 11,
20521 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20523 * Updated the README file.
20525 *Ralf S. Engelschall*
20527 * Added various .cvsignore files in the CVS repository subdirs
20528 to make a "cvs update" really silent.
20530 *Ralf S. Engelschall*
20532 * Recompiled the error-definition header files and added
20533 missing symbols to the Win32 linker tables.
20535 *Ralf S. Engelschall*
20537 * Cleaned up the top-level documents;
20538 o new files: CHANGES and LICENSE
20539 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20540 o merged COPYRIGHT into LICENSE
20541 o removed obsolete TODO file
20542 o renamed MICROSOFT to INSTALL.W32
20544 *Ralf S. Engelschall*
20546 * Removed dummy files from the 0.9.1b source tree:
20547 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20548 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20549 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20550 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20551 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20553 *Ralf S. Engelschall*
20555 * Added various platform portability fixes.
20559 * The Genesis of the OpenSSL rpject:
20560 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20561 Young and Tim J. Hudson created while they were working for C2Net until
20564 *The OpenSSL Project*
20566 ### Changes between 0.9.0b and 0.9.1b [not released]
20568 * Updated a few CA certificates under certs/
20572 * Changed some BIGNUM api stuff.
20576 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20577 DGUX x86, Linux Alpha, etc.
20581 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20582 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20587 * Add -strparse option to asn1pars program which parses nested
20590 *Dr Stephen Henson <shenson@bigfoot.com>*
20592 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20596 * DSA fix for "ca" program.
20600 * Added "-genkey" option to "dsaparam" program.
20604 * Added RIPE MD160 (rmd160) message digest.
20608 * Added -a (all) option to "ssleay version" command.
20612 * Added PLATFORM define which is the id given to Configure.
20616 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20620 * Extended the ASN.1 parser routines.
20624 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20628 * Added a BN_CTX to the BN library.
20632 * Fixed the weak key values in DES library
20636 * Changed API in EVP library for cipher aliases.
20640 * Added support for RC2/64bit cipher.
20644 * Converted the lhash library to the crypto/mem.c functions.
20648 * Added more recognized ASN.1 object ids.
20652 * Added more RSA padding checks for SSL/TLS.
20656 * Added BIO proxy/filter functionality.
20660 * Added extra_certs to SSL_CTX which can be used
20661 send extra CA certificates to the client in the CA cert chain sending
20662 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20666 * Now Fortezza is denied in the authentication phase because
20667 this is key exchange mechanism is not supported by SSLeay at all.
20671 * Additional PKCS1 checks.
20675 * Support the string "TLSv1" for all TLS v1 ciphers.
20679 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20680 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20684 * Fixed a few memory leaks.
20688 * Fixed various code and comment typos.
20692 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20693 bytes sent in the client random.
20695 *Edward Bishop <ebishop@spyglass.com>*
20699 [CVE-2024-2511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-2511
20700 [CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727
20701 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20702 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20703 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20704 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20705 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20706 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20707 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20708 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20709 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20710 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20711 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20712 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20713 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20714 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20715 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20716 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20717 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20718 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20719 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20720 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20721 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20722 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20723 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20724 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20725 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20726 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20727 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20728 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20729 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20730 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20731 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20732 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20733 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20734 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20735 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20736 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20737 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20738 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20739 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20740 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20741 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20742 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20743 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20744 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20745 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20746 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20747 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20748 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20749 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20750 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20751 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20752 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20753 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20754 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20755 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20756 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20757 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20758 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20759 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20760 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20761 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20762 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20763 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20764 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20765 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20766 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20767 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20768 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20769 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20770 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20771 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20772 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20773 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20774 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20775 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20776 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20777 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20778 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20779 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20780 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20781 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20782 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20783 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20784 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20785 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20786 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20787 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20788 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20789 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20790 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20791 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20792 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20793 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20794 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20795 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20796 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20797 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20798 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20799 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20800 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20801 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20802 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20803 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20804 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20805 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20806 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20807 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20808 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20809 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20810 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20811 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20812 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20813 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20814 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20815 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20816 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20817 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20818 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20819 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20820 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20821 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20822 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20823 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20824 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20825 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20826 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20827 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20828 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20829 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20830 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20831 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20832 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20833 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20834 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20835 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20836 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20837 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20838 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20839 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20840 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20841 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20842 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20843 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20844 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20845 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20846 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20847 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20848 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20849 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20850 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20851 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20852 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20853 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20854 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20855 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20856 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20857 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20858 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20859 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20860 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20861 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20862 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20863 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20864 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20865 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20866 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20867 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20868 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20869 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20870 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20871 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20872 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20873 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20874 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20875 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20876 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20877 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20878 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20879 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20880 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20881 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20882 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20883 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20884 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20885 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20886 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20887 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655