4 This is a detailed breakdown of significant changes. For a high-level overview
5 of changes in each release, see [NEWS.md](./NEWS.md).
7 For a full list of changes, see the [git commit log][log] and pick the
8 appropriate release branch.
10 [log]: https://github.com/openssl/openssl/commits/
15 - [OpenSSL 3.3](#openssl-33)
16 - [OpenSSL 3.2](#openssl-32)
17 - [OpenSSL 3.1](#openssl-31)
18 - [OpenSSL 3.0](#openssl-30)
19 - [OpenSSL 1.1.1](#openssl-111)
20 - [OpenSSL 1.1.0](#openssl-110)
21 - [OpenSSL 1.0.2](#openssl-102)
22 - [OpenSSL 1.0.1](#openssl-101)
23 - [OpenSSL 1.0.0](#openssl-100)
24 - [OpenSSL 0.9.x](#openssl-09x)
29 ### Changes between 3.2 and 3.3 [xx XXX xxxx]
31 * The activate and soft_load configuration settings for providers in
32 openssl.cnf have been updated to require a value of [1|yes|true|on]
33 (in lower or UPPER case) to enable the setting. Conversely a value
34 of [0|no|false|off] will disable the setting. All other values, or the
35 omission of a value for these settings will result in an error.
39 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
40 if called with a NULL stack argument.
44 * In `openssl speed`, changed the default hash function used with `hmac` from
49 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
50 - `certProfile` request message header and respective `-profile` CLI option
51 - support for delayed delivery of all types of response messages
55 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
56 be less hard coded in the build file templates, and to allow easier
57 addition of more exporters. With that, an exporter for CMake is also
62 * The BLAKE2s hash algorithm matches BLAKE2b's support
63 for configurable output length.
67 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
68 server to prefer session resumption using PSK-only key exchange over PSK
69 with DHE, if both are available.
71 *Markus Minichmayr, Tapkey GmbH*
76 ### Changes between 3.2.0 and 3.2.1 [xx XXX xxxx]
78 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
79 for PowerPC CPUs saves the contents of vector registers in different
80 order than they are restored. Thus the contents of some of these vector
81 registers is corrupted when returning to the caller. The vulnerable code is
82 used only on newer PowerPC processors supporting the PowerISA 2.07
85 The consequences of this kind of internal application state corruption can
86 be various - from no consequences, if the calling application does not
87 depend on the contents of non-volatile XMM registers at all, to the worst
88 consequences, where the attacker could get complete control of the
89 application process. However unless the compiler uses the vector registers
90 for storing pointers, the most likely consequence, if any, would be an
91 incorrect result of some application dependent calculations or a crash
92 leading to a denial of service.
98 * Disable building QUIC server utility when OpenSSL is configured with
103 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
105 * The EVP_PKEY_fromdata function has been augmented to allow for the derivation
106 of CRT (Chinese Remainder Theorem) parameters when requested. See the
107 OSSL_PKEY_PARAM_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
111 * The BLAKE2b hash algorithm supports a configurable output length
112 by setting the "size" parameter.
114 *Čestmír Kalina and Tomáš Mráz*
116 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
120 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
121 and the corresponding provider-storemgmt API function
122 OSSL_FUNC_store_delete().
126 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
127 a passphrase callback when opening a store.
131 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
132 from 8 bytes to 16 bytes.
133 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
134 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
135 requires a salt length of 128 bits. This affects OpenSSL command line
136 applications such as "genrsa" and "pkcs8" and API's such as
137 PEM_write_bio_PrivateKey() that are reliant on the default value.
138 The additional commandline option 'saltlen' has been added to the
139 OpenSSL command line applications for "pkcs8" and "enc" to allow the
140 salt length to be set to a non default value.
144 * Changed the default value of the `ess_cert_id_alg` configuration
145 option which is used to calculate the TSA's public key certificate
146 identifier. The default algorithm is updated to be sha256 instead
149 *Małgorzata Olszówka*
151 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
152 table for point multiplication of the base point, which increases the size of
153 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
154 been added to disable the precomputed table.
158 * Added client side support for QUIC
160 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
162 * Added multiple tutorials on the OpenSSL library and in particular
163 on writing various clients (using TLS and QUIC protocols) with libssl.
167 * Added secp384r1 implementation using Solinas' reduction to improve
168 speed of the NIST P-384 elliptic curve. To enable the implementation
169 the build option `enable-ec_nistp_64_gcc_128` must be used.
173 * Improved RFC7468 compliance of the asn1parse command.
175 *Matthias St. Pierre*
177 * Added SHA256/192 algorithm support.
181 * Added support for securely getting root CA certificate update in
186 * Improved contention on global write locks by using more read locks where
191 * Improved performance of OSSL_PARAM lookups in performance critical
196 * Added the SSL_get0_group_name() function to provide access to the
197 name of the group used for the TLS key exchange.
201 * Provide a new configure option `no-http` that can be used to disable the
202 HTTP support. Provide new configure options `no-apps` and `no-docs` to
203 disable building the openssl command line application and the documentation.
207 * Provide a new configure option `no-ecx` that can be used to disable the
208 X25519, X448, and EdDSA support.
212 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
213 the EVP_KDF_CTX_set_params() function they are now concatenated not just
214 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
218 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
219 the provider context as a parameter.
223 * TLS round-trip time calculation was added by a Brigham Young University
224 Capstone team partnering with Sandia National Laboratories. A new function
225 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
230 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
231 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
232 option. Use of the "advanced" s_client command command via the "-adv" option
237 * Added an "advanced" command mode to s_client. Use this with the "-adv"
238 option. The old "basic" command mode recognises certain letters that must
239 always appear at the start of a line and cannot be escaped. The advanced
240 command mode enables commands to be entered anywhere and there is an
241 escaping mechanism. After starting s_client with "-adv" type "{help}"
242 to show a list of available commands.
246 * Add Raw Public Key (RFC7250) support. Authentication is supported
247 by matching keys against either local policy (TLSA records synthesised
248 from the expected keys) or DANE (TLSA records obtained by the
249 application from DNS). TLSA records will also match the same key in
250 the server certificate, should RPK use not happen to be negotiated.
254 * Added support for modular exponentiation and CRT offloading for the
259 * Added further assembler code for the RISC-V architecture.
263 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
264 from a given EC_GROUP.
268 * Improved support for non-default library contexts and property queries
269 when parsing PKCS#12 files.
273 * Implemented support for all five instances of EdDSA from RFC8032:
274 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
275 The streaming is not yet supported for the HashEdDSA variants
276 (Ed25519ph and Ed448ph).
280 * Added SM4 optimization for ARM processors using ASIMD and AES HW
285 * Implemented SM4-XTS support.
289 * Added platform-agnostic OSSL_sleep() function.
293 * Implemented deterministic ECDSA signatures (RFC6979) support.
297 * Implemented AES-GCM-SIV (RFC8452) support.
301 * Added support for pluggable (provider-based) TLS signature algorithms.
302 This enables TLS 1.3 authentication operations with algorithms embedded
303 in providers not included by default in OpenSSL. In combination with
304 the already available pluggable KEM and X.509 support, this enables
305 for example suitable providers to deliver post-quantum or quantum-safe
306 cryptography to OpenSSL users.
310 * Added support for pluggable (provider-based) CMS signature algorithms.
311 This enables CMS sign and verify operations with algorithms embedded
312 in providers not included by default in OpenSSL.
316 * Added support for Hybrid Public Key Encryption (HPKE) as defined
317 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
318 Message Layer Security (MLS) and other IETF specifications.
319 HPKE can also be used by other applications that require
320 encrypting "to" an ECDH public key. External APIs are defined in
321 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
325 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
330 * Add support for certificate compression (RFC8879), including
331 library support for Brotli and Zstandard compression.
335 * Add the ability to add custom attributes to PKCS12 files. Add a new API
336 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
337 for a user specified callback and optional argument.
338 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
339 added to the existing STACK_OF attrs.
343 * Major refactor of the libssl record layer.
347 * Add a mac salt length option for the pkcs12 command.
351 * Add more SRTP protection profiles from RFC8723 and RFC8269.
355 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
357 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
359 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
360 supported and enabled.
364 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
365 to the list of ciphersuites providing Perfect Forward Secrecy as
366 required by SECLEVEL >= 3.
368 *Dmitry Belyavskiy, Nicola Tuveri*
370 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
371 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
372 SSL_get0_iana_groups() function-like macro, retrieves the list of
373 supported groups sent by the peer.
374 The function SSL_client_hello_get_extension_order() populates
375 a caller-supplied array with the list of extension types present in the
376 ClientHello, in order of appearance.
380 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
381 to make it possible to use empty passphrase strings.
385 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
389 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
390 to change functions used for allocating the memory of asynchronous call stack.
394 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
398 * A failure exit code is returned when using the openssl x509 command to check
399 certificate attributes and the checks fail.
403 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
404 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
405 of 160 bits and above and less than 224 bits were previously accepted by
406 default but are now no longer allowed. By default TLS compression was
407 already disabled in previous OpenSSL versions. At security level 2 it cannot
412 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
417 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
418 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
419 will need to load the legacy crypto provider.
423 * CCM8 cipher suites in TLS have been downgraded to security level zero
424 because they use a short authentication tag which lowers their strength.
428 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
433 * Add X.509 certificate codeSigning purpose and related checks on key usage and
434 extended key usage of the leaf certificate according to the CA/Browser Forum.
438 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
439 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
440 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
441 X.509 version 3 if the certificate information includes X.509 extensions.
445 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
446 such as adding a trace facility for debugging certificate chain building.
450 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
451 in particular supporting requests for central key generation, generalized
452 polling, and various types of genm/genp exchanges defined in CMP Updates.
456 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
457 like correcting the TLS and proxy support and adding tracing for debugging.
461 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
465 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
466 a certificate to be added is already present. `CMS_sign_ex()` and
467 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
468 and no longer throw an error for them.
472 * Fixed and extended `util/check-format.pl` for checking adherence to the
473 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
474 The checks are meanwhile more complete and yield fewer false positives.
478 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
479 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
480 calls. They can be used as the transport BIOs for QUIC.
482 *Hugo Landau, Matt Caswell and Tomáš Mráz*
484 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
485 sending and receiving multiple messages in a single call. An implementation
486 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
490 * Support for loading root certificates from the Windows certificate store
491 has been added. The support is in the form of a store which recognises the
492 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
493 arguments. This store is built by default and can be disabled using the new
494 compile-time option `no-winstore`. This store is not currently used by
495 default and must be loaded explicitly using the above store URI. It is
496 expected to be loaded by default in the future.
500 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
501 kernel versions that support KTLS have a known bug in CCM processing. That
502 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
503 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
508 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
513 * Zerocopy KTLS sendfile() support on Linux.
517 * The OBJ_ calls are now thread safe using a global lock.
521 * New parameter `-digest` for openssl cms command allowing signing
522 pre-computed digests and new CMS API functions supporting that
527 * OPENSSL_malloc() and other allocation functions now raise errors on
528 allocation failures. The callers do not need to explicitly raise errors
529 unless they want to for tracing purposes.
533 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
534 decryption as a protection against Bleichenbacher-like attacks.
535 The RSA decryption API will now return a randomly generated deterministic
536 message instead of an error in case it detects an error when checking
537 padding during PKCS#1 v1.5 decryption. This is a general protection against
538 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
540 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
541 on the RSA decryption context.
545 * Added support for Brainpool curves in TLS-1.3.
547 *Bernd Edlinger and Matt Caswell*
549 * Added OpenBSD specific build targets.
553 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
554 a basic thread pool implementation for select platforms.
561 ### Changes between 3.1.4 and 3.1.5 [xx XXX xxxx]
563 * Fix excessive time spent in DH check / generation with large Q parameter
566 Applications that use the functions DH_generate_key() to generate an
567 X9.42 DH key may experience long delays. Likewise, applications that use
568 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
569 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
570 Where the key or parameters that are being checked have been obtained from
571 an untrusted source this may lead to a Denial of Service.
577 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
579 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
580 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
581 that alter the key or IV length ([CVE-2023-5363]).
585 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
587 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
589 The POLY1305 MAC (message authentication code) implementation in OpenSSL
590 does not save the contents of non-volatile XMM registers on Windows 64
591 platform when calculating the MAC of data larger than 64 bytes. Before
592 returning to the caller all the XMM registers are set to zero rather than
593 restoring their previous content. The vulnerable code is used only on newer
594 x86_64 processors supporting the AVX512-IFMA instructions.
596 The consequences of this kind of internal application state corruption can
597 be various - from no consequences, if the calling application does not
598 depend on the contents of non-volatile XMM registers at all, to the worst
599 consequences, where the attacker could get complete control of the
600 application process. However given the contents of the registers are just
601 zeroized so the attacker cannot put arbitrary values inside, the most likely
602 consequence, if any, would be an incorrect result of some application
603 dependent calculations or a crash leading to a denial of service.
609 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
611 * Fix excessive time spent checking DH q parameter value.
613 The function DH_check() performs various checks on DH parameters. After
614 fixing CVE-2023-3446 it was discovered that a large q parameter value can
615 also trigger an overly long computation during some of these checks.
616 A correct q value, if present, cannot be larger than the modulus p
617 parameter, thus it is unnecessary to perform these checks if q is larger
620 If DH_check() is called with such q parameter value,
621 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
622 intensive checks are skipped.
628 * Fix DH_check() excessive time with over sized modulus.
630 The function DH_check() performs various checks on DH parameters. One of
631 those checks confirms that the modulus ("p" parameter) is not too large.
632 Trying to use a very large modulus is slow and OpenSSL will not normally use
633 a modulus which is over 10,000 bits in length.
635 However the DH_check() function checks numerous aspects of the key or
636 parameters that have been supplied. Some of those checks use the supplied
637 modulus value even if it has already been found to be too large.
639 A new limit has been added to DH_check of 32,768 bits. Supplying a
640 key/parameters with a modulus over this size will simply cause DH_check() to
647 * Do not ignore empty associated data entries with AES-SIV.
649 The AES-SIV algorithm allows for authentication of multiple associated
650 data entries along with the encryption. To authenticate empty data the
651 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
652 with NULL pointer as the output buffer and 0 as the input buffer length.
653 The AES-SIV implementation in OpenSSL just returns success for such call
654 instead of performing the associated data authentication operation.
655 The empty data thus will not be authenticated. ([CVE-2023-2975])
657 Thanks to Juerg Wullschleger (Google) for discovering the issue.
659 The fix changes the authentication tag value and the ciphertext for
660 applications that use empty associated data entries with AES-SIV.
661 To decrypt data encrypted with previous versions of OpenSSL the application
662 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
667 * When building with the `enable-fips` option and using the resulting
668 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
669 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
670 not operate with truncated digests (FIPS 140-3 IG G.R).
674 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
676 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
677 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
679 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
680 numeric text form. For gigantic sub-identifiers, this would take a very
681 long time, the time complexity being O(n^2) where n is the size of that
682 sub-identifier. ([CVE-2023-2650])
684 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
685 IDENTIFIER to canonical numeric text form if the size of that OBJECT
686 IDENTIFIER is 586 bytes or less, and fail otherwise.
688 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
689 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
690 most 128 sub-identifiers, and that the maximum value that each sub-
691 identifier may have is 2^32-1 (4294967295 decimal).
693 For each byte of every sub-identifier, only the 7 lower bits are part of
694 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
695 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
700 * Multiple algorithm implementation fixes for ARM BE platforms.
704 * Added a -pedantic option to fipsinstall that adjusts the various
705 settings to ensure strict FIPS compliance rather than backwards
710 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
711 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
712 trigger a crash of an application using AES-XTS decryption if the memory
713 just after the buffer being decrypted is not mapped.
714 Thanks to Anton Romanov (Amazon) for discovering the issue.
719 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
720 The previous fix for this timing side channel turned out to cause
721 a severe 2-3x performance regression in the typical use case
722 compared to 3.0.7. The new fix uses existing constant time
723 code paths, and restores the previous performance level while
724 fully eliminating all existing timing side channels.
725 The fix was developed by Bernd Edlinger with testing support
730 * Add FIPS provider configuration option to disallow the use of
731 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
732 The option '-no_drbg_truncated_digests' can optionally be
733 supplied to 'openssl fipsinstall'.
737 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
738 that it does not enable policy checking. Thanks to David Benjamin for
739 discovering this issue.
744 * Fixed an issue where invalid certificate policies in leaf certificates are
745 silently ignored by OpenSSL and other certificate policy checks are skipped
746 for that certificate. A malicious CA could use this to deliberately assert
747 invalid certificate policies in order to circumvent policy checking on the
748 certificate altogether.
753 * Limited the number of nodes created in a policy tree to mitigate
754 against CVE-2023-0464. The default limit is set to 1000 nodes, which
755 should be sufficient for most installations. If required, the limit
756 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
757 time define to a desired maximum number of nodes or zero to allow
763 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
765 * Add FIPS provider configuration option to enforce the
766 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
767 The option '-ems-check' can optionally be supplied to
768 'openssl fipsinstall'.
772 * The FIPS provider includes a few non-approved algorithms for
773 backward compatibility purposes and the "fips=yes" property query
774 must be used for all algorithm fetches to ensure FIPS compliance.
776 The algorithms that are included but not approved are Triple DES ECB,
777 Triple DES CBC and EdDSA.
781 * Added support for KMAC in KBKDF.
785 * RNDR and RNDRRS support in provider functions to provide
786 random number generation for Arm CPUs (aarch64).
790 * s_client and s_server apps now explicitly say when the TLS version
791 does not include the renegotiation mechanism. This avoids confusion
792 between that scenario versus when the TLS version includes secure
793 renegotiation but the peer lacks support for it.
797 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
799 *Tomasz Kantecki, Andrey Matyukov*
801 * The various OBJ_* functions have been made thread safe.
805 * Parallel dual-prime 1536/2048-bit modular exponentiation for
806 AVX512_IFMA capable processors.
808 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
810 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
811 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
812 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
813 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
814 `OPENSSL_NO_DEPRECATED_3_1`.
816 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
817 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
818 definitions for these functions regardless of whether
819 `OPENSSL_NO_DEPRECATED_3_1` is defined.
821 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
822 functions regardless of whether they are using them. It is recommended that
823 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
827 * When generating safe-prime DH parameters set the recommended private key
828 length equivalent to minimum key lengths as in RFC 7919.
832 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
833 maximum size that is smaller or equal to the digest length to comply with
834 FIPS 186-4 section 5. This is implemented by a new option
835 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
836 `rsa_pss_saltlen` parameter, which is now the default. Signature
837 verification is not affected by this change and continues to work as before.
844 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
845 listed here are only a brief description.
846 The migration guide contains more detailed information related to new features,
847 breaking changes, and mappings for the large list of deprecated functions.
849 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
851 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
853 * Fixed NULL dereference during PKCS7 data verification.
855 A NULL pointer can be dereferenced when signatures are being
856 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
857 algorithm used for the signature is known to the OpenSSL library but
858 the implementation of the hash algorithm is not available the digest
859 initialization will fail. There is a missing check for the return
860 value from the initialization function which later leads to invalid
861 usage of the digest API most likely leading to a crash.
864 PKCS7 data is processed by the SMIME library calls and also by the
865 time stamp (TS) library calls. The TLS implementation in OpenSSL does
866 not call these functions however third party applications would be
867 affected if they call these functions to verify signatures on untrusted
872 * Fixed X.400 address type confusion in X.509 GeneralName.
874 There is a type confusion vulnerability relating to X.400 address processing
875 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
876 but the public structure definition for GENERAL_NAME incorrectly specified
877 the type of the x400Address field as ASN1_TYPE. This field is subsequently
878 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
881 When CRL checking is enabled (i.e. the application sets the
882 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
883 pass arbitrary pointers to a memcmp call, enabling them to read memory
884 contents or enact a denial of service.
889 * Fixed NULL dereference validating DSA public key.
891 An invalid pointer dereference on read can be triggered when an
892 application tries to check a malformed DSA public key by the
893 EVP_PKEY_public_check() function. This will most likely lead
894 to an application crash. This function can be called on public
895 keys supplied from untrusted sources which could allow an attacker
896 to cause a denial of service attack.
898 The TLS implementation in OpenSSL does not call this function
899 but applications might call the function if there are additional
900 security requirements imposed by standards such as FIPS 140-3.
903 *Shane Lontis, Tomáš Mráz*
905 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
907 An invalid pointer dereference on read can be triggered when an
908 application tries to load malformed PKCS7 data with the
909 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
911 The result of the dereference is an application crash which could
912 lead to a denial of service attack. The TLS implementation in OpenSSL
913 does not call this function however third party applications might
914 call these functions on untrusted data.
919 * Fixed Use-after-free following BIO_new_NDEF.
921 The public API function BIO_new_NDEF is a helper function used for
922 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
923 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
924 be called directly by end user applications.
926 The function receives a BIO from the caller, prepends a new BIO_f_asn1
927 filter BIO onto the front of it to form a BIO chain, and then returns
928 the new head of the BIO chain to the caller. Under certain conditions,
929 for example if a CMS recipient public key is invalid, the new filter BIO
930 is freed and the function returns a NULL result indicating a failure.
931 However, in this case, the BIO chain is not properly cleaned up and the
932 BIO passed by the caller still retains internal pointers to the previously
933 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
934 then a use-after-free will occur. This will most likely result in a crash.
937 *Viktor Dukhovni, Matt Caswell*
939 * Fixed Double free after calling PEM_read_bio_ex.
941 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
942 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
943 data. If the function succeeds then the "name_out", "header" and "data"
944 arguments are populated with pointers to buffers containing the relevant
945 decoded data. The caller is responsible for freeing those buffers. It is
946 possible to construct a PEM file that results in 0 bytes of payload data.
947 In this case PEM_read_bio_ex() will return a failure code but will populate
948 the header argument with a pointer to a buffer that has already been freed.
949 If the caller also frees this buffer then a double free will occur. This
950 will most likely lead to a crash.
952 The functions PEM_read_bio() and PEM_read() are simple wrappers around
953 PEM_read_bio_ex() and therefore these functions are also directly affected.
955 These functions are also called indirectly by a number of other OpenSSL
956 functions including PEM_X509_INFO_read_bio_ex() and
957 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
958 internal uses of these functions are not vulnerable because the caller does
959 not free the header argument if PEM_read_bio_ex() returns a failure code.
962 *Kurt Roeckx, Matt Caswell*
964 * Fixed Timing Oracle in RSA Decryption.
966 A timing based side channel exists in the OpenSSL RSA Decryption
967 implementation which could be sufficient to recover a plaintext across
968 a network in a Bleichenbacher style attack. To achieve a successful
969 decryption an attacker would have to be able to send a very large number
970 of trial messages for decryption. The vulnerability affects all RSA padding
971 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
974 *Dmitry Belyavsky, Hubert Kario*
976 * Fixed X.509 Name Constraints Read Buffer Overflow.
978 A read buffer overrun can be triggered in X.509 certificate verification,
979 specifically in name constraint checking. The read buffer overrun might
980 result in a crash which could lead to a denial of service attack.
981 In a TLS client, this can be triggered by connecting to a malicious
982 server. In a TLS server, this can be triggered if the server requests
983 client authentication and a malicious client connects.
988 * Fixed X.509 Policy Constraints Double Locking security issue.
990 If an X.509 certificate contains a malformed policy constraint and
991 policy processing is enabled, then a write lock will be taken twice
992 recursively. On some operating systems (most widely: Windows) this
993 results in a denial of service when the affected process hangs. Policy
994 processing being enabled on a publicly facing server is not considered
995 to be a common setup.
1000 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1001 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1002 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1003 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1004 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1005 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1006 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1007 for legacy EC and SM2 keys is also changed similarly to honor the
1008 equivalent conversion format flag as specified in the underlying
1009 `EC_KEY` object being exported to a provider, when this function is
1010 called through `EVP_PKEY_export()`.
1014 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1016 * Fixed two buffer overflows in punycode decoding functions.
1018 A buffer overrun can be triggered in X.509 certificate verification,
1019 specifically in name constraint checking. Note that this occurs after
1020 certificate chain signature verification and requires either a CA to
1021 have signed the malicious certificate or for the application to continue
1022 certificate verification despite failure to construct a path to a trusted
1025 In a TLS client, this can be triggered by connecting to a malicious
1026 server. In a TLS server, this can be triggered if the server requests
1027 client authentication and a malicious client connects.
1029 An attacker can craft a malicious email address to overflow
1030 an arbitrary number of bytes containing the `.` character (decimal 46)
1031 on the stack. This buffer overflow could result in a crash (causing a
1035 An attacker can craft a malicious email address to overflow four
1036 attacker-controlled bytes on the stack. This buffer overflow could
1037 result in a crash (causing a denial of service) or potentially remote code
1038 execution depending on stack layout for any given platform/compiler.
1043 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1044 parameters in OpenSSL code.
1045 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1046 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1047 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1048 Using these invalid names may cause algorithms to use slower methods
1049 that ignore the CRT parameters.
1053 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1058 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1059 data to be signed before signing the certificate.
1063 * Added RIPEMD160 to the default provider.
1067 * Ensured that the key share group sent or accepted for the key exchange
1068 is allowed for the protocol version.
1072 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1074 * OpenSSL supports creating a custom cipher via the legacy
1075 EVP_CIPHER_meth_new() function and associated function calls. This function
1076 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1077 to use the new provider mechanism in order to implement custom ciphers.
1079 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1080 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1081 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1082 and decryption initialisation functions). Instead of using the custom cipher
1083 directly it incorrectly tries to fetch an equivalent cipher from the
1084 available providers. An equivalent cipher is found based on the NID passed to
1085 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1086 given cipher. However it is possible for an application to incorrectly pass
1087 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1088 is used in this way the OpenSSL encryption/decryption initialisation function
1089 will match the NULL cipher as being equivalent and will fetch this from the
1090 available providers. This will succeed if the default provider has been
1091 loaded (or if a third party provider has been loaded that offers this
1092 cipher). Using the NULL cipher means that the plaintext is emitted as the
1095 Applications are only affected by this issue if they call
1096 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1097 encryption/decryption initialisation function. Applications that only use
1098 SSL/TLS are not impacted by this issue.
1103 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1108 * Fixed the linux-mips64 Configure target which was missing the
1109 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1114 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1119 * Correctly handle a retransmitted ClientHello in DTLS
1123 * Fixed detection of ktls support in cross-compile environment on Linux
1127 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1132 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1133 report correct results in some cases
1137 * Fix UWP builds by defining VirtualLock
1141 * For known safe primes use the minimum key length according to RFC 7919.
1142 Longer private key sizes unnecessarily raise the cycles needed to compute the
1143 shared secret without any increase of the real security. This fixes a
1144 regression from 1.1.1 where these shorter keys were generated for the known
1149 * Added the loongarch64 target
1153 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1154 only passed to the FIPS provider and not to the default or legacy provider.
1158 * Fixed reported performance degradation on aarch64. Restored the
1159 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1160 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1161 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1162 The new algorithm is still used for 32 bit targets.
1166 * Added a missing header for memcmp that caused compilation failure on some
1171 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1173 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1174 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1175 This issue makes the RSA implementation with 2048 bit private keys
1176 incorrect on such machines and memory corruption will happen during
1177 the computation. As a consequence of the memory corruption an attacker
1178 may be able to trigger a remote code execution on the machine performing
1181 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1182 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1183 are affected by this issue.
1188 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1189 implementation would not encrypt the entirety of the data under some
1190 circumstances. This could reveal sixteen bytes of data that was
1191 preexisting in the memory that wasn't written. In the special case of
1192 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1194 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1195 they are both unaffected.
1198 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1200 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1202 * In addition to the c_rehash shell command injection identified in
1203 CVE-2022-1292, further bugs where the c_rehash script does not
1204 properly sanitise shell metacharacters to prevent command injection have been
1207 When the CVE-2022-1292 was fixed it was not discovered that there
1208 are other places in the script where the file names of certificates
1209 being hashed were possibly passed to a command executed through the shell.
1211 This script is distributed by some operating systems in a manner where
1212 it is automatically executed. On such operating systems, an attacker
1213 could execute arbitrary commands with the privileges of the script.
1215 Use of the c_rehash script is considered obsolete and should be replaced
1216 by the OpenSSL rehash command line tool.
1219 *Daniel Fiala, Tomáš Mráz*
1221 * Case insensitive string comparison no longer uses locales. It has instead
1222 been directly implemented.
1226 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1228 * Case insensitive string comparison is reimplemented via new locale-agnostic
1229 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1230 comparison. The previous implementation had problems when the Turkish locale
1235 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1236 metacharacters to prevent command injection. This script is distributed by
1237 some operating systems in a manner where it is automatically executed. On
1238 such operating systems, an attacker could execute arbitrary commands with the
1239 privileges of the script.
1241 Use of the c_rehash script is considered obsolete and should be replaced
1242 by the OpenSSL rehash command line tool.
1247 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1248 certificate on an OCSP response. The bug caused the function in the case
1249 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1250 response (meaning a successful verification) even in the case where the
1251 response signing certificate fails to verify.
1253 It is anticipated that most users of `OCSP_basic_verify` will not use the
1254 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1255 a negative value (indicating a fatal error) in the case of a certificate
1256 verification failure. The normal expected return value in this case would be
1259 This issue also impacts the command line OpenSSL "ocsp" application. When
1260 verifying an ocsp response with the "-no_cert_checks" option the command line
1261 application will report that the verification is successful even though it
1262 has in fact failed. In this case the incorrect successful response will also
1263 be accompanied by error messages showing the failure and contradicting the
1264 apparently successful result.
1269 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1270 AAD data as the MAC key. This made the MAC key trivially predictable.
1272 An attacker could exploit this issue by performing a man-in-the-middle attack
1273 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1274 that the modified data would still pass the MAC integrity check.
1276 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1277 endpoint will always be rejected by the recipient and the connection will
1278 fail at that point. Many application protocols require data to be sent from
1279 the client to the server first. Therefore, in such a case, only an OpenSSL
1280 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1282 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1283 sent in both directions. In this case both clients and servers could be
1284 affected, regardless of the application protocol.
1286 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1287 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1288 the handshake when using this ciphersuite.
1290 The confidentiality of data is not impacted by this issue, i.e. an attacker
1291 cannot decrypt data that has been encrypted using this ciphersuite - they can
1294 In order for this attack to work both endpoints must legitimately negotiate
1295 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1296 OpenSSL 3.0, and is not available within the default provider or the default
1297 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1298 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1299 following must have occurred:
1301 1) OpenSSL must have been compiled with the (non-default) compile time option
1302 enable-weak-ssl-ciphers
1304 2) OpenSSL must have had the legacy provider explicitly loaded (either
1305 through application code or via configuration)
1307 3) The ciphersuite must have been explicitly added to the ciphersuite list
1309 4) The libssl security level must have been set to 0 (default is 1)
1311 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1313 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1314 others that both endpoints have in common
1319 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1320 occupied by the removed hash table entries.
1322 This function is used when decoding certificates or keys. If a long lived
1323 process periodically decodes certificates or keys its memory usage will
1324 expand without bounds and the process might be terminated by the operating
1325 system causing a denial of service. Also traversing the empty hash table
1326 entries will take increasingly more time.
1328 Typically such long lived processes might be TLS clients or TLS servers
1329 configured to accept client certificate authentication.
1332 *Hugo Landau, Aliaksei Levin*
1334 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1335 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1336 statistics are no longer supported. For compatibility, these statistics are
1337 still listed in the output but are now always reported as zero.
1341 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1343 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1344 for non-prime moduli.
1346 Internally this function is used when parsing certificates that contain
1347 elliptic curve public keys in compressed form or explicit elliptic curve
1348 parameters with a base point encoded in compressed form.
1350 It is possible to trigger the infinite loop by crafting a certificate that
1351 has invalid explicit curve parameters.
1353 Since certificate parsing happens prior to verification of the certificate
1354 signature, any process that parses an externally supplied certificate may thus
1355 be subject to a denial of service attack. The infinite loop can also be
1356 reached when parsing crafted private keys as they can contain explicit
1357 elliptic curve parameters.
1359 Thus vulnerable situations include:
1361 - TLS clients consuming server certificates
1362 - TLS servers consuming client certificates
1363 - Hosting providers taking certificates or private keys from customers
1364 - Certificate authorities parsing certification requests from subscribers
1365 - Anything else which parses ASN.1 elliptic curve parameters
1367 Also any other applications that use the BN_mod_sqrt() where the attacker
1368 can control the parameter values are vulnerable to this DoS issue.
1373 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1374 to the list of ciphersuites providing Perfect Forward Secrecy as
1375 required by SECLEVEL >= 3.
1377 *Dmitry Belyavskiy, Nicola Tuveri*
1379 * Made the AES constant time code for no-asm configurations
1380 optional due to the resulting 95% performance degradation.
1381 The AES constant time code can be enabled, for no assembly
1382 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1386 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1391 * The negative return value handling of the certificate verification callback
1392 was reverted. The replacement is to set the verification retry state with
1393 the SSL_set_retry_verify() function.
1397 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1399 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1400 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1401 verify a certificate supplied by a server. That function may return a
1402 negative return value to indicate an internal error (for example out of
1403 memory). Such a negative return value is mishandled by OpenSSL and will cause
1404 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1405 success and a subsequent call to SSL_get_error() to return the value
1406 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1407 returned by OpenSSL if the application has previously called
1408 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1409 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1410 totally unexpected and applications may not behave correctly as a result. The
1411 exact behaviour will depend on the application but it could result in
1412 crashes, infinite loops or other similar incorrect responses.
1414 This issue is made more serious in combination with a separate bug in OpenSSL
1415 3.0 that will cause X509_verify_cert() to indicate an internal error when
1416 processing a certificate chain. This will occur where a certificate does not
1417 include the Subject Alternative Name extension but where a Certificate
1418 Authority has enforced name constraints. This issue can occur even with valid
1424 * Corrected a few file name and file reference bugs in the build,
1425 installation and setup scripts, which lead to installation verification
1426 failures. Slightly enhanced the installation verification script.
1430 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1435 * Fixed PVK encoder to properly query for the passphrase.
1439 * Multiple fixes in the OSSL_HTTP API functions.
1443 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1444 OSSL_PARAM_INTEGER data type and return error on negative numbers
1445 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1446 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1450 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1454 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1458 * Multiple threading fixes.
1462 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1466 * Allow fetching an operation from the provider that owns an unexportable key
1467 as a fallback if that is still allowed by the property query.
1471 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1473 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1478 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1479 S390X capability vector to zero. This simplifies testing of different code
1480 paths on S390X architecture.
1484 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1485 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1486 SP 800-38D". The communication will fail at this point.
1490 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1491 confidential in EC_GROUP data.
1495 * The byte order mark (BOM) character is ignored if encountered at the
1496 beginning of a PEM-formatted file.
1500 * Added CMS support for the Russian GOST algorithms.
1504 * Due to move of the implementation of cryptographic operations
1505 to the providers, validation of various operation parameters can
1506 be postponed until the actual operation is executed where previously
1507 it happened immediately when an operation parameter was set.
1509 For example when setting an unsupported curve with
1510 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1511 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1513 *OpenSSL team members and many third party contributors*
1515 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1516 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1517 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1518 instead to retrieve these algorithms from a provider.
1522 * On build targets where the multilib postfix is set in the build
1523 configuration the libdir directory was changing based on whether
1524 the lib directory with the multilib postfix exists on the system
1525 or not. This unpredictable behavior was removed and eventual
1526 multilib postfix is now always added to the default libdir. Use
1527 `--libdir=lib` to override the libdir if adding the postfix is
1532 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1533 no longer interoperable with OpenSSL 1.1.1.
1537 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1538 function codes, this function can only cause problems for calling
1543 * Add a configurable flag to output date formats as ISO 8601. Does not
1544 change the default date format.
1548 * Version of MSVC earlier than 1300 could get link warnings, which could
1549 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1550 Support for this flag has been removed.
1554 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1555 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1556 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1557 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1558 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1562 * The signatures of the functions to get and set options on SSL and
1563 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1564 Some source code changes may be required.
1568 * The public definitions of conf_method_st and conf_st have been
1569 deprecated. They will be made opaque in a future release.
1571 *Rich Salz and Tomáš Mráz*
1573 * Client-initiated renegotiation is disabled by default. To allow it, use
1574 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1575 flag, or the "ClientRenegotiation" config parameter as appropriate.
1579 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1580 or modify relative pathname inclusion.
1584 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1585 validated. Please consult the README-FIPS and
1586 README-PROVIDERS files, as well as the migration guide.
1588 *OpenSSL team members and many third party contributors*
1590 * For the key types DH and DHX the allowed settable parameters are now different.
1594 * The openssl commands that read keys, certificates, and CRLs now
1595 automatically detect the PEM or DER format of the input files.
1597 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1599 * Added enhanced PKCS#12 APIs which accept a library context.
1603 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1607 * Added support for Kernel TLS (KTLS).
1609 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1611 * Support for RFC 5746 secure renegotiation is now required by default for
1612 SSL or TLS connections to succeed.
1616 * The signature of the `copy` functional parameter of the
1617 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1618 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1619 the signature of the `pub_decode` functional parameter of the
1620 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1621 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1625 * The error return values from some control calls (ctrl) have changed.
1629 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1633 * Many functions in the EVP_ namespace that are getters of values from
1634 implementations or contexts were renamed to include get or get0 in their
1635 names. Old names are provided as macro aliases for compatibility and
1640 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1641 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1642 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1647 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1650 * The output from the command line applications may have minor
1655 * The output from numerous "printing" may have minor changes.
1659 * Windows thread synchronization uses read/write primitives (SRWLock) when
1660 supported by the OS, otherwise CriticalSection continues to be used.
1664 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1665 work on read only BIO source/sinks that do not support these functions.
1666 This allows piping or redirection of a file BIO using stdin to be buffered
1667 into memory. This is used internally in OSSL_DECODER_from_bio().
1671 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1672 this function would return one of the values OSSL_STORE_INFO_NAME,
1673 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1674 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1675 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1676 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1677 using this function should be amended to handle the changed return value.
1681 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1682 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1683 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1684 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1685 contains more than one certificate identifier: This means that all
1686 certificates referenced there MUST be part of the validation chain.
1690 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1691 RC5, DESX and DES have been moved to the legacy provider.
1695 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1696 RIPEMD-160 have been moved to the legacy provider.
1700 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1705 * The deprecated functions EVP_PKEY_get0_RSA(),
1706 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1707 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1708 well as the similarly named "get1" functions behave differently in
1713 * A number of functions handling low-level keys or engines were deprecated
1714 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1715 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1716 EVP_PKEY_get0_siphash().
1720 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1721 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1722 will need to load the legacy crypto provider. This includes these PBE
1723 algorithms which use this KDF:
1724 - NID_pbeWithMD2AndDES_CBC
1725 - NID_pbeWithMD5AndDES_CBC
1726 - NID_pbeWithSHA1AndRC2_CBC
1727 - NID_pbeWithMD2AndRC2_CBC
1728 - NID_pbeWithMD5AndRC2_CBC
1729 - NID_pbeWithSHA1AndDES_CBC
1733 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1734 BIO_debug_callback() functions.
1738 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1739 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1743 * The RAND_METHOD APIs have been deprecated.
1747 * The SRP APIs have been deprecated.
1751 * Add a compile time option to prevent the caching of provider fetched
1752 algorithms. This is enabled by including the no-cached-fetch option
1753 at configuration time.
1757 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1758 count of PKCS12_DEFAULT_ITER.
1760 *Tomáš Mráz and Sahana Prasad*
1762 * The openssl speed command does not use low-level API calls anymore.
1766 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1769 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1771 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1775 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1776 providers may supply their own group implementations (using either the "key
1777 exchange" or the "key encapsulation" methods) which will automatically be
1778 detected and used by libssl.
1780 *Matt Caswell, Nicola Tuveri*
1782 * The undocumented function X509_certificate_type() has been deprecated;
1786 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1790 * Removed RSA padding mode for SSLv23 (which was only used for
1791 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1792 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1797 * Deprecated the obsolete X9.31 RSA key generation related functions.
1799 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1800 is not allowed to return a value > 1, this is no more taken as failure.
1802 *Viktor Dukhovni and David von Oheimb*
1804 * Deprecated the obsolete X9.31 RSA key generation related functions
1805 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1806 BN_X931_generate_prime_ex().
1810 * The default key generation method for the regular 2-prime RSA keys was
1811 changed to the FIPS 186-4 B.3.6 method.
1815 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1819 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1823 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1824 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1826 *Rich Salz, Richard Levitte, and David von Oheimb*
1828 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1832 * Deprecated `OCSP_parse_url()`.
1836 * Validation of SM2 keys has been separated from the validation of regular EC
1841 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1842 switches: a validation failure triggers an early exit, returning a failure
1843 exit status to the parent process.
1847 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1848 to ignore unknown ciphers.
1852 * The `-cipher-commands` and `-digest-commands` options
1853 of the command line utility `list` have been deprecated.
1854 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1858 * Added convenience functions for generating asymmetric key pairs:
1859 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1860 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1864 * All of the low-level EC_KEY functions have been deprecated.
1866 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1868 * Deprecated all the libcrypto and libssl error string loading
1873 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1874 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1879 * The `-crypt` option to the `passwd` command line tool has been removed.
1883 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1888 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1892 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1893 EVP_PKEY_get1_tls_encodedpoint().
1897 * The security callback, which can be customised by application code, supports
1898 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1899 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1903 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1904 interface. Their functionality remains unchanged.
1908 * Added new option for 'openssl list', '-providers', which will display the
1909 list of loaded providers, their names, version and status. It optionally
1910 displays their gettable parameters.
1914 * Removed EVP_PKEY_set_alias_type().
1918 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1919 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1923 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1924 parameters are still checked for type safety at compile time via helper
1929 * Remove the RAND_DRBG API
1931 *Paul Dale and Matthias St. Pierre*
1933 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1934 as well as actual hostnames.
1938 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1939 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1940 conversely, silently ignore DTLS protocol version bounds when configuring
1941 TLS-based contexts. The commands can be repeated to set bounds of both
1942 types. The same applies with the corresponding "min_protocol" and
1943 "max_protocol" command-line switches, in case some application uses both TLS
1946 SSL_CTX instances that are created for a fixed protocol version (e.g.
1947 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1948 attempts to apply bounds to these protocol versions would result in an
1949 error. Now only the "version-flexible" SSL_CTX instances are subject to
1950 limits in configuration files in command-line options.
1954 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1959 * Reworked the recorded ERR codes to make better space for system errors.
1960 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1961 given code is a system error (true) or an OpenSSL error (false).
1965 * Reworked the test perl framework to better allow parallel testing.
1967 *Nicola Tuveri and David von Oheimb*
1969 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1970 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1974 * 'Configure' has been changed to figure out the configuration target if
1975 none is given on the command line. Consequently, the 'config' script is
1976 now only a mere wrapper. All documentation is changed to only mention
1979 *Rich Salz and Richard Levitte*
1981 * Added a library context `OSSL_LIB_CTX` that applications as well as
1982 other libraries can use to form a separate context within which
1983 libcrypto operations are performed.
1987 * Added various `_ex` functions to the OpenSSL API that support using
1988 a non-default `OSSL_LIB_CTX`.
1992 * Handshake now fails if Extended Master Secret extension is dropped
1997 * Dropped interactive mode from the `openssl` program.
2001 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2003 *David von Oheimb and Shane Lontis*
2005 * Deprecated `EC_METHOD_get_field_type()`.
2009 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2010 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2011 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2015 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2019 * Add CAdES-BES signature verification support, mostly derived
2020 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2022 *Filipe Raimundo da Silva*
2024 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2028 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2029 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2033 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2037 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2038 EC_KEY_precompute_mult().
2042 * Deprecated EC_POINTs_mul().
2046 * Removed FIPS_mode() and FIPS_mode_set().
2050 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2054 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2055 EC_POINT_get_Jprojective_coordinates_GFp().
2059 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2060 arrays to be more easily constructed via a series of utility functions.
2061 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2062 the various push functions and finally convert to a passable OSSL_PARAM
2063 array using OSSL_PARAM_BLD_to_param().
2067 * The security strength of SHA1 and MD5 based signatures in TLS has been
2072 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2073 contain a provider side internal key.
2077 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2081 * Project text documents not yet having a proper file name extension
2082 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2083 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2087 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2088 have been converted to Markdown with the goal to produce documents
2089 which not only look pretty when viewed online in the browser, but
2090 remain well readable inside a plain text editor.
2092 To achieve this goal, a 'minimalistic' Markdown style has been applied
2093 which avoids formatting elements that interfere too much with the
2094 reading flow in the text file. For example, it
2096 * avoids [ATX headings][] and uses [setext headings][] instead
2097 (which works for `<h1>` and `<h2>` headings only).
2098 * avoids [inline links][] and uses [reference links][] instead.
2099 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2101 [ATX headings]: https://github.github.com/gfm/#atx-headings
2102 [setext headings]: https://github.github.com/gfm/#setext-headings
2103 [inline links]: https://github.github.com/gfm/#inline-link
2104 [reference links]: https://github.github.com/gfm/#reference-link
2105 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2106 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2108 *Matthias St. Pierre*
2110 * The test suite is changed to preserve results of each test recipe.
2111 A new directory test-runs/ with subdirectories named like the
2112 test recipes are created in the build tree for this purpose.
2116 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2117 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2118 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2120 *David von Oheimb, Martin Peylo*
2122 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2123 It supports arbitrary request and response content types, GET redirection,
2124 TLS, connections via HTTP(S) proxies, connections and exchange via
2125 user-defined BIOs (allowing implicit connections), persistent connections,
2126 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2127 The legacy OCSP-focused (and only partly documented) API
2128 is retained for backward compatibility, while most of it is deprecated.
2132 * Added `util/check-format.pl`, a tool for checking adherence to the
2133 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2134 The checks performed are incomplete and yield some false positives.
2135 Still the tool should be useful for detecting most typical glitches.
2139 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2140 If domain name resolution yields multiple IP addresses all of them are tried
2141 after `connect()` failures.
2145 * All of the low-level RSA functions have been deprecated.
2149 * X509 certificates signed using SHA1 are no longer allowed at security
2154 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2155 modified to use PKEY APIs. These commands are now in maintenance mode
2156 and no new features will be added to them.
2160 * The command line utility rsautl has been deprecated.
2164 * The command line utilities genrsa and rsa have been modified to use PKEY
2165 APIs. They now write PKCS#8 keys by default. These commands are now in
2166 maintenance mode and no new features will be added to them.
2170 * All of the low-level DH functions have been deprecated.
2172 *Paul Dale and Matt Caswell*
2174 * All of the low-level DSA functions have been deprecated.
2178 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2179 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2183 * Deprecated low-level ECDH and ECDSA functions.
2187 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2191 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2192 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2193 a new formulation to include all the things it can be used for,
2194 as well as words of caution.
2198 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2202 * All of the low-level HMAC functions have been deprecated.
2204 *Paul Dale and David von Oheimb*
2206 * Over two thousand fixes were made to the documentation, including:
2207 - Common options (such as -rand/-writerand, TLS version control, etc)
2208 were refactored and point to newly-enhanced descriptions in openssl.pod.
2209 - Added style conformance for all options (with help from Richard Levitte),
2210 documented all reported missing options, added a CI build to check
2211 that all options are documented and that no unimplemented options
2213 - Documented some internals, such as all use of environment variables.
2214 - Addressed all internal broken L<> references.
2218 * All of the low-level CMAC functions have been deprecated.
2222 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2223 functions have been deprecated.
2225 *Paul Dale and David von Oheimb*
2227 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2228 set of functions. The documentation mentioned negative values for some
2229 errors, but this was never the case, so the mention of negative values
2232 Code that followed the documentation and thereby check with something
2233 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2237 * All of the low-level cipher functions have been deprecated.
2239 *Matt Caswell and Paul Dale*
2241 * Removed include/openssl/opensslconf.h.in and replaced it with
2242 include/openssl/configuration.h.in, which differs in not including
2243 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2244 was added to include both.
2246 This allows internal hacks where one might need to modify the set
2247 of configured macros, for example this if deprecated symbols are
2248 still supposed to be available internally:
2250 #include <openssl/configuration.h>
2252 #undef OPENSSL_NO_DEPRECATED
2253 #define OPENSSL_SUPPRESS_DEPRECATED
2255 #include <openssl/macros.h>
2257 This should not be used by applications that use the exported
2258 symbols, as that will lead to linking errors.
2262 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2263 used in exponentiation with 512-bit moduli. No EC algorithms are
2264 affected. Analysis suggests that attacks against 2-prime RSA1024,
2265 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2266 difficult to perform and are not believed likely. Attacks against DH512
2267 are considered just feasible. However, for an attack the target would
2268 have to reuse the DH512 private key, which is not recommended anyway.
2269 Also applications directly using the low-level API BN_mod_exp may be
2270 affected if they use BN_FLG_CONSTTIME.
2275 * Most memory-debug features have been deprecated, and the functionality
2276 replaced with no-ops.
2280 * Added documentation for the STACK API.
2284 * Introduced a new method type and API, OSSL_ENCODER, to represent
2285 generic encoders. These do the same sort of job that PEM writers
2286 and d2i functions do, but with support for methods supplied by
2287 providers, and the possibility for providers to support other
2292 * Introduced a new method type and API, OSSL_DECODER, to represent
2293 generic decoders. These do the same sort of job that PEM readers
2294 and i2d functions do, but with support for methods supplied by
2295 providers, and the possibility for providers to support other
2300 * Added a .pragma directive to the syntax of configuration files, to
2301 allow varying behavior in a supported and predictable manner.
2302 Currently added pragma:
2306 This allows dollar signs to be a keyword character unless it's
2307 followed by a opening brace or parenthesis. This is useful for
2308 platforms where dollar signs are commonly used in names, such as
2309 volume names and system directory names on VMS.
2313 * Added functionality to create an EVP_PKEY from user data.
2317 * Change the interpretation of the '--api' configuration option to
2318 mean that this is a desired API compatibility level with no
2319 further meaning. The previous interpretation, that this would
2320 also mean to remove all deprecated symbols up to and including
2321 the given version, no requires that 'no-deprecated' is also used
2322 in the configuration.
2324 When building applications, the desired API compatibility level
2325 can be set with the OPENSSL_API_COMPAT macro like before. For
2326 API compatibility version below 3.0, the old style numerical
2327 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2328 For version 3.0 and on, the value is expected to be the decimal
2329 value calculated from the major and minor version like this:
2331 MAJOR * 10000 + MINOR * 100
2335 -DOPENSSL_API_COMPAT=30000 For 3.0
2336 -DOPENSSL_API_COMPAT=30200 For 3.2
2338 To hide declarations that are deprecated up to and including the
2339 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2340 given when building the application as well.
2344 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2345 access to certificate and CRL stores via URIs and OSSL_STORE
2348 This adds the following functions:
2350 - X509_LOOKUP_store()
2351 - X509_STORE_load_file()
2352 - X509_STORE_load_path()
2353 - X509_STORE_load_store()
2354 - SSL_add_store_cert_subjects_to_stack()
2355 - SSL_CTX_set_default_verify_store()
2356 - SSL_CTX_load_verify_file()
2357 - SSL_CTX_load_verify_dir()
2358 - SSL_CTX_load_verify_store()
2362 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2363 The presence of this system service is determined at run-time.
2367 * Added functionality to create an EVP_PKEY context based on data
2368 for methods from providers. This takes an algorithm name and a
2369 property query string and simply stores them, with the intent
2370 that any operation that uses this context will use those strings
2371 to fetch the needed methods implicitly, thereby making the port
2372 of application written for pre-3.0 OpenSSL easier.
2376 * The undocumented function NCONF_WIN32() has been deprecated; for
2377 conversion details see the HISTORY section of doc/man5/config.pod
2381 * Introduced the new functions EVP_DigestSignInit_ex() and
2382 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2383 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2384 pages for further details.
2388 * Over two thousand fixes were made to the documentation, including:
2389 adding missing command flags, better style conformance, documentation
2392 *Rich Salz, Richard Levitte*
2394 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2395 X25519, X448, Ed25519 and Ed448.
2399 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2404 * Deprecated the public definition of `ERR_STATE` as well as the function
2405 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2410 * Added ERR functionality to give callers access to the stored function
2411 names that have replaced the older function code based functions.
2413 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2414 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2415 ERR_peek_error_all() and ERR_peek_last_error_all().
2417 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2418 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2419 ERR_func_error_string().
2423 * Extended testing to be verbose for failing tests only. The make variables
2424 VERBOSE_FAILURE or VF can be used to enable this:
2426 $ make VF=1 test # Unix
2427 $ mms /macro=(VF=1) test ! OpenVMS
2428 $ nmake VF=1 test # Windows
2432 * Added the `-copy_extensions` option to the `x509` command for use with
2433 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2434 all extensions in the request are copied to the certificate or vice versa.
2436 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2438 * Added the `-copy_extensions` option to the `req` command for use with
2439 `-x509`. When given with the `copy` or `copyall` argument,
2440 all extensions in the certification request are copied to the certificate.
2444 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2445 they generate are by default RFC 5280 compliant in the following sense:
2446 There is a subjectKeyIdentifier extension with a hash value of the public key
2447 and for not self-signed certs there is an authorityKeyIdentifier extension
2448 with a keyIdentifier field or issuer information identifying the signing key.
2449 This is done unless some configuration overrides the new default behavior,
2450 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2454 * Added several checks to `X509_verify_cert()` according to requirements in
2455 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2456 (which may be done by using the CLI option `-x509_strict`):
2457 * The basicConstraints of CA certificates must be marked critical.
2458 * CA certificates must explicitly include the keyUsage extension.
2459 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2460 * The issuer name of any certificate must not be empty.
2461 * The subject name of CA certs, certs with keyUsage crlSign,
2462 and certs without subjectAlternativeName must not be empty.
2463 * If a subjectAlternativeName extension is given it must not be empty.
2464 * The signatureAlgorithm field and the cert signature must be consistent.
2465 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2466 must not be marked critical.
2467 * The authorityKeyIdentifier must be given for X.509v3 certs
2468 unless they are self-signed.
2469 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2473 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2474 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2478 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2479 used even when parsing explicit parameters, when loading a encoded key
2480 or calling `EC_GROUP_new_from_ecpkparameters()`/
2481 `EC_GROUP_new_from_ecparameters()`.
2482 This prevents bypass of security hardening and performance gains,
2483 especially for curves with specialized EC_METHODs.
2484 By default, if a key encoded with explicit parameters is loaded and later
2485 encoded, the output is still encoded with explicit parameters, even if
2486 internally a "named" EC_GROUP is used for computation.
2490 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2491 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2492 NULL. After this change, only the cofactor parameter can be NULL. It also
2493 does some minimal sanity checks on the passed order.
2498 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2499 An attack is simple, if the first CMS_recipientInfo is valid but the
2500 second CMS_recipientInfo is chosen ciphertext. If the second
2501 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2502 encryption key will be replaced by garbage, and the message cannot be
2503 decoded, but if the RSA decryption fails, the correct encryption key is
2504 used and the recipient will not notice the attack.
2505 As a work around for this potential attack the length of the decrypted
2506 key must be equal to the cipher default key length, in case the
2507 certificate is not given and all recipientInfo are tried out.
2508 The old behaviour can be re-enabled in the CMS code by setting the
2509 CMS_DEBUG_DECRYPT flag.
2513 * Early start up entropy quality from the DEVRANDOM seed source has been
2514 improved for older Linux systems. The RAND subsystem will wait for
2515 /dev/random to be producing output before seeding from /dev/urandom.
2516 The seeded state is stored for future library initialisations using
2517 a system global shared memory segment. The shared memory identifier
2518 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2519 the desired value. The default identifier is 114.
2523 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2524 when primes for RSA keys are computed.
2525 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2526 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2527 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2528 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2529 This avoids possible fingerprinting of newly generated RSA modules.
2533 * Correct the extended master secret constant on EBCDIC systems. Without this
2534 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2535 negotiate EMS will fail. Unfortunately this also means that TLS connections
2536 between EBCDIC systems with this fix, and EBCDIC systems without this
2537 fix will fail if they negotiate EMS.
2541 * Changed the library initialisation so that the config file is now loaded
2542 by default. This was already the case for libssl. It now occurs for both
2543 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2544 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2548 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2549 where the former acts as a replacement for `ERR_put_error()`, and the
2550 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2551 `ERR_raise_data()` adds more flexibility by taking a format string and
2552 an arbitrary number of arguments following it, to be processed with
2557 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2558 to check if a named provider is loaded and available. When called, it
2559 will also activate all fallback providers if such are still present.
2563 * Enforce a minimum DH modulus size of 512 bits.
2567 * Changed DH parameters to generate the order q subgroup instead of 2q.
2568 Previously generated DH parameters are still accepted by DH_check
2569 but DH_generate_key works around that by clearing bit 0 of the
2570 private key for those. This avoids leaking bit 0 of the private key.
2574 * Significantly reduce secure memory usage by the randomness pools.
2578 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2583 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2584 algorithms. An implementation of a key exchange algorithm can be obtained
2585 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2586 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2587 the older EVP_PKEY_derive_init() function. See the man pages for the new
2588 functions for further details.
2592 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2596 * Removed the function names from error messages and deprecated the
2601 * Removed NextStep support and the macro OPENSSL_UNISTD
2605 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2606 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2607 Also removed "export var as function" capability; we do not export
2608 variables, only functions.
2612 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2613 an error and 1 indicating success. In previous versions of OpenSSL this
2614 was a void type. If a key was set longer than the maximum possible this
2619 * Support SM2 signing and verification schemes with X509 certificate.
2623 * Use SHA256 as the default digest for TS query in the `ts` app.
2627 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2631 * Default cipher lists/suites are now available via a function, the
2632 #defines are deprecated.
2636 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2637 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2638 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2642 * Join the directories crypto/x509 and crypto/x509v3
2646 * Added command 'openssl kdf' that uses the EVP_KDF API.
2650 * Added command 'openssl mac' that uses the EVP_MAC API.
2654 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2655 as default directories. Also added the command 'openssl info'
2656 for scripting purposes.
2660 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2665 * Add prediction resistance to the DRBG reseeding process.
2669 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2670 mandated by IEEE Std 1619-2018.
2674 * Added newline escaping functionality to a filename when using openssl dgst.
2675 This output format is to replicate the output format found in the `*sum`
2676 checksum programs. This aims to preserve backward compatibility.
2678 *Matt Eaton, Richard Levitte, and Paul Dale*
2680 * Removed the heartbeat message in DTLS feature, as it has very
2681 little usage and doesn't seem to fulfill a valuable purpose.
2682 The configuration option is now deprecated.
2686 * Changed the output of 'openssl {digestname} < file' to display the
2687 digest name in its output.
2691 * Added a new generic trace API which provides support for enabling
2692 instrumentation through trace output.
2694 *Richard Levitte & Matthias St. Pierre*
2696 * Added build tests for C++. These are generated files that only do one
2697 thing, to include one public OpenSSL head file each. This tests that
2698 the public header files can be usefully included in a C++ application.
2700 This test isn't enabled by default. It can be enabled with the option
2701 'enable-buildtest-c++'.
2705 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2709 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2713 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2717 * Added KMAC to EVP_MAC.
2721 * Added property based algorithm implementation selection framework to
2726 * Added SCA hardening for modular field inversion in EC_GROUP through
2727 a new dedicated field_inv() pointer in EC_METHOD.
2728 This also addresses a leakage affecting conversions from projective
2729 to affine coordinates.
2731 *Billy Bob Brumley, Nicola Tuveri*
2733 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2734 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2735 those algorithms that were already supported through the EVP_PKEY API
2736 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2737 and scrypt are now wrappers that call EVP_KDF.
2741 * Build devcrypto engine as a dynamic engine.
2743 *Eneas U de Queiroz*
2745 * Add keyed BLAKE2 to EVP_MAC.
2749 * Fix a bug in the computation of the endpoint-pair shared secret used
2750 by DTLS over SCTP. This breaks interoperability with older versions
2751 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2752 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2753 interoperability with such broken implementations. However, enabling
2754 this switch breaks interoperability with correct implementations.
2756 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2757 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2761 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2765 * Changed the license to the Apache License v2.0.
2769 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2771 - Major releases (indicated by incrementing the MAJOR release number)
2772 may introduce incompatible API/ABI changes.
2773 - Minor releases (indicated by incrementing the MINOR release number)
2774 may introduce new features but retain API/ABI compatibility.
2775 - Patch releases (indicated by incrementing the PATCH number)
2776 are intended for bug fixes and other improvements of existing
2777 features only (like improving performance or adding documentation)
2778 and retain API/ABI compatibility.
2782 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2786 * Remove the 'dist' target and add a tarball building script. The
2787 'dist' target has fallen out of use, and it shouldn't be
2788 necessary to configure just to create a source distribution.
2792 * Recreate the OS390-Unix config target. It no longer relies on a
2793 special script like it did for OpenSSL pre-1.1.0.
2797 * Instead of having the source directories listed in Configure, add
2798 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2803 * Add GMAC to EVP_MAC.
2807 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2811 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2812 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2813 to facilitate the continued use of MACs through raw private keys in
2814 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2818 * Deprecate ECDH_KDF_X9_62().
2822 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2823 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2824 are retained for backwards compatibility.
2828 * AES-XTS mode now enforces that its two keys are different to mitigate
2829 the attacked described in "Efficient Instantiations of Tweakable
2830 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2831 Details of this attack can be obtained from:
2832 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2836 * Rename the object files, i.e. give them other names than in previous
2837 versions. Their names now include the name of the final product, as
2838 well as its type mnemonic (bin, lib, shlib).
2842 * Added new option for 'openssl list', '-objects', which will display the
2843 list of built in objects, i.e. OIDs with names.
2847 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2848 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2853 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2854 improves application performance by removing data copies and providing
2855 applications with zero-copy system calls such as sendfile and splice.
2859 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2863 * `PKCS12_parse` now maintains the order of the parsed certificates
2864 when outputting them via `*ca` (rather than reversing it).
2868 * Deprecated pthread fork support methods.
2872 * Added support for FFDHE key exchange in TLS 1.3.
2876 * Added a new concept for OpenSSL plugability: providers. This
2877 functionality is designed to replace the ENGINE API and ENGINE
2878 implementations, and to be much more dynamic, allowing provider
2879 authors to introduce new algorithms among other things, as long as
2880 there's an API that supports the algorithm type.
2882 With this concept comes a new core API for interaction between
2883 libcrypto and provider implementations. Public libcrypto functions
2884 that want to use providers do so through this core API.
2886 The main documentation for this core API is found in
2887 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2888 refer to other manuals describing the API specific for supported
2889 algorithm types (also called operations).
2896 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2898 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2900 * Avoid loading of a dynamic engine twice.
2904 * Prioritise DANE TLSA issuer certs over peer certs
2908 * Fixed random API for MacOS prior to 10.12
2910 These MacOS versions don't support the CommonCrypto APIs
2914 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2916 * Fixed an SM2 Decryption Buffer Overflow.
2918 In order to decrypt SM2 encrypted data an application is expected to
2919 call the API function EVP_PKEY_decrypt(). Typically an application will
2920 call this function twice. The first time, on entry, the "out" parameter
2921 can be NULL and, on exit, the "outlen" parameter is populated with the
2922 buffer size required to hold the decrypted plaintext. The application
2923 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2924 again, but this time passing a non-NULL value for the "out" parameter.
2926 A bug in the implementation of the SM2 decryption code means that the
2927 calculation of the buffer size required to hold the plaintext returned
2928 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2929 size required by the second call. This can lead to a buffer overflow
2930 when EVP_PKEY_decrypt() is called by the application a second time with
2931 a buffer that is too small.
2933 A malicious attacker who is able present SM2 content for decryption to
2934 an application could cause attacker chosen data to overflow the buffer
2935 by up to a maximum of 62 bytes altering the contents of other data held
2936 after the buffer, possibly changing application behaviour or causing
2937 the application to crash. The location of the buffer is application
2938 dependent but is typically heap allocated.
2943 * Fixed various read buffer overruns processing ASN.1 strings
2945 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2946 structure which contains a buffer holding the string data and a field
2947 holding the buffer length. This contrasts with normal C strings which
2948 are represented as a buffer for the string data which is terminated
2949 with a NUL (0) byte.
2951 Although not a strict requirement, ASN.1 strings that are parsed using
2952 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2953 well as any string whose value has been set with the ASN1_STRING_set()
2954 function will additionally NUL terminate the byte array in the
2955 ASN1_STRING structure.
2957 However, it is possible for applications to directly construct valid
2958 ASN1_STRING structures which do not NUL terminate the byte array by
2959 directly setting the "data" and "length" fields in the ASN1_STRING
2960 array. This can also happen by using the ASN1_STRING_set0() function.
2962 Numerous OpenSSL functions that print ASN.1 data have been found to
2963 assume that the ASN1_STRING byte array will be NUL terminated, even
2964 though this is not guaranteed for strings that have been directly
2965 constructed. Where an application requests an ASN.1 structure to be
2966 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2967 been directly constructed by the application without NUL terminating
2968 the "data" field, then a read buffer overrun can occur.
2970 The same thing can also occur during name constraints processing
2971 of certificates (for example if a certificate has been directly
2972 constructed by the application instead of loading it via the OpenSSL
2973 parsing functions, and the certificate contains non NUL terminated
2974 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2975 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2977 If a malicious actor can cause an application to directly construct an
2978 ASN1_STRING and then process it through one of the affected OpenSSL
2979 functions then this issue could be hit. This might result in a crash
2980 (causing a Denial of Service attack). It could also result in the
2981 disclosure of private memory contents (such as private keys, or
2982 sensitive plaintext).
2987 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2989 * Fixed a problem with verifying a certificate chain when using the
2990 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2991 the certificates present in a certificate chain. It is not set by default.
2993 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2994 the chain that have explicitly encoded elliptic curve parameters was added
2995 as an additional strict check.
2997 An error in the implementation of this check meant that the result of a
2998 previous check to confirm that certificates in the chain are valid CA
2999 certificates was overwritten. This effectively bypasses the check
3000 that non-CA certificates must not be able to issue other certificates.
3002 If a "purpose" has been configured then there is a subsequent opportunity
3003 for checks that the certificate is a valid CA. All of the named "purpose"
3004 values implemented in libcrypto perform this check. Therefore, where
3005 a purpose is set the certificate chain will still be rejected even when the
3006 strict flag has been used. A purpose is set by default in libssl client and
3007 server certificate verification routines, but it can be overridden or
3008 removed by an application.
3010 In order to be affected, an application must explicitly set the
3011 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3012 for the certificate verification or, in the case of TLS client or server
3013 applications, override the default purpose.
3018 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3019 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3020 renegotiation ClientHello omits the signature_algorithms extension (where it
3021 was present in the initial ClientHello), but includes a
3022 signature_algorithms_cert extension then a NULL pointer dereference will
3023 result, leading to a crash and a denial of service attack.
3025 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3026 (which is the default configuration). OpenSSL TLS clients are not impacted by
3030 *Peter Kästle and Samuel Sapalski*
3032 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3034 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3035 create a unique hash value based on the issuer and serial number data
3036 contained within an X509 certificate. However, it was failing to correctly
3037 handle any errors that may occur while parsing the issuer field (which might
3038 occur if the issuer field is maliciously constructed). This may subsequently
3039 result in a NULL pointer deref and a crash leading to a potential denial of
3045 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3046 padding mode to correctly check for rollback attacks. This is considered a
3047 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3052 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3053 functions. Previously they could overflow the output length argument in some
3054 cases where the input length is close to the maximum permissible length for
3055 an integer on the platform. In such cases the return value from the function
3056 call would be 1 (indicating success), but the output length value would be
3057 negative. This could cause applications to behave incorrectly or crash.
3062 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3063 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3064 could be exploited in a side channel attack to recover the password. Since
3065 the attack is local host only this is outside of the current OpenSSL
3066 threat model and therefore no CVE is assigned.
3068 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3073 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3075 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3076 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3077 If an attacker can control both items being compared then this could lead
3078 to a possible denial of service attack. OpenSSL itself uses the
3079 GENERAL_NAME_cmp function for two purposes:
3080 1) Comparing CRL distribution point names between an available CRL and a
3081 CRL distribution point embedded in an X509 certificate
3082 2) When verifying that a timestamp response token signer matches the
3083 timestamp authority name (exposed via the API functions
3084 TS_RESP_verify_response and TS_RESP_verify_token)
3089 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3091 * Certificates with explicit curve parameters are now disallowed in
3092 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3096 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3097 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3098 conversely, silently ignore DTLS protocol version bounds when configuring
3099 TLS-based contexts. The commands can be repeated to set bounds of both
3100 types. The same applies with the corresponding "min_protocol" and
3101 "max_protocol" command-line switches, in case some application uses both TLS
3104 SSL_CTX instances that are created for a fixed protocol version (e.g.
3105 TLSv1_server_method()) also silently ignore version bounds. Previously
3106 attempts to apply bounds to these protocol versions would result in an
3107 error. Now only the "version-flexible" SSL_CTX instances are subject to
3108 limits in configuration files in command-line options.
3112 * Handshake now fails if Extended Master Secret extension is dropped
3117 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3119 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3121 * Fixed segmentation fault in SSL_check_chain()
3122 Server or client applications that call the SSL_check_chain() function
3123 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3124 dereference as a result of incorrect handling of the
3125 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3126 or unrecognised signature algorithm is received from the peer. This could
3127 be exploited by a malicious peer in a Denial of Service attack.
3132 * Added AES consttime code for no-asm configurations
3133 an optional constant time support for AES was added
3134 when building openssl for no-asm.
3135 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3136 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3137 At this time this feature is by default disabled.
3138 It will be enabled by default in 3.0.
3142 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3144 * Revert the change of EOF detection while reading in libssl to avoid
3145 regressions in applications depending on the current way of reporting
3146 the EOF. As the existing method is not fully accurate the change to
3147 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3148 branch and will be present in the 3.0 release.
3152 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3153 when primes for RSA keys are computed.
3154 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3155 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3156 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3157 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3158 This avoids possible fingerprinting of newly generated RSA modules.
3162 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3164 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3165 while reading in libssl then we would report an error back to the
3166 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3167 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3168 therefore give a hint as to what went wrong.
3172 * Check that ed25519 and ed448 are allowed by the security level. Previously
3173 signature algorithms not using an MD were not being checked that they were
3174 allowed by the security level.
3178 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3179 was not quite right. The behaviour was not consistent between resumption
3180 and normal handshakes, and also not quite consistent with historical
3181 behaviour. The behaviour in various scenarios has been clarified and
3182 it has been updated to make it match historical behaviour as closely as
3187 * *[VMS only]* The header files that the VMS compilers include automatically,
3188 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3189 that the C++ compiler doesn't understand. This is a shortcoming in the
3190 compiler, but can be worked around with `__cplusplus` guards.
3192 C++ applications that use OpenSSL libraries must be compiled using the
3193 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3194 functions. Otherwise, only functions with symbols of less than 31
3195 characters can be used, as the linker will not be able to successfully
3196 resolve symbols with longer names.
3200 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3201 The presence of this system service is determined at run-time.
3205 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3210 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3212 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3213 number generator (RNG). This was intended to include protection in the
3214 event of a fork() system call in order to ensure that the parent and child
3215 processes did not share the same RNG state. However, this protection was not
3216 being used in the default case.
3218 A partial mitigation for this issue is that the output from a high
3219 precision timer is mixed into the RNG state so the likelihood of a parent
3220 and child process sharing state is significantly reduced.
3222 If an application already calls OPENSSL_init_crypto() explicitly using
3223 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3226 *Matthias St. Pierre*
3228 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3229 used even when parsing explicit parameters, when loading a encoded key
3230 or calling `EC_GROUP_new_from_ecpkparameters()`/
3231 `EC_GROUP_new_from_ecparameters()`.
3232 This prevents bypass of security hardening and performance gains,
3233 especially for curves with specialized EC_METHODs.
3234 By default, if a key encoded with explicit parameters is loaded and later
3235 encoded, the output is still encoded with explicit parameters, even if
3236 internally a "named" EC_GROUP is used for computation.
3240 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3241 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3242 NULL. After this change, only the cofactor parameter can be NULL. It also
3243 does some minimal sanity checks on the passed order.
3248 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3249 An attack is simple, if the first CMS_recipientInfo is valid but the
3250 second CMS_recipientInfo is chosen ciphertext. If the second
3251 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3252 encryption key will be replaced by garbage, and the message cannot be
3253 decoded, but if the RSA decryption fails, the correct encryption key is
3254 used and the recipient will not notice the attack.
3255 As a work around for this potential attack the length of the decrypted
3256 key must be equal to the cipher default key length, in case the
3257 certificate is not given and all recipientInfo are tried out.
3258 The old behaviour can be re-enabled in the CMS code by setting the
3259 CMS_DEBUG_DECRYPT flag.
3264 * Early start up entropy quality from the DEVRANDOM seed source has been
3265 improved for older Linux systems. The RAND subsystem will wait for
3266 /dev/random to be producing output before seeding from /dev/urandom.
3267 The seeded state is stored for future library initialisations using
3268 a system global shared memory segment. The shared memory identifier
3269 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3270 the desired value. The default identifier is 114.
3274 * Correct the extended master secret constant on EBCDIC systems. Without this
3275 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3276 negotiate EMS will fail. Unfortunately this also means that TLS connections
3277 between EBCDIC systems with this fix, and EBCDIC systems without this
3278 fix will fail if they negotiate EMS.
3282 * Use Windows installation paths in the mingw builds
3284 Mingw isn't a POSIX environment per se, which means that Windows
3285 paths should be used for installation.
3290 * Changed DH_check to accept parameters with order q and 2q subgroups.
3291 With order 2q subgroups the bit 0 of the private key is not secret
3292 but DH_generate_key works around that by clearing bit 0 of the
3293 private key for those. This avoids leaking bit 0 of the private key.
3297 * Significantly reduce secure memory usage by the randomness pools.
3301 * Revert the DEVRANDOM_WAIT feature for Linux systems
3303 The DEVRANDOM_WAIT feature added a select() call to wait for the
3304 /dev/random device to become readable before reading from the
3305 /dev/urandom device.
3307 It turned out that this change had negative side effects on
3308 performance which were not acceptable. After some discussion it
3309 was decided to revert this feature and leave it up to the OS
3310 resp. the platform maintainer to ensure a proper initialization
3311 during early boot time.
3313 *Matthias St. Pierre*
3315 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3317 * Add build tests for C++. These are generated files that only do one
3318 thing, to include one public OpenSSL head file each. This tests that
3319 the public header files can be usefully included in a C++ application.
3321 This test isn't enabled by default. It can be enabled with the option
3322 'enable-buildtest-c++'.
3326 * Enable SHA3 pre-hashing for ECDSA and DSA.
3330 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3331 This changes the size when using the `genpkey` command when no size is given.
3332 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3333 generation commands to use 2048 bits by default.
3337 * Reorganize the manual pages to consistently have RETURN VALUES,
3338 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3339 util/fix-doc-nits accordingly.
3341 *Paul Yang, Joshua Lock*
3343 * Add the missing accessor EVP_PKEY_get0_engine()
3347 * Have commands like `s_client` and `s_server` output the signature scheme
3348 along with other cipher suite parameters when debugging.
3352 * Make OPENSSL_config() error agnostic again.
3356 * Do the error handling in RSA decryption constant time.
3360 * Prevent over long nonces in ChaCha20-Poly1305.
3362 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3363 for every encryption operation. RFC 7539 specifies that the nonce value
3364 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3365 and front pads the nonce with 0 bytes if it is less than 12
3366 bytes. However it also incorrectly allows a nonce to be set of up to 16
3367 bytes. In this case only the last 12 bytes are significant and any
3368 additional leading bytes are ignored.
3370 It is a requirement of using this cipher that nonce values are
3371 unique. Messages encrypted using a reused nonce value are susceptible to
3372 serious confidentiality and integrity attacks. If an application changes
3373 the default nonce length to be longer than 12 bytes and then makes a
3374 change to the leading bytes of the nonce expecting the new value to be a
3375 new unique nonce then such an application could inadvertently encrypt
3376 messages with a reused nonce.
3378 Additionally the ignored bytes in a long nonce are not covered by the
3379 integrity guarantee of this cipher. Any application that relies on the
3380 integrity of these ignored leading bytes of a long nonce may be further
3381 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3382 is safe because no such use sets such a long nonce value. However user
3383 applications that use this cipher directly and set a non-default nonce
3384 length to be longer than 12 bytes may be vulnerable.
3386 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3392 * Add DEVRANDOM_WAIT feature for Linux systems
3394 On older Linux systems where the getrandom() system call is not available,
3395 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3396 Contrary to getrandom(), the /dev/urandom device will not block during
3397 early boot when the kernel CSPRNG has not been seeded yet.
3399 To mitigate this known weakness, use select() to wait for /dev/random to
3400 become readable before reading from /dev/urandom.
3402 * Ensure that SM2 only uses SM3 as digest algorithm
3406 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3408 * Change the info callback signals for the start and end of a post-handshake
3409 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3410 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3411 confused by this and assume that a TLSv1.2 renegotiation has started. This
3412 can break KeyUpdate handling. Instead we no longer signal the start and end
3413 of a post handshake message exchange (although the messages themselves are
3414 still signalled). This could break some applications that were expecting
3415 the old signals. However without this KeyUpdate is not usable for many
3420 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3422 * Timing vulnerability in DSA signature generation
3424 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3425 timing side channel attack. An attacker could use variations in the signing
3426 algorithm to recover the private key.
3428 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3433 * Timing vulnerability in ECDSA signature generation
3435 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3436 timing side channel attack. An attacker could use variations in the signing
3437 algorithm to recover the private key.
3439 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3444 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3445 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3446 of two gigabytes and the error handling improved.
3448 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3449 categorized as a normal bug, not a security issue, because the DRBG reseeds
3450 automatically and is fully functional even without additional randomness
3451 provided by the application.
3453 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3455 * Add a new ClientHello callback. Provides a callback interface that gives
3456 the application the ability to adjust the nascent SSL object at the
3457 earliest stage of ClientHello processing, immediately after extensions have
3458 been collected but before they have been processed. In particular, this
3459 callback can adjust the supported TLS versions in response to the contents
3464 * Add SM2 base algorithm support.
3468 * s390x assembly pack: add (improved) hardware-support for the following
3469 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3470 aes-cfb/cfb8, aes-ecb.
3474 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3475 parameter is no longer accepted, as it leads to a corrupt table. NULL
3476 pem_str is reserved for alias entries only.
3480 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3481 step for prime curves. The new implementation is based on formulae from
3482 differential addition-and-doubling in homogeneous projective coordinates
3483 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3484 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3485 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3486 to work in projective coordinates.
3488 *Billy Bob Brumley, Nicola Tuveri*
3490 * Change generating and checking of primes so that the error rate of not
3491 being prime depends on the intended use based on the size of the input.
3492 For larger primes this will result in more rounds of Miller-Rabin.
3493 The maximal error rate for primes with more than 1080 bits is lowered
3496 *Kurt Roeckx, Annie Yousar*
3498 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3502 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3503 moving between systems, and to avoid confusion when a Windows build is
3504 done with mingw vs with MSVC. For POSIX installs, there's still a
3505 symlink or copy named 'tsget' to avoid that confusion as well.
3509 * Revert blinding in ECDSA sign and instead make problematic addition
3510 length-invariant. Switch even to fixed-length Montgomery multiplication.
3514 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3515 step for binary curves. The new implementation is based on formulae from
3516 differential addition-and-doubling in mixed Lopez-Dahab projective
3517 coordinates, modified to independently blind the operands.
3519 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3521 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3522 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3523 EC_METHODs to implement their own specialized "ladder step", to take
3524 advantage of more favorable coordinate systems or more efficient
3525 differential addition-and-doubling algorithms.
3527 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3529 * Modified the random device based seed sources to keep the relevant
3530 file descriptors open rather than reopening them on each access.
3531 This allows such sources to operate in a chroot() jail without
3532 the associated device nodes being available. This behaviour can be
3533 controlled using RAND_keep_random_devices_open().
3537 * Numerous side-channel attack mitigations have been applied. This may have
3538 performance impacts for some algorithms for the benefit of improved
3539 security. Specific changes are noted in this change log by their respective
3544 * AIX shared library support overhaul. Switch to AIX "natural" way of
3545 handling shared libraries, which means collecting shared objects of
3546 different versions and bitnesses in one common archive. This allows to
3547 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3548 doesn't affect the way 3rd party applications are linked, only how
3549 multi-version installation is managed.
3553 * Make ec_group_do_inverse_ord() more robust and available to other
3554 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3555 mitigations are applied to the fallback BN_mod_inverse().
3556 When using this function rather than BN_mod_inverse() directly, new
3557 EC cryptosystem implementations are then safer-by-default.
3561 * Add coordinate blinding for EC_POINT and implement projective
3562 coordinate blinding for generic prime curves as a countermeasure to
3563 chosen point SCA attacks.
3565 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3567 * Add blinding to ECDSA and DSA signatures to protect against side channel
3568 attacks discovered by Keegan Ryan (NCC Group).
3572 * Enforce checking in the `pkeyutl` command to ensure that the input
3573 length does not exceed the maximum supported digest length when performing
3574 a sign, verify or verifyrecover operation.
3578 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3579 I/O in combination with something like select() or poll() will hang. This
3580 can be turned off again using SSL_CTX_clear_mode().
3581 Many applications do not properly handle non-application data records, and
3582 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3583 around the problems in those applications, but can also break some.
3584 It's recommended to read the manpages about SSL_read(), SSL_write(),
3585 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3586 SSL_CTX_set_read_ahead() again.
3590 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3591 now allow empty (zero character) pass phrases.
3595 * Apply blinding to binary field modular inversion and remove patent
3596 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3600 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3601 binary and prime elliptic curves.
3605 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3606 constant time fixed point multiplication.
3610 * Revise elliptic curve scalar multiplication with timing attack
3611 defenses: ec_wNAF_mul redirects to a constant time implementation
3612 when computing fixed point and variable point multiplication (which
3613 in OpenSSL are mostly used with secret scalars in keygen, sign,
3614 ECDH derive operations).
3615 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3618 * Updated CONTRIBUTING
3622 * Updated DRBG / RAND to request nonce and additional low entropy
3623 randomness from the system.
3625 *Matthias St. Pierre*
3627 * Updated 'openssl rehash' to use OpenSSL consistent default.
3631 * Moved the load of the ssl_conf module to libcrypto, which helps
3632 loading engines that libssl uses before libssl is initialised.
3636 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3640 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3642 *Ingo Schwarze, Rich Salz*
3644 * Added output of accepting IP address and port for 'openssl s_server'
3648 * Added a new API for TLSv1.3 ciphersuites:
3649 SSL_CTX_set_ciphersuites()
3650 SSL_set_ciphersuites()
3654 * Memory allocation failures consistently add an error to the error
3659 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3660 in libcrypto when run as setuid/setgid.
3664 * Load any config file by default when libssl is used.
3668 * Added new public header file <openssl/rand_drbg.h> and documentation
3669 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3671 *Matthias St. Pierre*
3673 * QNX support removed (cannot find contributors to get their approval
3674 for the license change).
3678 * TLSv1.3 replay protection for early data has been implemented. See the
3679 SSL_read_early_data() man page for further details.
3683 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3684 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3685 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3686 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3687 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3688 configuration has been separated out. See the ciphers man page or the
3689 SSL_CTX_set_ciphersuites() man page for more information.
3693 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3694 in responder mode now supports the new "-multi" option, which
3695 spawns the specified number of child processes to handle OCSP
3696 requests. The "-timeout" option now also limits the OCSP
3697 responder's patience to wait to receive the full client request
3698 on a newly accepted connection. Child processes are respawned
3699 as needed, and the CA index file is automatically reloaded
3700 when changed. This makes it possible to run the "ocsp" responder
3701 as a long-running service, making the OpenSSL CA somewhat more
3702 feature-complete. In this mode, most diagnostic messages logged
3703 after entering the event loop are logged via syslog(3) rather than
3708 * Added support for X448 and Ed448. Heavily based on original work by
3713 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3714 objects loaded. This adds the functions OSSL_STORE_expect() and
3715 OSSL_STORE_find() as well as needed tools to construct searches and
3716 get the search data out of them.
3720 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3721 version of OpenSSL should review their configuration settings to ensure
3722 that they are still appropriate for TLSv1.3. For further information see:
3723 <https://wiki.openssl.org/index.php/TLS1.3>
3727 * Grand redesign of the OpenSSL random generator
3729 The default RAND method now utilizes an AES-CTR DRBG according to
3730 NIST standard SP 800-90Ar1. The new random generator is essentially
3731 a port of the default random generator from the OpenSSL FIPS 2.0
3732 object module. It is a hybrid deterministic random bit generator
3733 using an AES-CTR bit stream and which seeds and reseeds itself
3734 automatically using trusted system entropy sources.
3736 Some of its new features are:
3737 - Support for multiple DRBG instances with seed chaining.
3738 - The default RAND method makes use of a DRBG.
3739 - There is a public and private DRBG instance.
3740 - The DRBG instances are fork-safe.
3741 - Keep all global DRBG instances on the secure heap if it is enabled.
3742 - The public and private DRBG instance are per thread for lock free
3745 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3747 * Changed Configure so it only says what it does and doesn't dump
3748 so much data. Instead, ./configdata.pm should be used as a script
3749 to display all sorts of configuration data.
3753 * Added processing of "make variables" to Configure.
3757 * Added SHA512/224 and SHA512/256 algorithm support.
3761 * The last traces of Netware support, first removed in 1.1.0, have
3766 * Get rid of Makefile.shared, and in the process, make the processing
3767 of certain files (rc.obj, or the .def/.map/.opt files produced from
3768 the ordinal files) more visible and hopefully easier to trace and
3769 debug (or make silent).
3773 * Make it possible to have environment variable assignments as
3774 arguments to config / Configure.
3778 * Add multi-prime RSA (RFC 8017) support.
3782 * Add SM3 implemented according to GB/T 32905-2016
3783 *Jack Lloyd <jack.lloyd@ribose.com>,*
3784 *Ronald Tse <ronald.tse@ribose.com>,*
3785 *Erick Borsboom <erick.borsboom@ribose.com>*
3787 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3788 as documented in RFC6066.
3789 Based on a patch from Tomasz Moń
3791 *Filipe Raimundo da Silva*
3793 * Add SM4 implemented according to GB/T 32907-2016.
3794 *Jack Lloyd <jack.lloyd@ribose.com>,*
3795 *Ronald Tse <ronald.tse@ribose.com>,*
3796 *Erick Borsboom <erick.borsboom@ribose.com>*
3798 * Reimplement -newreq-nodes and ERR_error_string_n; the
3799 original author does not agree with the license change.
3803 * Add ARIA AEAD TLS support.
3807 * Some macro definitions to support VS6 have been removed. Visual
3808 Studio 6 has not worked since 1.1.0
3812 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3813 without clearing the errors.
3817 * Add "atfork" functions. If building on a system that without
3818 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3819 requirements. The RAND facility now uses/requires this.
3827 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3828 not possible to disable entirely. However, it's still possible to
3829 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3832 To disable, configure with 'no-ui-console'. 'no-ui' is still
3833 possible to use as an alias. Check at compile time with the
3834 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3835 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3839 * Add a STORE module, which implements a uniform and URI based reader of
3840 stores that can contain keys, certificates, CRLs and numerous other
3841 objects. The main API is loosely based on a few stdio functions,
3842 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3843 OSSL_STORE_error and OSSL_STORE_close.
3844 The implementation uses backends called "loaders" to implement arbitrary
3845 URI schemes. There is one built in "loader" for the 'file' scheme.
3849 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3850 then adjusted to work on FreeBSD 8.4 as well.
3851 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3852 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3856 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3857 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3858 error code calls like this:
3860 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3862 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3863 that can be encoded in C. For the foreseeable future, this will only
3866 *Richard Levitte and Tim Hudson*
3868 * Removed BSD cryptodev engine.
3872 * Add a build target 'build_all_generated', to build all generated files
3873 and only that. This can be used to prepare everything that requires
3874 things like perl for a system that lacks perl and then move everything
3875 to that system and do the rest of the build there.
3879 * In the UI interface, make it possible to duplicate the user data. This
3880 can be used by engines that need to retain the data for a longer time
3881 than just the call where this user data is passed.
3885 * Ignore the '-named_curve auto' value for compatibility of applications
3888 *Tomáš Mráz <tmraz@fedoraproject.org>*
3890 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3891 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3892 alerts across multiple records (some of which could be empty). In practice
3893 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3894 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3895 support this at all. Supporting it adds significant complexity to the
3896 record layer, and its removal is unlikely to cause interoperability
3901 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3902 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3903 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3908 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3909 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3911 *Richard Levitte, Andy Polyakov*
3913 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3918 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3919 platform rather than 'mingw'.
3923 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3924 success if they are asked to add an object which already exists
3925 in the store. This change cascades to other functions which load
3926 certificates and CRLs.
3930 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3931 facilitate stack unwinding even from assembly subroutines.
3935 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3936 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3940 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3941 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3942 which is the minimum version we support.
3946 * Certificate time validation (X509_cmp_time) enforces stricter
3947 compliance with RFC 5280. Fractional seconds and timezone offsets
3948 are no longer allowed.
3952 * Add support for ARIA
3956 * s_client will now send the Server Name Indication (SNI) extension by
3957 default unless the new "-noservername" option is used. The server name is
3958 based on the host provided to the "-connect" option unless overridden by
3959 using "-servername".
3963 * Add support for SipHash
3967 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3968 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3969 prevent issues where no progress is being made and the peer continually
3970 sends unrecognised record types, using up resources processing them.
3974 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3975 using the algorithm defined in
3976 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3980 * Heartbeat support has been removed; the ABI is changed for now.
3982 *Richard Levitte, Rich Salz*
3984 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3988 * The RSA "null" method, which was partially supported to avoid patent
3989 issues, has been replaced to always returns NULL.
3996 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3998 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3999 used even when parsing explicit parameters, when loading a encoded key
4000 or calling `EC_GROUP_new_from_ecpkparameters()`/
4001 `EC_GROUP_new_from_ecparameters()`.
4002 This prevents bypass of security hardening and performance gains,
4003 especially for curves with specialized EC_METHODs.
4004 By default, if a key encoded with explicit parameters is loaded and later
4005 encoded, the output is still encoded with explicit parameters, even if
4006 internally a "named" EC_GROUP is used for computation.
4010 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4011 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4012 NULL. After this change, only the cofactor parameter can be NULL. It also
4013 does some minimal sanity checks on the passed order.
4018 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4019 An attack is simple, if the first CMS_recipientInfo is valid but the
4020 second CMS_recipientInfo is chosen ciphertext. If the second
4021 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4022 encryption key will be replaced by garbage, and the message cannot be
4023 decoded, but if the RSA decryption fails, the correct encryption key is
4024 used and the recipient will not notice the attack.
4025 As a work around for this potential attack the length of the decrypted
4026 key must be equal to the cipher default key length, in case the
4027 certificate is not given and all recipientInfo are tried out.
4028 The old behaviour can be re-enabled in the CMS code by setting the
4029 CMS_DEBUG_DECRYPT flag.
4034 * Use Windows installation paths in the mingw builds
4036 Mingw isn't a POSIX environment per se, which means that Windows
4037 paths should be used for installation.
4042 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4044 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4045 This changes the size when using the `genpkey` command when no size is given.
4046 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4047 generation commands to use 2048 bits by default.
4051 * Prevent over long nonces in ChaCha20-Poly1305.
4053 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4054 for every encryption operation. RFC 7539 specifies that the nonce value
4055 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4056 and front pads the nonce with 0 bytes if it is less than 12
4057 bytes. However it also incorrectly allows a nonce to be set of up to 16
4058 bytes. In this case only the last 12 bytes are significant and any
4059 additional leading bytes are ignored.
4061 It is a requirement of using this cipher that nonce values are
4062 unique. Messages encrypted using a reused nonce value are susceptible to
4063 serious confidentiality and integrity attacks. If an application changes
4064 the default nonce length to be longer than 12 bytes and then makes a
4065 change to the leading bytes of the nonce expecting the new value to be a
4066 new unique nonce then such an application could inadvertently encrypt
4067 messages with a reused nonce.
4069 Additionally the ignored bytes in a long nonce are not covered by the
4070 integrity guarantee of this cipher. Any application that relies on the
4071 integrity of these ignored leading bytes of a long nonce may be further
4072 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4073 is safe because no such use sets such a long nonce value. However user
4074 applications that use this cipher directly and set a non-default nonce
4075 length to be longer than 12 bytes may be vulnerable.
4077 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4083 * Added SCA hardening for modular field inversion in EC_GROUP through
4084 a new dedicated field_inv() pointer in EC_METHOD.
4085 This also addresses a leakage affecting conversions from projective
4086 to affine coordinates.
4088 *Billy Bob Brumley, Nicola Tuveri*
4090 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4091 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4095 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4099 * Remove the 'dist' target and add a tarball building script. The
4100 'dist' target has fallen out of use, and it shouldn't be
4101 necessary to configure just to create a source distribution.
4105 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4107 * Timing vulnerability in DSA signature generation
4109 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4110 timing side channel attack. An attacker could use variations in the signing
4111 algorithm to recover the private key.
4113 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4118 * Timing vulnerability in ECDSA signature generation
4120 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4121 timing side channel attack. An attacker could use variations in the signing
4122 algorithm to recover the private key.
4124 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4129 * Add coordinate blinding for EC_POINT and implement projective
4130 coordinate blinding for generic prime curves as a countermeasure to
4131 chosen point SCA attacks.
4133 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4135 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4137 * Client DoS due to large DH parameter
4139 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4140 malicious server can send a very large prime value to the client. This will
4141 cause the client to spend an unreasonably long period of time generating a
4142 key for this prime resulting in a hang until the client has finished. This
4143 could be exploited in a Denial Of Service attack.
4145 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4150 * Cache timing vulnerability in RSA Key Generation
4152 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4153 a cache timing side channel attack. An attacker with sufficient access to
4154 mount cache timing attacks during the RSA key generation process could
4155 recover the private key.
4157 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4158 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4163 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4164 parameter is no longer accepted, as it leads to a corrupt table. NULL
4165 pem_str is reserved for alias entries only.
4169 * Revert blinding in ECDSA sign and instead make problematic addition
4170 length-invariant. Switch even to fixed-length Montgomery multiplication.
4174 * Change generating and checking of primes so that the error rate of not
4175 being prime depends on the intended use based on the size of the input.
4176 For larger primes this will result in more rounds of Miller-Rabin.
4177 The maximal error rate for primes with more than 1080 bits is lowered
4180 *Kurt Roeckx, Annie Yousar*
4182 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4186 * Add blinding to ECDSA and DSA signatures to protect against side channel
4187 attacks discovered by Keegan Ryan (NCC Group).
4191 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4192 now allow empty (zero character) pass phrases.
4196 * Certificate time validation (X509_cmp_time) enforces stricter
4197 compliance with RFC 5280. Fractional seconds and timezone offsets
4198 are no longer allowed.
4202 * Fixed a text canonicalisation bug in CMS
4204 Where a CMS detached signature is used with text content the text goes
4205 through a canonicalisation process first prior to signing or verifying a
4206 signature. This process strips trailing space at the end of lines, converts
4207 line terminators to CRLF and removes additional trailing line terminators
4208 at the end of a file. A bug in the canonicalisation process meant that
4209 some characters, such as form-feed, were incorrectly treated as whitespace
4210 and removed. This is contrary to the specification (RFC5485). This fix
4211 could mean that detached text data signed with an earlier version of
4212 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4213 signed with a fixed OpenSSL may fail to verify with an earlier version of
4214 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4215 and use the "-binary" flag (for the "cms" command line application) or set
4216 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4220 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4222 * Constructed ASN.1 types with a recursive definition could exceed the stack
4224 Constructed ASN.1 types with a recursive definition (such as can be found
4225 in PKCS7) could eventually exceed the stack given malicious input with
4226 excessive recursion. This could result in a Denial Of Service attack. There
4227 are no such structures used within SSL/TLS that come from untrusted sources
4228 so this is considered safe.
4230 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4236 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4238 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4239 effectively reduced to only comparing the least significant bit of each
4240 byte. This allows an attacker to forge messages that would be considered as
4241 authenticated in an amount of tries lower than that guaranteed by the
4242 security claims of the scheme. The module can only be compiled by the
4243 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4245 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4251 * Add a build target 'build_all_generated', to build all generated files
4252 and only that. This can be used to prepare everything that requires
4253 things like perl for a system that lacks perl and then move everything
4254 to that system and do the rest of the build there.
4258 * Backport SSL_OP_NO_RENGOTIATION
4260 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4261 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4262 changes this is no longer possible in 1.1.0. Therefore, the new
4263 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4264 1.1.0 to provide equivalent functionality.
4266 Note that if an application built against 1.1.0h headers (or above) is run
4267 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4268 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4272 * Removed the OS390-Unix config target. It relied on a script that doesn't
4277 * rsaz_1024_mul_avx2 overflow bug on x86_64
4279 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4280 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4281 Analysis suggests that attacks against RSA and DSA as a result of this
4282 defect would be very difficult to perform and are not believed likely.
4283 Attacks against DH1024 are considered just feasible, because most of the
4284 work necessary to deduce information about a private key may be performed
4285 offline. The amount of resources required for such an attack would be
4286 significant. However, for an attack on TLS to be meaningful, the server
4287 would have to share the DH1024 private key among multiple clients, which is
4288 no longer an option since CVE-2016-0701.
4290 This only affects processors that support the AVX2 but not ADX extensions
4291 like Intel Haswell (4th generation).
4293 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4294 was originally found via the OSS-Fuzz project.
4299 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4301 * bn_sqrx8x_internal carry bug on x86_64
4303 There is a carry propagating bug in the x86_64 Montgomery squaring
4304 procedure. No EC algorithms are affected. Analysis suggests that attacks
4305 against RSA and DSA as a result of this defect would be very difficult to
4306 perform and are not believed likely. Attacks against DH are considered just
4307 feasible (although very difficult) because most of the work necessary to
4308 deduce information about a private key may be performed offline. The amount
4309 of resources required for such an attack would be very significant and
4310 likely only accessible to a limited number of attackers. An attacker would
4311 additionally need online access to an unpatched system using the target
4312 private key in a scenario with persistent DH parameters and a private
4313 key that is shared between multiple clients.
4315 This only affects processors that support the BMI1, BMI2 and ADX extensions
4316 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4318 This issue was reported to OpenSSL by the OSS-Fuzz project.
4323 * Malformed X.509 IPAddressFamily could cause OOB read
4325 If an X.509 certificate has a malformed IPAddressFamily extension,
4326 OpenSSL could do a one-byte buffer overread. The most likely result
4327 would be an erroneous display of the certificate in text format.
4329 This issue was reported to OpenSSL by the OSS-Fuzz project.
4334 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4336 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4337 platform rather than 'mingw'.
4341 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4342 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4343 which is the minimum version we support.
4347 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4349 * Encrypt-Then-Mac renegotiation crash
4351 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4352 negotiated where it was not in the original handshake (or vice-versa) then
4353 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4354 and servers are affected.
4356 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4361 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4363 * Truncated packet could crash via OOB read
4365 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4366 cipher is being used, then a truncated packet can cause that host to
4367 perform an out-of-bounds read, usually resulting in a crash.
4369 This issue was reported to OpenSSL by Robert Święcki of Google.
4374 * Bad (EC)DHE parameters cause a client crash
4376 If a malicious server supplies bad parameters for a DHE or ECDHE key
4377 exchange then this can result in the client attempting to dereference a
4378 NULL pointer leading to a client crash. This could be exploited in a Denial
4381 This issue was reported to OpenSSL by Guido Vranken.
4386 * BN_mod_exp may produce incorrect results on x86_64
4388 There is a carry propagating bug in the x86_64 Montgomery squaring
4389 procedure. No EC algorithms are affected. Analysis suggests that attacks
4390 against RSA and DSA as a result of this defect would be very difficult to
4391 perform and are not believed likely. Attacks against DH are considered just
4392 feasible (although very difficult) because most of the work necessary to
4393 deduce information about a private key may be performed offline. The amount
4394 of resources required for such an attack would be very significant and
4395 likely only accessible to a limited number of attackers. An attacker would
4396 additionally need online access to an unpatched system using the target
4397 private key in a scenario with persistent DH parameters and a private
4398 key that is shared between multiple clients. For example this can occur by
4399 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4400 similar to CVE-2015-3193 but must be treated as a separate problem.
4402 This issue was reported to OpenSSL by the OSS-Fuzz project.
4407 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4409 * ChaCha20/Poly1305 heap-buffer-overflow
4411 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4412 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4413 crash. This issue is not considered to be exploitable beyond a DoS.
4415 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4420 * CMS Null dereference
4422 Applications parsing invalid CMS structures can crash with a NULL pointer
4423 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4424 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4425 structure callback if an attempt is made to free certain invalid encodings.
4426 Only CHOICE structures using a callback which do not handle NULL value are
4429 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4434 * Montgomery multiplication may produce incorrect results
4436 There is a carry propagating bug in the Broadwell-specific Montgomery
4437 multiplication procedure that handles input lengths divisible by, but
4438 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4439 and DH private keys are impossible. This is because the subroutine in
4440 question is not used in operations with the private key itself and an input
4441 of the attacker's direct choice. Otherwise the bug can manifest itself as
4442 transient authentication and key negotiation failures or reproducible
4443 erroneous outcome of public-key operations with specially crafted input.
4444 Among EC algorithms only Brainpool P-512 curves are affected and one
4445 presumably can attack ECDH key negotiation. Impact was not analyzed in
4446 detail, because pre-requisites for attack are considered unlikely. Namely
4447 multiple clients have to choose the curve in question and the server has to
4448 share the private key among them, neither of which is default behaviour.
4449 Even then only clients that chose the curve will be affected.
4451 This issue was publicly reported as transient failures and was not
4452 initially recognized as a security issue. Thanks to Richard Morgan for
4453 providing reproducible case.
4458 * Removed automatic addition of RPATH in shared libraries and executables,
4459 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4463 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4465 * Fix Use After Free for large message sizes
4467 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4468 message larger than approx 16k is received then the underlying buffer to
4469 store the incoming message is reallocated and moved. Unfortunately a
4470 dangling pointer to the old location is left which results in an attempt to
4471 write to the previously freed location. This is likely to result in a
4472 crash, however it could potentially lead to execution of arbitrary code.
4474 This issue only affects OpenSSL 1.1.0a.
4476 This issue was reported to OpenSSL by Robert Święcki.
4481 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4483 * OCSP Status Request extension unbounded memory growth
4485 A malicious client can send an excessively large OCSP Status Request
4486 extension. If that client continually requests renegotiation, sending a
4487 large OCSP Status Request extension each time, then there will be unbounded
4488 memory growth on the server. This will eventually lead to a Denial Of
4489 Service attack through memory exhaustion. Servers with a default
4490 configuration are vulnerable even if they do not support OCSP. Builds using
4491 the "no-ocsp" build time option are not affected.
4493 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4498 * SSL_peek() hang on empty record
4500 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4501 sends an empty record. This could be exploited by a malicious peer in a
4502 Denial Of Service attack.
4504 This issue was reported to OpenSSL by Alex Gaynor.
4509 * Excessive allocation of memory in tls_get_message_header() and
4510 dtls1_preprocess_fragment()
4512 A (D)TLS message includes 3 bytes for its length in the header for the
4513 message. This would allow for messages up to 16Mb in length. Messages of
4514 this length are excessive and OpenSSL includes a check to ensure that a
4515 peer is sending reasonably sized messages in order to avoid too much memory
4516 being consumed to service a connection. A flaw in the logic of version
4517 1.1.0 means that memory for the message is allocated too early, prior to
4518 the excessive message length check. Due to way memory is allocated in
4519 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4520 to service a connection. This could lead to a Denial of Service through
4521 memory exhaustion. However, the excessive message length check still takes
4522 place, and this would cause the connection to immediately fail. Assuming
4523 that the application calls SSL_free() on the failed connection in a timely
4524 manner then the 21Mb of allocated memory will then be immediately freed
4525 again. Therefore, the excessive memory allocation will be transitory in
4526 nature. This then means that there is only a security impact if:
4528 1) The application does not call SSL_free() in a timely manner in the event
4529 that the connection fails
4531 2) The application is working in a constrained environment where there is
4532 very little free memory
4534 3) The attacker initiates multiple connection attempts such that there are
4535 multiple connections in a state where memory has been allocated for the
4536 connection; SSL_free() has not yet been called; and there is insufficient
4537 memory to service the multiple requests.
4539 Except in the instance of (1) above any Denial Of Service is likely to be
4540 transitory because as soon as the connection fails the memory is
4541 subsequently freed again in the SSL_free() call. However there is an
4542 increased risk during this period of application crashes due to the lack of
4543 memory - which would then mean a more serious Denial of Service.
4545 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4546 (CVE-2016-6307 and CVE-2016-6308)
4550 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4551 had to be removed. Primary reason is that vendor assembler can't
4552 assemble our modules with -KPIC flag. As result it, assembly
4553 support, was not even available as option. But its lack means
4554 lack of side-channel resistant code, which is incompatible with
4555 security by todays standards. Fortunately gcc is readily available
4556 prepackaged option, which we firmly point at...
4560 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4562 * Windows command-line tool supports UTF-8 opt-in option for arguments
4563 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4564 (to any value) allows Windows user to access PKCS#12 file generated
4565 with Windows CryptoAPI and protected with non-ASCII password, as well
4566 as files generated under UTF-8 locale on Linux also protected with
4571 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4572 have been disabled by default and removed from DEFAULT, just like RC4.
4573 See the RC4 item below to re-enable both.
4577 * The method for finding the storage location for the Windows RAND seed file
4578 has changed. First we check %RANDFILE%. If that is not set then we check
4579 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4580 all else fails we fall back to C:\.
4584 * The EVP_EncryptUpdate() function has had its return type changed from void
4585 to int. A return of 0 indicates and error while a return of 1 indicates
4590 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4591 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4592 off the constant time implementation for RSA, DSA and DH have been made
4593 no-ops and deprecated.
4597 * Windows RAND implementation was simplified to only get entropy by
4598 calling CryptGenRandom(). Various other RAND-related tickets
4601 *Joseph Wylie Yandle, Rich Salz*
4603 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4604 and `OPENSSL_LH_`, respectively. The old names are available
4605 with API compatibility. They new names are now completely documented.
4609 * Unify TYPE_up_ref(obj) methods signature.
4610 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4611 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4612 int (instead of void) like all others TYPE_up_ref() methods.
4613 So now these methods also check the return value of CRYPTO_atomic_add(),
4614 and the validity of object reference counter.
4616 *fdasilvayy@gmail.com*
4618 * With Windows Visual Studio builds, the .pdb files are installed
4619 alongside the installed libraries and executables. For a static
4620 library installation, ossl_static.pdb is the associate compiler
4621 generated .pdb file to be used when linking programs.
4625 * Remove openssl.spec. Packaging files belong with the packagers.
4629 * Automatic Darwin/OSX configuration has had a refresh, it will now
4630 recognise x86_64 architectures automatically. You can still decide
4631 to build for a different bitness with the environment variable
4632 KERNEL_BITS (can be 32 or 64), for example:
4634 KERNEL_BITS=32 ./config
4638 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4639 256 bit AES and HMAC with SHA256.
4643 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4647 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4651 * To enable users to have their own config files and build file templates,
4652 Configure looks in the directory indicated by the environment variable
4653 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4654 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4655 name and is used as is.
4659 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4660 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4661 X509_CERT_FILE_CTX was removed.
4665 * "shared" builds are now the default. To create only static libraries use
4666 the "no-shared" Configure option.
4670 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4671 All of these option have not worked for some while and are fundamental
4676 * Make various cleanup routines no-ops and mark them as deprecated. Most
4677 global cleanup functions are no longer required because they are handled
4678 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4679 Explicitly de-initing can cause problems (e.g. where a library that uses
4680 OpenSSL de-inits, but an application is still using it). The affected
4681 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4682 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4683 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4684 COMP_zlib_cleanup().
4688 * --strict-warnings no longer enables runtime debugging options
4689 such as REF_DEBUG. Instead, debug options are automatically
4690 enabled with '--debug' builds.
4692 *Andy Polyakov, Emilia Käsper*
4694 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4695 have been moved out of the public header files. New functions for managing
4696 these have been added.
4700 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4701 objects have been moved out of the public header files. New
4702 functions for managing these have been added.
4706 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4707 have been moved out of the public header files. New functions for managing
4708 these have been added.
4712 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4713 moved out of the public header files. New functions for managing these
4718 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4722 * Removed the mk1mf build scripts.
4726 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4727 it is always safe to #include a header now.
4731 * Removed the aged BC-32 config and all its supporting scripts
4735 * Removed support for Ultrix, Netware, and OS/2.
4739 * Add support for HKDF.
4741 *Alessandro Ghedini*
4743 * Add support for blake2b and blake2s
4747 * Added support for "pipelining". Ciphers that have the
4748 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4749 encryptions/decryptions simultaneously. There are currently no built-in
4750 ciphers with this property but the expectation is that engines will be able
4751 to offer it to significantly improve throughput. Support has been extended
4752 into libssl so that multiple records for a single connection can be
4753 processed in one go (for >=TLS 1.1).
4757 * Added the AFALG engine. This is an async capable engine which is able to
4758 offload work to the Linux kernel. In this initial version it only supports
4759 AES128-CBC. The kernel must be version 4.1.0 or greater.
4763 * OpenSSL now uses a new threading API. It is no longer necessary to
4764 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4765 are two supported threading models: pthreads and windows threads. It is
4766 also possible to configure OpenSSL at compile time for "no-threads". The
4767 old threading API should no longer be used. The functions have been
4768 replaced with "no-op" compatibility macros.
4770 *Alessandro Ghedini, Matt Caswell*
4772 * Modify behavior of ALPN to invoke callback after SNI/servername
4773 callback, such that updates to the SSL_CTX affect ALPN.
4777 * Add SSL_CIPHER queries for authentication and key-exchange.
4781 * Changes to the DEFAULT cipherlist:
4782 - Prefer (EC)DHE handshakes over plain RSA.
4783 - Prefer AEAD ciphers over legacy ciphers.
4784 - Prefer ECDSA over RSA when both certificates are available.
4785 - Prefer TLSv1.2 ciphers/PRF.
4786 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4791 * Change the ECC default curve list to be this, in order: x25519,
4792 secp256r1, secp521r1, secp384r1.
4796 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4797 disabled by default. They can be re-enabled using the
4798 enable-weak-ssl-ciphers option to Configure.
4802 * If the server has ALPN configured, but supports no protocols that the
4803 client advertises, send a fatal "no_application_protocol" alert.
4804 This behaviour is SHALL in RFC 7301, though it isn't universally
4805 implemented by other servers.
4809 * Add X25519 support.
4810 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4811 for public and private key encoding using the format documented in
4812 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4813 key generation and key derivation.
4815 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4820 * Deprecate SRP_VBASE_get_by_user.
4821 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4822 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4823 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4824 seed, even if the seed is configured.
4826 Users should use SRP_VBASE_get1_by_user instead. Note that in
4827 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4828 also that even though configuring the SRP seed attempts to hide
4829 invalid usernames by continuing the handshake with fake
4830 credentials, this behaviour is not constant time and no strong
4831 guarantees are made that the handshake is indistinguishable from
4832 that of a valid user.
4836 * Configuration change; it's now possible to build dynamic engines
4837 without having to build shared libraries and vice versa. This
4838 only applies to the engines in `engines/`, those in `crypto/engine/`
4839 will always be built into libcrypto (i.e. "static").
4841 Building dynamic engines is enabled by default; to disable, use
4842 the configuration option "disable-dynamic-engine".
4844 The only requirements for building dynamic engines are the
4845 presence of the DSO module and building with position independent
4846 code, so they will also automatically be disabled if configuring
4847 with "disable-dso" or "disable-pic".
4849 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4850 are also taken away from openssl/opensslconf.h, as they are
4855 * Configuration change; if there is a known flag to compile
4856 position independent code, it will always be applied on the
4857 libcrypto and libssl object files, and never on the application
4858 object files. This means other libraries that use routines from
4859 libcrypto / libssl can be made into shared libraries regardless
4860 of how OpenSSL was configured.
4862 If this isn't desirable, the configuration options "disable-pic"
4863 or "no-pic" can be used to disable the use of PIC. This will
4864 also disable building shared libraries and dynamic engines.
4868 * Removed JPAKE code. It was experimental and has no wide use.
4872 * The INSTALL_PREFIX Makefile variable has been renamed to
4873 DESTDIR. That makes for less confusion on what this variable
4874 is for. Also, the configuration option --install_prefix is
4879 * Heartbeat for TLS has been removed and is disabled by default
4880 for DTLS; configure with enable-heartbeats. Code that uses the
4881 old #define's might need to be updated.
4883 *Emilia Käsper, Rich Salz*
4885 * Rename REF_CHECK to REF_DEBUG.
4889 * New "unified" build system
4891 The "unified" build system is aimed to be a common system for all
4892 platforms we support. With it comes new support for VMS.
4894 This system builds supports building in a different directory tree
4895 than the source tree. It produces one Makefile (for unix family
4896 or lookalikes), or one descrip.mms (for VMS).
4898 The source of information to make the Makefile / descrip.mms is
4899 small files called 'build.info', holding the necessary
4900 information for each directory with source to compile, and a
4901 template in Configurations, like unix-Makefile.tmpl or
4904 With this change, the library names were also renamed on Windows
4905 and on VMS. They now have names that are closer to the standard
4906 on Unix, and include the major version number, and in certain
4907 cases, the architecture they are built for. See "Notes on shared
4908 libraries" in INSTALL.
4910 We rely heavily on the perl module Text::Template.
4914 * Added support for auto-initialisation and de-initialisation of the library.
4915 OpenSSL no longer requires explicit init or deinit routines to be called,
4916 except in certain circumstances. See the OPENSSL_init_crypto() and
4917 OPENSSL_init_ssl() man pages for further information.
4921 * The arguments to the DTLSv1_listen function have changed. Specifically the
4922 "peer" argument is now expected to be a BIO_ADDR object.
4924 * Rewrite of BIO networking library. The BIO library lacked consistent
4925 support of IPv6, and adding it required some more extensive
4926 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4927 which hold all types of addresses and chains of address information.
4928 It also introduces a new API, with functions like BIO_socket,
4929 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4930 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4931 have been adapted accordingly.
4935 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4940 * CRIME protection: disable compression by default, even if OpenSSL is
4941 compiled with zlib enabled. Applications can still enable compression
4942 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4943 using the SSL_CONF library to configure compression.
4947 * The signature of the session callback configured with
4948 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4949 was explicitly marked as `const unsigned char*` instead of
4954 * Always DPURIFY. Remove the use of uninitialized memory in the
4955 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4959 * Removed many obsolete configuration items, including
4960 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4961 MD2_CHAR, MD2_INT, MD2_LONG
4963 IDEA_SHORT, IDEA_LONG
4964 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4966 *Rich Salz, with advice from Andy Polyakov*
4968 * Many BN internals have been moved to an internal header file.
4970 *Rich Salz with help from Andy Polyakov*
4972 * Configuration and writing out the results from it has changed.
4973 Files such as Makefile include/openssl/opensslconf.h and are now
4974 produced through general templates, such as Makefile.in and
4975 crypto/opensslconf.h.in and some help from the perl module
4978 Also, the center of configuration information is no longer
4979 Makefile. Instead, Configure produces a perl module in
4980 configdata.pm which holds most of the config data (in the hash
4981 table %config), the target data that comes from the target
4982 configuration in one of the `Configurations/*.conf` files (in
4987 * To clarify their intended purposes, the Configure options
4988 --prefix and --openssldir change their semantics, and become more
4989 straightforward and less interdependent.
4991 --prefix shall be used exclusively to give the location INSTALLTOP
4992 where programs, scripts, libraries, include files and manuals are
4993 going to be installed. The default is now /usr/local.
4995 --openssldir shall be used exclusively to give the default
4996 location OPENSSLDIR where certificates, private keys, CRLs are
4997 managed. This is also where the default openssl.cnf gets
4999 If the directory given with this option is a relative path, the
5000 values of both the --prefix value and the --openssldir value will
5001 be combined to become OPENSSLDIR.
5002 The default for --openssldir is INSTALLTOP/ssl.
5004 Anyone who uses --openssldir to specify where OpenSSL is to be
5005 installed MUST change to use --prefix instead.
5009 * The GOST engine was out of date and therefore it has been removed. An up
5010 to date GOST engine is now being maintained in an external repository.
5011 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5012 support for GOST ciphersuites (these are only activated if a GOST engine
5017 * EGD is no longer supported by default; use enable-egd when
5020 *Ben Kaduk and Rich Salz*
5022 * The distribution now has Makefile.in files, which are used to
5023 create Makefile's when Configure is run. *Configure must be run
5024 before trying to build now.*
5028 * The return value for SSL_CIPHER_description() for error conditions
5033 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5035 Obtaining and performing DNSSEC validation of TLSA records is
5036 the application's responsibility. The application provides
5037 the TLSA records of its choice to OpenSSL, and these are then
5038 used to authenticate the peer.
5040 The TLSA records need not even come from DNS. They can, for
5041 example, be used to implement local end-entity certificate or
5042 trust-anchor "pinning", where the "pin" data takes the form
5043 of TLSA records, which can augment or replace verification
5044 based on the usual WebPKI public certification authorities.
5048 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5049 continues to support deprecated interfaces in default builds.
5050 However, applications are strongly advised to compile their
5051 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5052 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5053 or the 1.1.0 releases.
5055 In environments in which all applications have been ported to
5056 not use any deprecated interfaces OpenSSL's Configure script
5057 should be used with the --api=1.1.0 option to entirely remove
5058 support for the deprecated features from the library and
5059 unconditionally disable them in the installed headers.
5060 Essentially the same effect can be achieved with the "no-deprecated"
5061 argument to Configure, except that this will always restrict
5062 the build to just the latest API, rather than a fixed API
5065 As applications are ported to future revisions of the API,
5066 they should update their compile-time OPENSSL_API_COMPAT define
5067 accordingly, but in most cases should be able to continue to
5068 compile with later releases.
5070 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5071 0x10000000L and 0x00908000L, respectively. However those
5072 versions did not support the OPENSSL_API_COMPAT feature, and
5073 so applications are not typically tested for explicit support
5074 of just the undeprecated features of either release.
5078 * Add support for setting the minimum and maximum supported protocol.
5079 It can bet set via the SSL_set_min_proto_version() and
5080 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5081 MaxProtocol. It's recommended to use the new APIs to disable
5082 protocols instead of disabling individual protocols using
5083 SSL_set_options() or SSL_CONF's Protocol. This change also
5084 removes support for disabling TLS 1.2 in the OpenSSL TLS
5085 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5089 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5093 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5094 and integrates ECDSA and ECDH functionality into EC. Implementations can
5095 now redirect key generation and no longer need to convert to or from
5098 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5099 include the ec.h header file instead.
5103 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5104 ciphers who are no longer supported and drops support the ephemeral RSA key
5105 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5109 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5110 opaque. For HMAC_CTX, the following constructors and destructors
5113 HMAC_CTX *HMAC_CTX_new(void);
5114 void HMAC_CTX_free(HMAC_CTX *ctx);
5116 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5117 destroy such methods has been added. See EVP_MD_meth_new(3) and
5118 EVP_CIPHER_meth_new(3) for documentation.
5121 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5122 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5123 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5124 an already created structure.
5125 2) For consistency with the majority of our object creators and
5126 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5127 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5128 for deprecated builds.
5132 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5133 cryptographic operations to be performed asynchronously as long as an
5134 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5135 further details. Libssl has also had this capability integrated with the
5136 introduction of the new mode SSL_MODE_ASYNC and associated error
5137 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5138 pages. This work was developed in partnership with Intel Corp.
5142 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5143 always enabled now. If you want to disable the support you should
5144 exclude it using the list of supported ciphers. This also means that the
5145 "-no_ecdhe" option has been removed from s_server.
5149 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5150 SSL_{CTX_}set1_curves() which can set a list.
5154 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5155 curve you want to support using SSL_{CTX_}set1_curves().
5159 * State machine rewrite. The state machine code has been significantly
5160 refactored in order to remove much duplication of code and solve issues
5161 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5162 further details). This change does have some associated API changes.
5163 Notably the SSL_state() function has been removed and replaced by
5164 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5165 SSL_set_state() has been removed altogether. The previous handshake states
5166 defined in ssl.h and ssl3.h have also been removed.
5170 * All instances of the string "ssleay" in the public API were replaced
5171 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5172 Some error codes related to internal RSA_eay API's were renamed.
5176 * The demo files in crypto/threads were moved to demo/threads.
5180 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5183 *Matt Caswell, Rich Salz*
5185 * New ASN.1 embed macro.
5187 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5188 structure is not allocated: it is part of the parent. That is instead of
5196 This reduces memory fragmentation and make it impossible to accidentally
5197 set a mandatory field to NULL.
5199 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5200 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5201 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5206 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5210 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5211 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5212 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5213 DES and RC4 ciphersuites.
5217 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5218 This changes the decoding behaviour for some invalid messages,
5219 though the change is mostly in the more lenient direction, and
5220 legacy behaviour is preserved as much as possible.
5224 * Fix no-stdio build.
5225 *David Woodhouse <David.Woodhouse@intel.com> and also*
5226 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5228 * New testing framework
5229 The testing framework has been largely rewritten and is now using
5230 perl and the perl modules Test::Harness and an extended variant of
5231 Test::More called OpenSSL::Test to do its work. All test scripts in
5232 test/ have been rewritten into test recipes, and all direct calls to
5233 executables in test/Makefile have become individual recipes using the
5234 simplified testing OpenSSL::Test::Simple.
5236 For documentation on our testing modules, do:
5238 perldoc test/testlib/OpenSSL/Test/Simple.pm
5239 perldoc test/testlib/OpenSSL/Test.pm
5243 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5244 are used; the latter aborts on memory leaks (usually checked on exit).
5245 Some undocumented "set malloc, etc., hooks" functions were removed
5246 and others were changed. All are now documented.
5250 * In DSA_generate_parameters_ex, if the provided seed is too short,
5253 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5255 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5256 from RFC4279, RFC4785, RFC5487, RFC5489.
5258 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5259 original RSA_PSK patch.
5263 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5264 era flag was never set throughout the codebase (only read). Also removed
5265 SSL3_FLAGS_POP_BUFFER which was only used if
5266 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5270 * Changed the default name options in the "ca", "crl", "req" and "x509"
5271 to be "oneline" instead of "compat".
5275 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5276 not aware of clients that still exhibit this bug, and the workaround
5277 hasn't been working properly for a while.
5281 * The return type of BIO_number_read() and BIO_number_written() as well as
5282 the corresponding num_read and num_write members in the BIO structure has
5283 changed from unsigned long to uint64_t. On platforms where an unsigned
5284 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5289 * Given the pervasive nature of TLS extensions it is inadvisable to run
5290 OpenSSL without support for them. It also means that maintaining
5291 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5292 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5296 * Removed support for the two export grade static DH ciphersuites
5297 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5298 were newly added (along with a number of other static DH ciphersuites) to
5299 1.0.2. However the two export ones have *never* worked since they were
5300 introduced. It seems strange in any case to be adding new export
5301 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5305 * Version negotiation has been rewritten. In particular SSLv23_method(),
5306 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5307 and turned into macros which simply call the new preferred function names
5308 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5309 should use the new names instead. Also as part of this change the ssl23.h
5310 header file has been removed.
5314 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5315 code and the associated standard is no longer considered fit-for-purpose.
5319 * RT2547 was closed. When generating a private key, try to make the
5320 output file readable only by the owner. This behavior change might
5321 be noticeable when interacting with other software.
5323 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5328 * Added HTTP GET support to the ocsp command.
5332 * Changed default digest for the dgst and enc commands from MD5 to
5337 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5341 * Added support for TLS extended master secret from
5342 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5343 initial patch which was a great help during development.
5347 * All libssl internal structures have been removed from the public header
5348 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5349 now redundant). Users should not attempt to access internal structures
5350 directly. Instead they should use the provided API functions.
5354 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5355 Access to deprecated functions can be re-enabled by running config with
5356 "enable-deprecated". In addition applications wishing to use deprecated
5357 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5358 will, by default, disable some transitive includes that previously existed
5359 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5363 * Added support for OCB mode. OpenSSL has been granted a patent license
5364 compatible with the OpenSSL license for use of OCB. Details are available
5365 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5366 for OCB can be removed by calling config with no-ocb.
5370 * SSLv2 support has been removed. It still supports receiving an SSLv2
5371 compatible client hello.
5375 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5376 done while fixing the error code for the key-too-small case.
5378 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5380 * CA.sh has been removed; use CA.pl instead.
5384 * Removed old DES API.
5388 * Remove various unsupported platforms:
5394 Sinix/ReliantUNIX RM400
5399 16-bit platforms such as WIN16
5403 * Clean up OPENSSL_NO_xxx #define's
5404 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5405 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5406 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5407 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5408 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5409 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5410 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5411 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5412 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5413 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5417 * Cleaned up dead code
5418 Remove all but one '#ifdef undef' which is to be looked at.
5422 * Clean up calling of xxx_free routines.
5423 Just like free(), fix most of the xxx_free routines to accept
5424 NULL. Remove the non-null checks from callers. Save much code.
5428 * Add secure heap for storage of private keys (when possible).
5429 Add BIO_s_secmem(), CBIGNUM, etc.
5430 Contributed by Akamai Technologies under our Corporate CLA.
5434 * Experimental support for a new, fast, unbiased prime candidate generator,
5435 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5437 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5439 * New output format NSS in the sess_id command line tool. This allows
5440 exporting the session id and the master key in NSS keylog format.
5442 *Martin Kaiser <martin@kaiser.cx>*
5444 * Harmonize version and its documentation. -f flag is used to display
5447 *mancha <mancha1@zoho.com>*
5449 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5450 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5452 *mancha <mancha1@zoho.com>*
5454 * Fix some double frees. These are not thought to be exploitable.
5456 *mancha <mancha1@zoho.com>*
5458 * A missing bounds check in the handling of the TLS heartbeat extension
5459 can be used to reveal up to 64k of memory to a connected client or
5462 Thanks for Neel Mehta of Google Security for discovering this bug and to
5463 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5464 preparing the fix ([CVE-2014-0160])
5466 *Adam Langley, Bodo Moeller*
5468 * Fix for the attack described in the paper "Recovering OpenSSL
5469 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5470 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5471 <http://eprint.iacr.org/2014/140>
5473 Thanks to Yuval Yarom and Naomi Benger for discovering this
5474 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5476 *Yuval Yarom and Naomi Benger*
5478 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5479 this fixes a limitation in previous versions of OpenSSL.
5483 * Experimental encrypt-then-mac support.
5485 Experimental support for encrypt then mac from
5486 draft-gutmann-tls-encrypt-then-mac-02.txt
5488 To enable it set the appropriate extension number (0x42 for the test
5489 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5491 For non-compliant peers (i.e. just about everything) this should have no
5494 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5498 * Add EVP support for key wrapping algorithms, to avoid problems with
5499 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5500 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5501 algorithms and include tests cases.
5505 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5510 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5511 MGF1 digest and OAEP label.
5515 * Make openssl verify return errors.
5517 *Chris Palmer <palmer@google.com> and Ben Laurie*
5519 * New function ASN1_TIME_diff to calculate the difference between two
5520 ASN1_TIME structures or one structure and the current time.
5524 * Update fips_test_suite to support multiple command line options. New
5525 test to induce all self test errors in sequence and check expected
5530 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5531 sign or verify all in one operation.
5535 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5536 test programs and fips_test_suite. Includes functionality to parse
5537 the minimal script output of fipsalgest.pl directly.
5541 * Add authorisation parameter to FIPS_module_mode_set().
5545 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5549 * Use separate DRBG fields for internal and external flags. New function
5550 FIPS_drbg_health_check() to perform on demand health checking. Add
5551 generation tests to fips_test_suite with reduced health check interval to
5552 demonstrate periodic health checking. Add "nodh" option to
5553 fips_test_suite to skip very slow DH test.
5557 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5562 * More extensive health check for DRBG checking many more failure modes.
5563 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5564 combination: call this in fips_test_suite.
5568 * Add support for canonical generation of DSA parameter 'g'. See
5571 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5572 POST to handle HMAC cases.
5576 * Add functions FIPS_module_version() and FIPS_module_version_text()
5577 to return numerical and string versions of the FIPS module number.
5581 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5582 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5583 outside the validated module in the FIPS capable OpenSSL.
5587 * Minor change to DRBG entropy callback semantics. In some cases
5588 there is no multiple of the block length between min_len and
5589 max_len. Allow the callback to return more than max_len bytes
5590 of entropy but discard any extra: it is the callback's responsibility
5591 to ensure that the extra data discarded does not impact the
5592 requested amount of entropy.
5596 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5597 information in FIPS186-3, SP800-57 and SP800-131A.
5601 * CCM support via EVP. Interface is very similar to GCM case except we
5602 must supply all data in one chunk (i.e. no update, final) and the
5603 message length must be supplied if AAD is used. Add algorithm test
5608 * Initial version of POST overhaul. Add POST callback to allow the status
5609 of POST to be monitored and/or failures induced. Modify fips_test_suite
5610 to use callback. Always run all selftests even if one fails.
5614 * XTS support including algorithm test driver in the fips_gcmtest program.
5615 Note: this does increase the maximum key length from 32 to 64 bytes but
5616 there should be no binary compatibility issues as existing applications
5617 will never use XTS mode.
5621 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5622 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5623 performs algorithm blocking for unapproved PRNG types. Also do not
5624 set PRNG type in FIPS_mode_set(): leave this to the application.
5625 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5626 the standard OpenSSL PRNG: set additional data to a date time vector.
5630 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5631 This shouldn't present any incompatibility problems because applications
5632 shouldn't be using these directly and any that are will need to rethink
5633 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5637 * Extensive self tests and health checking required by SP800-90 DRBG.
5638 Remove strength parameter from FIPS_drbg_instantiate and always
5639 instantiate at maximum supported strength.
5643 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5647 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5651 * New function DH_compute_key_padded() to compute a DH key and pad with
5652 leading zeroes if needed: this complies with SP800-56A et al.
5656 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5657 anything, incomplete, subject to change and largely untested at present.
5661 * Modify fipscanisteronly build option to only build the necessary object
5662 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5666 * Add experimental option FIPSSYMS to give all symbols in
5667 fipscanister.o and FIPS or fips prefix. This will avoid
5668 conflicts with future versions of OpenSSL. Add perl script
5669 util/fipsas.pl to preprocess assembly language source files
5670 and rename any affected symbols.
5674 * Add selftest checks and algorithm block of non-fips algorithms in
5675 FIPS mode. Remove DES2 from selftests.
5679 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5680 return internal method without any ENGINE dependencies. Add new
5681 tiny fips sign and verify functions.
5685 * New build option no-ec2m to disable characteristic 2 code.
5689 * New build option "fipscanisteronly". This only builds fipscanister.o
5690 and (currently) associated fips utilities. Uses the file Makefile.fips
5691 instead of Makefile.org as the prototype.
5695 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5696 Update fips_gcmtest to use IV generator.
5700 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5701 setting output buffer to NULL. The `*Final` function must be
5702 called although it will not retrieve any additional data. The tag
5703 can be set or retrieved with a ctrl. The IV length is by default 12
5704 bytes (96 bits) but can be set to an alternative value. If the IV
5705 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5710 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5711 underlying do_cipher function handles all cipher semantics itself
5712 including padding and finalisation. This is useful if (for example)
5713 an ENGINE cipher handles block padding itself. The behaviour of
5714 do_cipher is subtly changed if this flag is set: the return value
5715 is the number of characters written to the output buffer (zero is
5716 no longer an error code) or a negative error code. Also if the
5717 input buffer is NULL and length 0 finalisation should be performed.
5721 * If a candidate issuer certificate is already part of the constructed
5722 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5726 * Improve forward-security support: add functions
5728 void SSL_CTX_set_not_resumable_session_callback(
5729 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5730 void SSL_set_not_resumable_session_callback(
5731 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5733 for use by SSL/TLS servers; the callback function will be called whenever a
5734 new session is created, and gets to decide whether the session may be
5735 cached to make it resumable (return 0) or not (return 1). (As by the
5736 SSL/TLS protocol specifications, the session_id sent by the server will be
5737 empty to indicate that the session is not resumable; also, the server will
5738 not generate RFC 4507 (RFC 5077) session tickets.)
5740 A simple reasonable callback implementation is to return is_forward_secure.
5741 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5742 by the SSL/TLS server library, indicating whether it can provide forward
5745 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5747 * New -verify_name option in command line utilities to set verification
5752 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5753 Add CMAC pkey methods.
5757 * Experimental renegotiation in s_server -www mode. If the client
5758 browses /reneg connection is renegotiated. If /renegcert it is
5759 renegotiated requesting a certificate.
5763 * Add an "external" session cache for debugging purposes to s_server. This
5764 should help trace issues which normally are only apparent in deployed
5765 multi-process servers.
5769 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5770 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5771 BIO_set_cipher() and some obscure PEM functions were changed so they
5772 can now return an error. The RAND changes required a change to the
5773 RAND_METHOD structure.
5777 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5778 a gcc attribute to warn if the result of a function is ignored. This
5779 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5780 whose return value is often ignored.
5784 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5785 These allow SCTs (signed certificate timestamps) to be requested and
5786 validated when establishing a connection.
5788 *Rob Percival <robpercival@google.com>*
5793 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5795 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5796 used even when parsing explicit parameters, when loading a encoded key
5797 or calling `EC_GROUP_new_from_ecpkparameters()`/
5798 `EC_GROUP_new_from_ecparameters()`.
5799 This prevents bypass of security hardening and performance gains,
5800 especially for curves with specialized EC_METHODs.
5801 By default, if a key encoded with explicit parameters is loaded and later
5802 encoded, the output is still encoded with explicit parameters, even if
5803 internally a "named" EC_GROUP is used for computation.
5807 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5808 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5809 NULL. After this change, only the cofactor parameter can be NULL. It also
5810 does some minimal sanity checks on the passed order.
5815 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5816 An attack is simple, if the first CMS_recipientInfo is valid but the
5817 second CMS_recipientInfo is chosen ciphertext. If the second
5818 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5819 encryption key will be replaced by garbage, and the message cannot be
5820 decoded, but if the RSA decryption fails, the correct encryption key is
5821 used and the recipient will not notice the attack.
5822 As a work around for this potential attack the length of the decrypted
5823 key must be equal to the cipher default key length, in case the
5824 certificate is not given and all recipientInfo are tried out.
5825 The old behaviour can be re-enabled in the CMS code by setting the
5826 CMS_DEBUG_DECRYPT flag.
5831 * Document issue with installation paths in diverse Windows builds
5833 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5834 binaries and run-time config file.
5839 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5841 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5842 This changes the size when using the `genpkey` command when no size is given.
5843 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5844 generation commands to use 2048 bits by default.
5848 * Add FIPS support for Android Arm 64-bit
5850 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5851 Module in Version 2.0.10. For some reason, the corresponding target
5852 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5853 built with FIPS support on Android Arm 64-bit. This omission has been
5856 *Matthias St. Pierre*
5858 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5860 * 0-byte record padding oracle
5862 If an application encounters a fatal protocol error and then calls
5863 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5864 then OpenSSL can respond differently to the calling application if a 0 byte
5865 record is received with invalid padding compared to if a 0 byte record is
5866 received with an invalid MAC. If the application then behaves differently
5867 based on that in a way that is detectable to the remote peer, then this
5868 amounts to a padding oracle that could be used to decrypt data.
5870 In order for this to be exploitable "non-stitched" ciphersuites must be in
5871 use. Stitched ciphersuites are optimised implementations of certain
5872 commonly used ciphersuites. Also the application must call SSL_shutdown()
5873 twice even if a protocol error has occurred (applications should not do
5874 this but some do anyway).
5876 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5877 Aviram, with additional investigation by Steven Collison and Andrew
5878 Hourselt. It was reported to OpenSSL on 10th December 2018.
5883 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5887 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5889 * Microarchitecture timing vulnerability in ECC scalar multiplication
5891 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5892 shown to be vulnerable to a microarchitecture timing side channel attack.
5893 An attacker with sufficient access to mount local timing attacks during
5894 ECDSA signature generation could recover the private key.
5896 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5897 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5903 * Timing vulnerability in DSA signature generation
5905 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5906 timing side channel attack. An attacker could use variations in the signing
5907 algorithm to recover the private key.
5909 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5914 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5915 Module, accidentally introduced while backporting security fixes from the
5916 development branch and hindering the use of ECC in FIPS mode.
5920 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5922 * Client DoS due to large DH parameter
5924 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5925 malicious server can send a very large prime value to the client. This will
5926 cause the client to spend an unreasonably long period of time generating a
5927 key for this prime resulting in a hang until the client has finished. This
5928 could be exploited in a Denial Of Service attack.
5930 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5935 * Cache timing vulnerability in RSA Key Generation
5937 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5938 a cache timing side channel attack. An attacker with sufficient access to
5939 mount cache timing attacks during the RSA key generation process could
5940 recover the private key.
5942 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5943 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5948 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5949 parameter is no longer accepted, as it leads to a corrupt table. NULL
5950 pem_str is reserved for alias entries only.
5954 * Revert blinding in ECDSA sign and instead make problematic addition
5955 length-invariant. Switch even to fixed-length Montgomery multiplication.
5959 * Change generating and checking of primes so that the error rate of not
5960 being prime depends on the intended use based on the size of the input.
5961 For larger primes this will result in more rounds of Miller-Rabin.
5962 The maximal error rate for primes with more than 1080 bits is lowered
5965 *Kurt Roeckx, Annie Yousar*
5967 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5971 * Add blinding to ECDSA and DSA signatures to protect against side channel
5972 attacks discovered by Keegan Ryan (NCC Group).
5976 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5977 now allow empty (zero character) pass phrases.
5981 * Certificate time validation (X509_cmp_time) enforces stricter
5982 compliance with RFC 5280. Fractional seconds and timezone offsets
5983 are no longer allowed.
5987 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5989 * Constructed ASN.1 types with a recursive definition could exceed the stack
5991 Constructed ASN.1 types with a recursive definition (such as can be found
5992 in PKCS7) could eventually exceed the stack given malicious input with
5993 excessive recursion. This could result in a Denial Of Service attack. There
5994 are no such structures used within SSL/TLS that come from untrusted sources
5995 so this is considered safe.
5997 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6003 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6005 * Read/write after SSL object in error state
6007 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6008 mechanism. The intent was that if a fatal error occurred during a handshake
6009 then OpenSSL would move into the error state and would immediately fail if
6010 you attempted to continue the handshake. This works as designed for the
6011 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6012 SSL_connect()), however due to a bug it does not work correctly if
6013 SSL_read() or SSL_write() is called directly. In that scenario, if the
6014 handshake fails then a fatal error will be returned in the initial function
6015 call. If SSL_read()/SSL_write() is subsequently called by the application
6016 for the same SSL object then it will succeed and the data is passed without
6017 being decrypted/encrypted directly from the SSL/TLS record layer.
6019 In order to exploit this issue an application bug would have to be present
6020 that resulted in a call to SSL_read()/SSL_write() being issued after having
6021 already received a fatal error.
6023 This issue was reported to OpenSSL by David Benjamin (Google).
6028 * rsaz_1024_mul_avx2 overflow bug on x86_64
6030 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6031 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6032 Analysis suggests that attacks against RSA and DSA as a result of this
6033 defect would be very difficult to perform and are not believed likely.
6034 Attacks against DH1024 are considered just feasible, because most of the
6035 work necessary to deduce information about a private key may be performed
6036 offline. The amount of resources required for such an attack would be
6037 significant. However, for an attack on TLS to be meaningful, the server
6038 would have to share the DH1024 private key among multiple clients, which is
6039 no longer an option since CVE-2016-0701.
6041 This only affects processors that support the AVX2 but not ADX extensions
6042 like Intel Haswell (4th generation).
6044 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6045 was originally found via the OSS-Fuzz project.
6050 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6052 * bn_sqrx8x_internal carry bug on x86_64
6054 There is a carry propagating bug in the x86_64 Montgomery squaring
6055 procedure. No EC algorithms are affected. Analysis suggests that attacks
6056 against RSA and DSA as a result of this defect would be very difficult to
6057 perform and are not believed likely. Attacks against DH are considered just
6058 feasible (although very difficult) because most of the work necessary to
6059 deduce information about a private key may be performed offline. The amount
6060 of resources required for such an attack would be very significant and
6061 likely only accessible to a limited number of attackers. An attacker would
6062 additionally need online access to an unpatched system using the target
6063 private key in a scenario with persistent DH parameters and a private
6064 key that is shared between multiple clients.
6066 This only affects processors that support the BMI1, BMI2 and ADX extensions
6067 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6069 This issue was reported to OpenSSL by the OSS-Fuzz project.
6074 * Malformed X.509 IPAddressFamily could cause OOB read
6076 If an X.509 certificate has a malformed IPAddressFamily extension,
6077 OpenSSL could do a one-byte buffer overread. The most likely result
6078 would be an erroneous display of the certificate in text format.
6080 This issue was reported to OpenSSL by the OSS-Fuzz project.
6084 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6086 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6087 platform rather than 'mingw'.
6091 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6093 * Truncated packet could crash via OOB read
6095 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6096 cipher is being used, then a truncated packet can cause that host to
6097 perform an out-of-bounds read, usually resulting in a crash.
6099 This issue was reported to OpenSSL by Robert Święcki of Google.
6104 * BN_mod_exp may produce incorrect results on x86_64
6106 There is a carry propagating bug in the x86_64 Montgomery squaring
6107 procedure. No EC algorithms are affected. Analysis suggests that attacks
6108 against RSA and DSA as a result of this defect would be very difficult to
6109 perform and are not believed likely. Attacks against DH are considered just
6110 feasible (although very difficult) because most of the work necessary to
6111 deduce information about a private key may be performed offline. The amount
6112 of resources required for such an attack would be very significant and
6113 likely only accessible to a limited number of attackers. An attacker would
6114 additionally need online access to an unpatched system using the target
6115 private key in a scenario with persistent DH parameters and a private
6116 key that is shared between multiple clients. For example this can occur by
6117 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6118 similar to CVE-2015-3193 but must be treated as a separate problem.
6120 This issue was reported to OpenSSL by the OSS-Fuzz project.
6125 * Montgomery multiplication may produce incorrect results
6127 There is a carry propagating bug in the Broadwell-specific Montgomery
6128 multiplication procedure that handles input lengths divisible by, but
6129 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6130 and DH private keys are impossible. This is because the subroutine in
6131 question is not used in operations with the private key itself and an input
6132 of the attacker's direct choice. Otherwise the bug can manifest itself as
6133 transient authentication and key negotiation failures or reproducible
6134 erroneous outcome of public-key operations with specially crafted input.
6135 Among EC algorithms only Brainpool P-512 curves are affected and one
6136 presumably can attack ECDH key negotiation. Impact was not analyzed in
6137 detail, because pre-requisites for attack are considered unlikely. Namely
6138 multiple clients have to choose the curve in question and the server has to
6139 share the private key among them, neither of which is default behaviour.
6140 Even then only clients that chose the curve will be affected.
6142 This issue was publicly reported as transient failures and was not
6143 initially recognized as a security issue. Thanks to Richard Morgan for
6144 providing reproducible case.
6149 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6150 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6151 prevent issues where no progress is being made and the peer continually
6152 sends unrecognised record types, using up resources processing them.
6156 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6158 * Missing CRL sanity check
6160 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6161 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6162 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6164 This issue only affects the OpenSSL 1.0.2i
6169 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6171 * OCSP Status Request extension unbounded memory growth
6173 A malicious client can send an excessively large OCSP Status Request
6174 extension. If that client continually requests renegotiation, sending a
6175 large OCSP Status Request extension each time, then there will be unbounded
6176 memory growth on the server. This will eventually lead to a Denial Of
6177 Service attack through memory exhaustion. Servers with a default
6178 configuration are vulnerable even if they do not support OCSP. Builds using
6179 the "no-ocsp" build time option are not affected.
6181 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6186 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6189 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6195 * OOB write in MDC2_Update()
6197 An overflow can occur in MDC2_Update() either if called directly or
6198 through the EVP_DigestUpdate() function using MDC2. If an attacker
6199 is able to supply very large amounts of input data after a previous
6200 call to EVP_EncryptUpdate() with a partial block then a length check
6201 can overflow resulting in a heap corruption.
6203 The amount of data needed is comparable to SIZE_MAX which is impractical
6206 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6211 * Malformed SHA512 ticket DoS
6213 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6214 DoS attack where a malformed ticket will result in an OOB read which will
6217 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6218 a custom server callback and ticket lookup mechanism.
6220 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6225 * OOB write in BN_bn2dec()
6227 The function BN_bn2dec() does not check the return value of BN_div_word().
6228 This can cause an OOB write if an application uses this function with an
6229 overly large BIGNUM. This could be a problem if an overly large certificate
6230 or CRL is printed out from an untrusted source. TLS is not affected because
6231 record limits will reject an oversized certificate before it is parsed.
6233 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6238 * OOB read in TS_OBJ_print_bio()
6240 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6241 the total length the OID text representation would use and not the amount
6242 of data written. This will result in OOB reads when large OIDs are
6245 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6250 * Pointer arithmetic undefined behaviour
6252 Avoid some undefined pointer arithmetic
6254 A common idiom in the codebase is to check limits in the following manner:
6257 Where "p" points to some malloc'd data of SIZE bytes and
6260 "len" here could be from some externally supplied data (e.g. from a TLS
6263 The rules of C pointer arithmetic are such that "p + len" is only well
6264 defined where len <= SIZE. Therefore the above idiom is actually
6265 undefined behaviour.
6267 For example this could cause problems if some malloc implementation
6268 provides an address for "p" such that "p + len" actually overflows for
6269 values of len that are too big and therefore p + len < limit.
6271 This issue was reported to OpenSSL by Guido Vranken
6276 * Constant time flag not preserved in DSA signing
6278 Operations in the DSA signing algorithm should run in constant time in
6279 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6280 implementation means that a non-constant time codepath is followed for
6281 certain operations. This has been demonstrated through a cache-timing
6282 attack to be sufficient for an attacker to recover the private DSA key.
6284 This issue was reported by César Pereida (Aalto University), Billy Brumley
6285 (Tampere University of Technology), and Yuval Yarom (The University of
6286 Adelaide and NICTA).
6291 * DTLS buffered message DoS
6293 In a DTLS connection where handshake messages are delivered out-of-order
6294 those messages that OpenSSL is not yet ready to process will be buffered
6295 for later use. Under certain circumstances, a flaw in the logic means that
6296 those messages do not get removed from the buffer even though the handshake
6297 has been completed. An attacker could force up to approx. 15 messages to
6298 remain in the buffer when they are no longer required. These messages will
6299 be cleared when the DTLS connection is closed. The default maximum size for
6300 a message is 100k. Therefore, the attacker could force an additional 1500k
6301 to be consumed per connection. By opening many simultaneous connections an
6302 attacker could cause a DoS attack through memory exhaustion.
6304 This issue was reported to OpenSSL by Quan Luo.
6309 * DTLS replay protection DoS
6311 A flaw in the DTLS replay attack protection mechanism means that records
6312 that arrive for future epochs update the replay protection "window" before
6313 the MAC for the record has been validated. This could be exploited by an
6314 attacker by sending a record for the next epoch (which does not have to
6315 decrypt or have a valid MAC), with a very large sequence number. This means
6316 that all subsequent legitimate packets are dropped causing a denial of
6317 service for a specific DTLS connection.
6319 This issue was reported to OpenSSL by the OCAP audit team.
6324 * Certificate message OOB reads
6326 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6327 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6328 theoretical DoS risk but this has not been observed in practice on common
6331 The messages affected are client certificate, client certificate request
6332 and server certificate. As a result the attack can only be performed
6333 against a client or a server which enables client authentication.
6335 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6340 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6342 * Prevent padding oracle in AES-NI CBC MAC check
6344 A MITM attacker can use a padding oracle attack to decrypt traffic
6345 when the connection uses an AES CBC cipher and the server support
6348 This issue was introduced as part of the fix for Lucky 13 padding
6349 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6350 constant time by making sure that always the same bytes are read and
6351 compared against either the MAC or padding bytes. But it no longer
6352 checked that there was enough data to have both the MAC and padding
6355 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6359 * Fix EVP_EncodeUpdate overflow
6361 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6362 Base64 encoding of binary data. If an attacker is able to supply very large
6363 amounts of input data then a length check can overflow resulting in a heap
6366 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6367 the `PEM_write_bio*` family of functions. These are mainly used within the
6368 OpenSSL command line applications, so any application which processes data
6369 from an untrusted source and outputs it as a PEM file should be considered
6370 vulnerable to this issue. User applications that call these APIs directly
6371 with large amounts of untrusted data may also be vulnerable.
6373 This issue was reported by Guido Vranken.
6378 * Fix EVP_EncryptUpdate overflow
6380 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6381 is able to supply very large amounts of input data after a previous call to
6382 EVP_EncryptUpdate() with a partial block then a length check can overflow
6383 resulting in a heap corruption. Following an analysis of all OpenSSL
6384 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6385 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6386 the first called function after an EVP_EncryptInit(), and therefore that
6387 specific call must be safe. The second form is where the length passed to
6388 EVP_EncryptUpdate() can be seen from the code to be some small value and
6389 therefore there is no possibility of an overflow. Since all instances are
6390 one of these two forms, it is believed that there can be no overflows in
6391 internal code due to this problem. It should be noted that
6392 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6393 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6394 of these calls have also been analysed too and it is believed there are no
6395 instances in internal usage where an overflow could occur.
6397 This issue was reported by Guido Vranken.
6402 * Prevent ASN.1 BIO excessive memory allocation
6404 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6405 a short invalid encoding can cause allocation of large amounts of memory
6406 potentially consuming excessive resources or exhausting memory.
6408 Any application parsing untrusted data through d2i BIO functions is
6409 affected. The memory based functions such as d2i_X509() are *not* affected.
6410 Since the memory based functions are used by the TLS library, TLS
6411 applications are not affected.
6413 This issue was reported by Brian Carpenter.
6420 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6421 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6422 in arbitrary stack data being returned in the buffer.
6424 This issue was reported by Guido Vranken.
6429 * Modify behavior of ALPN to invoke callback after SNI/servername
6430 callback, such that updates to the SSL_CTX affect ALPN.
6434 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6439 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6440 methods are enabled and ssl2 is disabled the methods return NULL.
6444 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6446 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6447 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6448 provide any "EXPORT" or "LOW" strength ciphers.
6452 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6453 is by default disabled at build-time. Builds that are not configured with
6454 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6455 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6456 will need to explicitly call either of:
6458 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6460 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6462 as appropriate. Even if either of those is used, or the application
6463 explicitly uses the version-specific SSLv2_method() or its client and
6464 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6465 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6466 ciphers, and SSLv2 56-bit DES are no longer available.
6471 * Fix a double-free in DSA code
6473 A double free bug was discovered when OpenSSL parses malformed DSA private
6474 keys and could lead to a DoS attack or memory corruption for applications
6475 that receive DSA private keys from untrusted sources. This scenario is
6478 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6484 * Disable SRP fake user seed to address a server memory leak.
6486 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6488 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6489 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6490 was changed to ignore the "fake user" SRP seed, even if the seed
6493 Users should use SRP_VBASE_get1_by_user instead. Note that in
6494 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6495 also that even though configuring the SRP seed attempts to hide
6496 invalid usernames by continuing the handshake with fake
6497 credentials, this behaviour is not constant time and no strong
6498 guarantees are made that the handshake is indistinguishable from
6499 that of a valid user.
6504 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6506 In the BN_hex2bn function the number of hex digits is calculated using an
6507 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6508 large values of `i` this can result in `bn_expand` not allocating any
6509 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6510 field as NULL leading to a subsequent NULL ptr deref. For very large values
6511 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6512 In this case memory is allocated to the internal BIGNUM data field, but it
6513 is insufficiently sized leading to heap corruption. A similar issue exists
6514 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6515 is ever called by user applications with very large untrusted hex/dec data.
6516 This is anticipated to be a rare occurrence.
6518 All OpenSSL internal usage of these functions use data that is not expected
6519 to be untrusted, e.g. config file data or application command line
6520 arguments. If user developed applications generate config file data based
6521 on untrusted data then it is possible that this could also lead to security
6522 consequences. This is also anticipated to be rare.
6524 This issue was reported to OpenSSL by Guido Vranken.
6529 * Fix memory issues in `BIO_*printf` functions
6531 The internal `fmtstr` function used in processing a "%s" format string in
6532 the `BIO_*printf` functions could overflow while calculating the length of a
6533 string and cause an OOB read when printing very long strings.
6535 Additionally the internal `doapr_outch` function can attempt to write to an
6536 OOB memory location (at an offset from the NULL pointer) in the event of a
6537 memory allocation failure. In 1.0.2 and below this could be caused where
6538 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6539 could be in processing a very long "%s" format string. Memory leaks can
6542 The first issue may mask the second issue dependent on compiler behaviour.
6543 These problems could enable attacks where large amounts of untrusted data
6544 is passed to the `BIO_*printf` functions. If applications use these functions
6545 in this way then they could be vulnerable. OpenSSL itself uses these
6546 functions when printing out human-readable dumps of ASN.1 data. Therefore
6547 applications that print this data could be vulnerable if the data is from
6548 untrusted sources. OpenSSL command line applications could also be
6549 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6550 as command line arguments.
6552 Libssl is not considered directly vulnerable. Additionally certificates etc
6553 received via remote connections via libssl are also unlikely to be able to
6554 trigger these issues because of message size limits enforced within libssl.
6556 This issue was reported to OpenSSL Guido Vranken.
6561 * Side channel attack on modular exponentiation
6563 A side-channel attack was found which makes use of cache-bank conflicts on
6564 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6565 of RSA keys. The ability to exploit this issue is limited as it relies on
6566 an attacker who has control of code in a thread running on the same
6567 hyper-threaded core as the victim thread which is performing decryptions.
6569 This issue was reported to OpenSSL by Yuval Yarom, The University of
6570 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6571 Nadia Heninger, University of Pennsylvania with more information at
6572 <http://cachebleed.info>.
6577 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6578 if no keysize is specified with default_bits. This fixes an
6579 omission in an earlier change that changed all RSA/DSA key generation
6580 commands to use 2048 bits by default.
6584 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6586 * DH small subgroups
6588 Historically OpenSSL only ever generated DH parameters based on "safe"
6589 primes. More recently (in version 1.0.2) support was provided for
6590 generating X9.42 style parameter files such as those required for RFC 5114
6591 support. The primes used in such files may not be "safe". Where an
6592 application is using DH configured with parameters based on primes that are
6593 not "safe" then an attacker could use this fact to find a peer's private
6594 DH exponent. This attack requires that the attacker complete multiple
6595 handshakes in which the peer uses the same private DH exponent. For example
6596 this could be used to discover a TLS server's private DH exponent if it's
6597 reusing the private DH exponent or it's using a static DH ciphersuite.
6599 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6600 TLS. It is not on by default. If the option is not set then the server
6601 reuses the same private DH exponent for the life of the server process and
6602 would be vulnerable to this attack. It is believed that many popular
6603 applications do set this option and would therefore not be at risk.
6605 The fix for this issue adds an additional check where a "q" parameter is
6606 available (as is the case in X9.42 based parameters). This detects the
6607 only known attack, and is the only possible defense for static DH
6608 ciphersuites. This could have some performance impact.
6610 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6611 default and cannot be disabled. This could have some performance impact.
6613 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6618 * SSLv2 doesn't block disabled ciphers
6620 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6621 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6622 been disabled, provided that the SSLv2 protocol was not also disabled via
6625 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6626 and Sebastian Schinzel.
6631 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6633 * BN_mod_exp may produce incorrect results on x86_64
6635 There is a carry propagating bug in the x86_64 Montgomery squaring
6636 procedure. No EC algorithms are affected. Analysis suggests that attacks
6637 against RSA and DSA as a result of this defect would be very difficult to
6638 perform and are not believed likely. Attacks against DH are considered just
6639 feasible (although very difficult) because most of the work necessary to
6640 deduce information about a private key may be performed offline. The amount
6641 of resources required for such an attack would be very significant and
6642 likely only accessible to a limited number of attackers. An attacker would
6643 additionally need online access to an unpatched system using the target
6644 private key in a scenario with persistent DH parameters and a private
6645 key that is shared between multiple clients. For example this can occur by
6646 default in OpenSSL DHE based SSL/TLS ciphersuites.
6648 This issue was reported to OpenSSL by Hanno Böck.
6653 * Certificate verify crash with missing PSS parameter
6655 The signature verification routines will crash with a NULL pointer
6656 dereference if presented with an ASN.1 signature using the RSA PSS
6657 algorithm and absent mask generation function parameter. Since these
6658 routines are used to verify certificate signature algorithms this can be
6659 used to crash any certificate verification operation and exploited in a
6660 DoS attack. Any application which performs certificate verification is
6661 vulnerable including OpenSSL clients and servers which enable client
6664 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6669 * X509_ATTRIBUTE memory leak
6671 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6672 memory. This structure is used by the PKCS#7 and CMS routines so any
6673 application which reads PKCS#7 or CMS data from untrusted sources is
6674 affected. SSL/TLS is not affected.
6676 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6682 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6683 This changes the decoding behaviour for some invalid messages,
6684 though the change is mostly in the more lenient direction, and
6685 legacy behaviour is preserved as much as possible.
6689 * In DSA_generate_parameters_ex, if the provided seed is too short,
6692 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6694 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6696 * Alternate chains certificate forgery
6698 During certificate verification, OpenSSL will attempt to find an
6699 alternative certificate chain if the first attempt to build such a chain
6700 fails. An error in the implementation of this logic can mean that an
6701 attacker could cause certain checks on untrusted certificates to be
6702 bypassed, such as the CA flag, enabling them to use a valid leaf
6703 certificate to act as a CA and "issue" an invalid certificate.
6705 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6710 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6712 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6713 incompatibility in the handling of HMAC. The previous ABI has now been
6718 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6720 * Malformed ECParameters causes infinite loop
6722 When processing an ECParameters structure OpenSSL enters an infinite loop
6723 if the curve specified is over a specially malformed binary polynomial
6726 This can be used to perform denial of service against any
6727 system which processes public keys, certificate requests or
6728 certificates. This includes TLS clients and TLS servers with
6729 client authentication enabled.
6731 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6736 * Exploitable out-of-bounds read in X509_cmp_time
6738 X509_cmp_time does not properly check the length of the ASN1_TIME
6739 string and can read a few bytes out of bounds. In addition,
6740 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6743 An attacker can use this to craft malformed certificates and CRLs of
6744 various sizes and potentially cause a segmentation fault, resulting in
6745 a DoS on applications that verify certificates or CRLs. TLS clients
6746 that verify CRLs are affected. TLS clients and servers with client
6747 authentication enabled may be affected if they use custom verification
6750 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6751 independently by Hanno Böck.
6756 * PKCS7 crash with missing EnvelopedContent
6758 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6759 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6760 with missing content and trigger a NULL pointer dereference on parsing.
6762 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6763 structures from untrusted sources are affected. OpenSSL clients and
6764 servers are not affected.
6766 This issue was reported to OpenSSL by Michal Zalewski (Google).
6771 * CMS verify infinite loop with unknown hash function
6773 When verifying a signedData message the CMS code can enter an infinite loop
6774 if presented with an unknown hash function OID. This can be used to perform
6775 denial of service against any system which verifies signedData messages using
6777 This issue was reported to OpenSSL by Johannes Bauer.
6782 * Race condition handling NewSessionTicket
6784 If a NewSessionTicket is received by a multi-threaded client when attempting to
6785 reuse a previous ticket then a race condition can occur potentially leading to
6786 a double free of the ticket data.
6791 * Only support 256-bit or stronger elliptic curves with the
6792 'ecdh_auto' setting (server) or by default (client). Of supported
6793 curves, prefer P-256 (both).
6797 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6799 * ClientHello sigalgs DoS fix
6801 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6802 invalid signature algorithms extension a NULL pointer dereference will
6803 occur. This can be exploited in a DoS attack against the server.
6805 This issue was was reported to OpenSSL by David Ramos of Stanford
6809 *Stephen Henson and Matt Caswell*
6811 * Multiblock corrupted pointer fix
6813 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6814 feature only applies on 64 bit x86 architecture platforms that support AES
6815 NI instructions. A defect in the implementation of "multiblock" can cause
6816 OpenSSL's internal write buffer to become incorrectly set to NULL when
6817 using non-blocking IO. Typically, when the user application is using a
6818 socket BIO for writing, this will only result in a failed connection.
6819 However if some other BIO is used then it is likely that a segmentation
6820 fault will be triggered, thus enabling a potential DoS attack.
6822 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6827 * Segmentation fault in DTLSv1_listen fix
6829 The DTLSv1_listen function is intended to be stateless and processes the
6830 initial ClientHello from many peers. It is common for user code to loop
6831 over the call to DTLSv1_listen until a valid ClientHello is received with
6832 an associated cookie. A defect in the implementation of DTLSv1_listen means
6833 that state is preserved in the SSL object from one invocation to the next
6834 that can lead to a segmentation fault. Errors processing the initial
6835 ClientHello can trigger this scenario. An example of such an error could be
6836 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6839 This issue was reported to OpenSSL by Per Allansson.
6844 * Segmentation fault in ASN1_TYPE_cmp fix
6846 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6847 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6848 certificate signature algorithm consistency this can be used to crash any
6849 certificate verification operation and exploited in a DoS attack. Any
6850 application which performs certificate verification is vulnerable including
6851 OpenSSL clients and servers which enable client authentication.
6856 * Segmentation fault for invalid PSS parameters fix
6858 The signature verification routines will crash with a NULL pointer
6859 dereference if presented with an ASN.1 signature using the RSA PSS
6860 algorithm and invalid parameters. Since these routines are used to verify
6861 certificate signature algorithms this can be used to crash any
6862 certificate verification operation and exploited in a DoS attack. Any
6863 application which performs certificate verification is vulnerable including
6864 OpenSSL clients and servers which enable client authentication.
6866 This issue was was reported to OpenSSL by Brian Carpenter.
6871 * ASN.1 structure reuse memory corruption fix
6873 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6874 memory corruption via an invalid write. Such reuse is and has been
6875 strongly discouraged and is believed to be rare.
6877 Applications that parse structures containing CHOICE or ANY DEFINED BY
6878 components may be affected. Certificate parsing (d2i_X509 and related
6879 functions) are however not affected. OpenSSL clients and servers are
6885 * PKCS7 NULL pointer dereferences fix
6887 The PKCS#7 parsing code does not handle missing outer ContentInfo
6888 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6889 missing content and trigger a NULL pointer dereference on parsing.
6891 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6892 otherwise parse PKCS#7 structures from untrusted sources are
6893 affected. OpenSSL clients and servers are not affected.
6895 This issue was reported to OpenSSL by Michal Zalewski (Google).
6900 * DoS via reachable assert in SSLv2 servers fix
6902 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6903 servers that both support SSLv2 and enable export cipher suites by sending
6904 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6906 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6907 (OpenSSL development team).
6912 * Empty CKE with client auth and DHE fix
6914 If client auth is used then a server can seg fault in the event of a DHE
6915 ciphersuite being selected and a zero length ClientKeyExchange message
6916 being sent by the client. This could be exploited in a DoS attack.
6921 * Handshake with unseeded PRNG fix
6923 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6924 with an unseeded PRNG. The conditions are:
6925 - The client is on a platform where the PRNG has not been seeded
6926 automatically, and the user has not seeded manually
6927 - A protocol specific client method version has been used (i.e. not
6928 SSL_client_methodv23)
6929 - A ciphersuite is used that does not require additional random data from
6930 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6932 If the handshake succeeds then the client random that has been used will
6933 have been generated from a PRNG with insufficient entropy and therefore the
6934 output may be predictable.
6936 For example using the following command with an unseeded openssl will
6937 succeed on an unpatched platform:
6939 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6944 * Use After Free following d2i_ECPrivatekey error fix
6946 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6947 could cause a use after free condition. This, in turn, could cause a double
6948 free in several private key parsing functions (such as d2i_PrivateKey
6949 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6950 for applications that receive EC private keys from untrusted
6951 sources. This scenario is considered rare.
6953 This issue was discovered by the BoringSSL project and fixed in their
6959 * X509_to_X509_REQ NULL pointer deref fix
6961 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6962 the certificate key is invalid. This function is rarely used in practice.
6964 This issue was discovered by Brian Carpenter.
6969 * Removed the export ciphers from the DEFAULT ciphers
6973 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6975 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6976 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6977 So far those who have to target multiple platforms would compromise
6978 and argue that binary targeting say ARMv5 would still execute on
6979 ARMv8. "Universal" build resolves this compromise by providing
6980 near-optimal performance even on newer platforms.
6984 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6985 (other platforms pending).
6987 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6989 * Add support for the SignedCertificateTimestampList certificate and
6990 OCSP response extensions from RFC6962.
6994 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6995 for corner cases. (Certain input points at infinity could lead to
6996 bogus results, with non-infinity inputs mapped to infinity too.)
7000 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7001 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7002 common cases are optimized and there still is room for further
7003 improvements. Vector Permutation AES for Altivec is also added.
7007 * Add support for little-endian ppc64 Linux target.
7009 *Marcelo Cerri (IBM)*
7011 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7012 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7013 are optimized and there still is room for further improvements.
7014 Both 32- and 64-bit modes are supported.
7016 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7018 * Improved ARMv7 NEON support.
7022 * Support for SPARC Architecture 2011 crypto extensions, first
7023 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7024 SHA256/512, MD5, GHASH and modular exponentiation.
7026 *Andy Polyakov, David Miller*
7028 * Accelerated modular exponentiation for Intel processors, a.k.a.
7031 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7033 * Support for new and upcoming Intel processors, including AVX2,
7034 BMI and SHA ISA extensions. This includes additional "stitched"
7035 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7038 This work was sponsored by Intel Corp.
7042 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7043 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7044 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7048 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7049 this fixes a limitation in previous versions of OpenSSL.
7053 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7054 MGF1 digest and OAEP label.
7058 * Add EVP support for key wrapping algorithms, to avoid problems with
7059 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7060 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7061 algorithms and include tests cases.
7065 * Add functions to allocate and set the fields of an ECDSA_METHOD
7068 *Douglas E. Engert, Steve Henson*
7070 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7071 difference in days and seconds between two tm or ASN1_TIME structures.
7075 * Add -rev test option to s_server to just reverse order of characters
7076 received by client and send back to server. Also prints an abbreviated
7077 summary of the connection parameters.
7081 * New option -brief for s_client and s_server to print out a brief summary
7082 of connection parameters.
7086 * Add callbacks for arbitrary TLS extensions.
7088 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7090 * New option -crl_download in several openssl utilities to download CRLs
7091 from CRLDP extension in certificates.
7095 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7099 * New function X509_CRL_diff to generate a delta CRL from the difference
7100 of two full CRLs. Add support to "crl" utility.
7104 * New functions to set lookup_crls function and to retrieve
7105 X509_STORE from X509_STORE_CTX.
7109 * Print out deprecated issuer and subject unique ID fields in
7114 * Extend OCSP I/O functions so they can be used for simple general purpose
7115 HTTP as well as OCSP. New wrapper function which can be used to download
7116 CRLs using the OCSP API.
7120 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7124 * `SSL_CONF*` functions. These provide a common framework for application
7125 configuration using configuration files or command lines.
7129 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7130 message callback and prints the results. Needs compile time option
7131 "enable-ssl-trace". New options to s_client and s_server to enable
7136 * New ctrl and macro to retrieve supported points extensions.
7137 Print out extension in s_server and s_client.
7141 * New functions to retrieve certificate signature and signature
7146 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7151 * New Suite B modes for TLS code. These use and enforce the requirements
7152 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7153 only use Suite B curves. The Suite B modes can be set by using the
7154 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7158 * New chain verification flags for Suite B levels of security. Check
7159 algorithms are acceptable when flags are set in X509_verify_cert.
7163 * Make tls1_check_chain return a set of flags indicating checks passed
7164 by a certificate chain. Add additional tests to handle client
7165 certificates: checks for matching certificate type and issuer name
7170 * If an attempt is made to use a signature algorithm not in the peer
7171 preference list abort the handshake. If client has no suitable
7172 signature algorithms in response to a certificate request do not
7173 use the certificate.
7177 * If server EC tmp key is not in client preference list abort handshake.
7181 * Add support for certificate stores in CERT structure. This makes it
7182 possible to have different stores per SSL structure or one store in
7183 the parent SSL_CTX. Include distinct stores for certificate chain
7184 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7185 to build and store a certificate chain in CERT structure: returning
7186 an error if the chain cannot be built: this will allow applications
7187 to test if a chain is correctly configured.
7189 Note: if the CERT based stores are not set then the parent SSL_CTX
7190 store is used to retain compatibility with existing behaviour.
7194 * New function ssl_set_client_disabled to set a ciphersuite disabled
7195 mask based on the current session, check mask when sending client
7196 hello and checking the requested ciphersuite.
7200 * New ctrls to retrieve and set certificate types in a certificate
7201 request message. Print out received values in s_client. If certificate
7202 types is not set with custom values set sensible values based on
7203 supported signature algorithms.
7207 * Support for distinct client and server supported signature algorithms.
7211 * Add certificate callback. If set this is called whenever a certificate
7212 is required by client or server. An application can decide which
7213 certificate chain to present based on arbitrary criteria: for example
7214 supported signature algorithms. Add very simple example to s_server.
7215 This fixes many of the problems and restrictions of the existing client
7216 certificate callback: for example you can now clear an existing
7217 certificate and specify the whole chain.
7221 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7222 the certificate can be used for (if anything). Set valid_flags field
7223 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7224 to have similar checks in it.
7226 Add new "cert_flags" field to CERT structure and include a "strict mode".
7227 This enforces some TLS certificate requirements (such as only permitting
7228 certificate signature algorithms contained in the supported algorithms
7229 extension) which some implementations ignore: this option should be used
7230 with caution as it could cause interoperability issues.
7234 * Update and tidy signature algorithm extension processing. Work out
7235 shared signature algorithms based on preferences and peer algorithms
7236 and print them out in s_client and s_server. Abort handshake if no
7237 shared signature algorithms.
7241 * Add new functions to allow customised supported signature algorithms
7242 for SSL and SSL_CTX structures. Add options to s_client and s_server
7247 * New function SSL_certs_clear() to delete all references to certificates
7248 from an SSL structure. Before this once a certificate had been added
7249 it couldn't be removed.
7253 * Integrate hostname, email address and IP address checking with certificate
7254 verification. New verify options supporting checking in openssl utility.
7258 * Fixes and wildcard matching support to hostname and email checking
7259 functions. Add manual page.
7261 *Florian Weimer (Red Hat Product Security Team)*
7263 * New functions to check a hostname email or IP address against a
7264 certificate. Add options x509 utility to print results of checks against
7269 * Fix OCSP checking.
7271 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7273 * Initial experimental support for explicitly trusted non-root CAs.
7274 OpenSSL still tries to build a complete chain to a root but if an
7275 intermediate CA has a trust setting included that is used. The first
7276 setting is used: whether to trust (e.g., -addtrust option to the x509
7281 * Add -trusted_first option which attempts to find certificates in the
7282 trusted store even if an untrusted chain is also supplied.
7286 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7287 platform support for Linux and Android.
7291 * Support for linux-x32, ILP32 environment in x86_64 framework.
7295 * Experimental multi-implementation support for FIPS capable OpenSSL.
7296 When in FIPS mode the approved implementations are used as normal,
7297 when not in FIPS mode the internal unapproved versions are used instead.
7298 This means that the FIPS capable OpenSSL isn't forced to use the
7299 (often lower performance) FIPS implementations outside FIPS mode.
7303 * Transparently support X9.42 DH parameters when calling
7304 PEM_read_bio_DHparameters. This means existing applications can handle
7305 the new parameter format automatically.
7309 * Initial experimental support for X9.42 DH parameter format: mainly
7310 to support use of 'q' parameter for RFC5114 parameters.
7314 * Add DH parameters from RFC5114 including test data to dhtest.
7318 * Support for automatic EC temporary key parameter selection. If enabled
7319 the most preferred EC parameters are automatically used instead of
7320 hardcoded fixed parameters. Now a server just has to call:
7321 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7322 support ECDH and use the most appropriate parameters.
7326 * Enhance and tidy EC curve and point format TLS extension code. Use
7327 static structures instead of allocation if default values are used.
7328 New ctrls to set curves we wish to support and to retrieve shared curves.
7329 Print out shared curves in s_server. New options to s_server and s_client
7330 to set list of supported curves.
7334 * New ctrls to retrieve supported signature algorithms and
7335 supported curve values as an array of NIDs. Extend openssl utility
7336 to print out received values.
7340 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7341 between NIDs and the more common NIST names such as "P-256". Enhance
7342 ecparam utility and ECC method to recognise the NIST names for curves.
7346 * Enhance SSL/TLS certificate chain handling to support different
7347 chains for each certificate instead of one chain in the parent SSL_CTX.
7351 * Support for fixed DH ciphersuite client authentication: where both
7352 server and client use DH certificates with common parameters.
7356 * Support for fixed DH ciphersuites: those requiring DH server
7361 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7363 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7364 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7365 X509_CINF_get_signature were reverted post internal team review.
7370 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7372 * OCSP Status Request extension unbounded memory growth
7374 A malicious client can send an excessively large OCSP Status Request
7375 extension. If that client continually requests renegotiation, sending a
7376 large OCSP Status Request extension each time, then there will be unbounded
7377 memory growth on the server. This will eventually lead to a Denial Of
7378 Service attack through memory exhaustion. Servers with a default
7379 configuration are vulnerable even if they do not support OCSP. Builds using
7380 the "no-ocsp" build time option are not affected.
7382 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7387 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7390 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7396 * OOB write in MDC2_Update()
7398 An overflow can occur in MDC2_Update() either if called directly or
7399 through the EVP_DigestUpdate() function using MDC2. If an attacker
7400 is able to supply very large amounts of input data after a previous
7401 call to EVP_EncryptUpdate() with a partial block then a length check
7402 can overflow resulting in a heap corruption.
7404 The amount of data needed is comparable to SIZE_MAX which is impractical
7407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7412 * Malformed SHA512 ticket DoS
7414 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7415 DoS attack where a malformed ticket will result in an OOB read which will
7418 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7419 a custom server callback and ticket lookup mechanism.
7421 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7426 * OOB write in BN_bn2dec()
7428 The function BN_bn2dec() does not check the return value of BN_div_word().
7429 This can cause an OOB write if an application uses this function with an
7430 overly large BIGNUM. This could be a problem if an overly large certificate
7431 or CRL is printed out from an untrusted source. TLS is not affected because
7432 record limits will reject an oversized certificate before it is parsed.
7434 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7439 * OOB read in TS_OBJ_print_bio()
7441 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7442 the total length the OID text representation would use and not the amount
7443 of data written. This will result in OOB reads when large OIDs are
7446 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7451 * Pointer arithmetic undefined behaviour
7453 Avoid some undefined pointer arithmetic
7455 A common idiom in the codebase is to check limits in the following manner:
7458 Where "p" points to some malloc'd data of SIZE bytes and
7461 "len" here could be from some externally supplied data (e.g. from a TLS
7464 The rules of C pointer arithmetic are such that "p + len" is only well
7465 defined where len <= SIZE. Therefore, the above idiom is actually
7466 undefined behaviour.
7468 For example this could cause problems if some malloc implementation
7469 provides an address for "p" such that "p + len" actually overflows for
7470 values of len that are too big and therefore p + len < limit.
7472 This issue was reported to OpenSSL by Guido Vranken
7477 * Constant time flag not preserved in DSA signing
7479 Operations in the DSA signing algorithm should run in constant time in
7480 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7481 implementation means that a non-constant time codepath is followed for
7482 certain operations. This has been demonstrated through a cache-timing
7483 attack to be sufficient for an attacker to recover the private DSA key.
7485 This issue was reported by César Pereida (Aalto University), Billy Brumley
7486 (Tampere University of Technology), and Yuval Yarom (The University of
7487 Adelaide and NICTA).
7492 * DTLS buffered message DoS
7494 In a DTLS connection where handshake messages are delivered out-of-order
7495 those messages that OpenSSL is not yet ready to process will be buffered
7496 for later use. Under certain circumstances, a flaw in the logic means that
7497 those messages do not get removed from the buffer even though the handshake
7498 has been completed. An attacker could force up to approx. 15 messages to
7499 remain in the buffer when they are no longer required. These messages will
7500 be cleared when the DTLS connection is closed. The default maximum size for
7501 a message is 100k. Therefore, the attacker could force an additional 1500k
7502 to be consumed per connection. By opening many simultaneous connections an
7503 attacker could cause a DoS attack through memory exhaustion.
7505 This issue was reported to OpenSSL by Quan Luo.
7510 * DTLS replay protection DoS
7512 A flaw in the DTLS replay attack protection mechanism means that records
7513 that arrive for future epochs update the replay protection "window" before
7514 the MAC for the record has been validated. This could be exploited by an
7515 attacker by sending a record for the next epoch (which does not have to
7516 decrypt or have a valid MAC), with a very large sequence number. This means
7517 that all subsequent legitimate packets are dropped causing a denial of
7518 service for a specific DTLS connection.
7520 This issue was reported to OpenSSL by the OCAP audit team.
7525 * Certificate message OOB reads
7527 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7528 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7529 theoretical DoS risk but this has not been observed in practice on common
7532 The messages affected are client certificate, client certificate request
7533 and server certificate. As a result the attack can only be performed
7534 against a client or a server which enables client authentication.
7536 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7541 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7543 * Prevent padding oracle in AES-NI CBC MAC check
7545 A MITM attacker can use a padding oracle attack to decrypt traffic
7546 when the connection uses an AES CBC cipher and the server support
7549 This issue was introduced as part of the fix for Lucky 13 padding
7550 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7551 constant time by making sure that always the same bytes are read and
7552 compared against either the MAC or padding bytes. But it no longer
7553 checked that there was enough data to have both the MAC and padding
7556 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7561 * Fix EVP_EncodeUpdate overflow
7563 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7564 Base64 encoding of binary data. If an attacker is able to supply very large
7565 amounts of input data then a length check can overflow resulting in a heap
7568 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7569 the `PEM_write_bio*` family of functions. These are mainly used within the
7570 OpenSSL command line applications, so any application which processes data
7571 from an untrusted source and outputs it as a PEM file should be considered
7572 vulnerable to this issue. User applications that call these APIs directly
7573 with large amounts of untrusted data may also be vulnerable.
7575 This issue was reported by Guido Vranken.
7580 * Fix EVP_EncryptUpdate overflow
7582 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7583 is able to supply very large amounts of input data after a previous call to
7584 EVP_EncryptUpdate() with a partial block then a length check can overflow
7585 resulting in a heap corruption. Following an analysis of all OpenSSL
7586 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7587 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7588 the first called function after an EVP_EncryptInit(), and therefore that
7589 specific call must be safe. The second form is where the length passed to
7590 EVP_EncryptUpdate() can be seen from the code to be some small value and
7591 therefore there is no possibility of an overflow. Since all instances are
7592 one of these two forms, it is believed that there can be no overflows in
7593 internal code due to this problem. It should be noted that
7594 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7595 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7596 of these calls have also been analysed too and it is believed there are no
7597 instances in internal usage where an overflow could occur.
7599 This issue was reported by Guido Vranken.
7604 * Prevent ASN.1 BIO excessive memory allocation
7606 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7607 a short invalid encoding can casuse allocation of large amounts of memory
7608 potentially consuming excessive resources or exhausting memory.
7610 Any application parsing untrusted data through d2i BIO functions is
7611 affected. The memory based functions such as d2i_X509() are *not* affected.
7612 Since the memory based functions are used by the TLS library, TLS
7613 applications are not affected.
7615 This issue was reported by Brian Carpenter.
7622 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7623 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7624 in arbitrary stack data being returned in the buffer.
7626 This issue was reported by Guido Vranken.
7631 * Modify behavior of ALPN to invoke callback after SNI/servername
7632 callback, such that updates to the SSL_CTX affect ALPN.
7636 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7641 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7642 methods are enabled and ssl2 is disabled the methods return NULL.
7646 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7648 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7649 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7650 provide any "EXPORT" or "LOW" strength ciphers.
7654 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7655 is by default disabled at build-time. Builds that are not configured with
7656 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7657 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7658 will need to explicitly call either of:
7660 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7662 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7664 as appropriate. Even if either of those is used, or the application
7665 explicitly uses the version-specific SSLv2_method() or its client and
7666 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7667 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7668 ciphers, and SSLv2 56-bit DES are no longer available.
7673 * Fix a double-free in DSA code
7675 A double free bug was discovered when OpenSSL parses malformed DSA private
7676 keys and could lead to a DoS attack or memory corruption for applications
7677 that receive DSA private keys from untrusted sources. This scenario is
7680 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7686 * Disable SRP fake user seed to address a server memory leak.
7688 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7690 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7691 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7692 was changed to ignore the "fake user" SRP seed, even if the seed
7695 Users should use SRP_VBASE_get1_by_user instead. Note that in
7696 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7697 also that even though configuring the SRP seed attempts to hide
7698 invalid usernames by continuing the handshake with fake
7699 credentials, this behaviour is not constant time and no strong
7700 guarantees are made that the handshake is indistinguishable from
7701 that of a valid user.
7706 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7708 In the BN_hex2bn function the number of hex digits is calculated using an
7709 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7710 large values of `i` this can result in `bn_expand` not allocating any
7711 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7712 field as NULL leading to a subsequent NULL ptr deref. For very large values
7713 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7714 In this case memory is allocated to the internal BIGNUM data field, but it
7715 is insufficiently sized leading to heap corruption. A similar issue exists
7716 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7717 is ever called by user applications with very large untrusted hex/dec data.
7718 This is anticipated to be a rare occurrence.
7720 All OpenSSL internal usage of these functions use data that is not expected
7721 to be untrusted, e.g. config file data or application command line
7722 arguments. If user developed applications generate config file data based
7723 on untrusted data then it is possible that this could also lead to security
7724 consequences. This is also anticipated to be rare.
7726 This issue was reported to OpenSSL by Guido Vranken.
7731 * Fix memory issues in `BIO_*printf` functions
7733 The internal `fmtstr` function used in processing a "%s" format string in
7734 the `BIO_*printf` functions could overflow while calculating the length of a
7735 string and cause an OOB read when printing very long strings.
7737 Additionally the internal `doapr_outch` function can attempt to write to an
7738 OOB memory location (at an offset from the NULL pointer) in the event of a
7739 memory allocation failure. In 1.0.2 and below this could be caused where
7740 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7741 could be in processing a very long "%s" format string. Memory leaks can
7744 The first issue may mask the second issue dependent on compiler behaviour.
7745 These problems could enable attacks where large amounts of untrusted data
7746 is passed to the `BIO_*printf` functions. If applications use these functions
7747 in this way then they could be vulnerable. OpenSSL itself uses these
7748 functions when printing out human-readable dumps of ASN.1 data. Therefore
7749 applications that print this data could be vulnerable if the data is from
7750 untrusted sources. OpenSSL command line applications could also be
7751 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7752 as command line arguments.
7754 Libssl is not considered directly vulnerable. Additionally certificates etc
7755 received via remote connections via libssl are also unlikely to be able to
7756 trigger these issues because of message size limits enforced within libssl.
7758 This issue was reported to OpenSSL Guido Vranken.
7763 * Side channel attack on modular exponentiation
7765 A side-channel attack was found which makes use of cache-bank conflicts on
7766 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7767 of RSA keys. The ability to exploit this issue is limited as it relies on
7768 an attacker who has control of code in a thread running on the same
7769 hyper-threaded core as the victim thread which is performing decryptions.
7771 This issue was reported to OpenSSL by Yuval Yarom, The University of
7772 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7773 Nadia Heninger, University of Pennsylvania with more information at
7774 <http://cachebleed.info>.
7779 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7780 if no keysize is specified with default_bits. This fixes an
7781 omission in an earlier change that changed all RSA/DSA key generation
7782 commands to use 2048 bits by default.
7786 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7788 * Protection for DH small subgroup attacks
7790 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7791 switched on by default and cannot be disabled. This could have some
7796 * SSLv2 doesn't block disabled ciphers
7798 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7799 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7800 been disabled, provided that the SSLv2 protocol was not also disabled via
7803 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7804 and Sebastian Schinzel.
7809 * Reject DH handshakes with parameters shorter than 1024 bits.
7813 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7815 * Certificate verify crash with missing PSS parameter
7817 The signature verification routines will crash with a NULL pointer
7818 dereference if presented with an ASN.1 signature using the RSA PSS
7819 algorithm and absent mask generation function parameter. Since these
7820 routines are used to verify certificate signature algorithms this can be
7821 used to crash any certificate verification operation and exploited in a
7822 DoS attack. Any application which performs certificate verification is
7823 vulnerable including OpenSSL clients and servers which enable client
7826 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7831 * X509_ATTRIBUTE memory leak
7833 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7834 memory. This structure is used by the PKCS#7 and CMS routines so any
7835 application which reads PKCS#7 or CMS data from untrusted sources is
7836 affected. SSL/TLS is not affected.
7838 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7844 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7845 This changes the decoding behaviour for some invalid messages,
7846 though the change is mostly in the more lenient direction, and
7847 legacy behaviour is preserved as much as possible.
7851 * In DSA_generate_parameters_ex, if the provided seed is too short,
7852 use a random seed, as already documented.
7854 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7856 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7858 * Alternate chains certificate forgery
7860 During certificate verification, OpenSSL will attempt to find an
7861 alternative certificate chain if the first attempt to build such a chain
7862 fails. An error in the implementation of this logic can mean that an
7863 attacker could cause certain checks on untrusted certificates to be
7864 bypassed, such as the CA flag, enabling them to use a valid leaf
7865 certificate to act as a CA and "issue" an invalid certificate.
7867 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7873 * Race condition handling PSK identify hint
7875 If PSK identity hints are received by a multi-threaded client then
7876 the values are wrongly updated in the parent SSL_CTX structure. This can
7877 result in a race condition potentially leading to a double free of the
7883 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7885 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7886 incompatibility in the handling of HMAC. The previous ABI has now been
7889 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7891 * Malformed ECParameters causes infinite loop
7893 When processing an ECParameters structure OpenSSL enters an infinite loop
7894 if the curve specified is over a specially malformed binary polynomial
7897 This can be used to perform denial of service against any
7898 system which processes public keys, certificate requests or
7899 certificates. This includes TLS clients and TLS servers with
7900 client authentication enabled.
7902 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7907 * Exploitable out-of-bounds read in X509_cmp_time
7909 X509_cmp_time does not properly check the length of the ASN1_TIME
7910 string and can read a few bytes out of bounds. In addition,
7911 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7914 An attacker can use this to craft malformed certificates and CRLs of
7915 various sizes and potentially cause a segmentation fault, resulting in
7916 a DoS on applications that verify certificates or CRLs. TLS clients
7917 that verify CRLs are affected. TLS clients and servers with client
7918 authentication enabled may be affected if they use custom verification
7921 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7922 independently by Hanno Böck.
7927 * PKCS7 crash with missing EnvelopedContent
7929 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7930 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7931 with missing content and trigger a NULL pointer dereference on parsing.
7933 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7934 structures from untrusted sources are affected. OpenSSL clients and
7935 servers are not affected.
7937 This issue was reported to OpenSSL by Michal Zalewski (Google).
7942 * CMS verify infinite loop with unknown hash function
7944 When verifying a signedData message the CMS code can enter an infinite loop
7945 if presented with an unknown hash function OID. This can be used to perform
7946 denial of service against any system which verifies signedData messages using
7948 This issue was reported to OpenSSL by Johannes Bauer.
7953 * Race condition handling NewSessionTicket
7955 If a NewSessionTicket is received by a multi-threaded client when attempting to
7956 reuse a previous ticket then a race condition can occur potentially leading to
7957 a double free of the ticket data.
7962 * Reject DH handshakes with parameters shorter than 768 bits.
7964 *Kurt Roeckx and Emilia Kasper*
7966 * dhparam: generate 2048-bit parameters by default.
7968 *Kurt Roeckx and Emilia Kasper*
7970 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7972 * Segmentation fault in ASN1_TYPE_cmp fix
7974 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7975 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7976 certificate signature algorithm consistency this can be used to crash any
7977 certificate verification operation and exploited in a DoS attack. Any
7978 application which performs certificate verification is vulnerable including
7979 OpenSSL clients and servers which enable client authentication.
7984 * ASN.1 structure reuse memory corruption fix
7986 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7987 memory corruption via an invalid write. Such reuse is and has been
7988 strongly discouraged and is believed to be rare.
7990 Applications that parse structures containing CHOICE or ANY DEFINED BY
7991 components may be affected. Certificate parsing (d2i_X509 and related
7992 functions) are however not affected. OpenSSL clients and servers are
7998 * PKCS7 NULL pointer dereferences fix
8000 The PKCS#7 parsing code does not handle missing outer ContentInfo
8001 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8002 missing content and trigger a NULL pointer dereference on parsing.
8004 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8005 otherwise parse PKCS#7 structures from untrusted sources are
8006 affected. OpenSSL clients and servers are not affected.
8008 This issue was reported to OpenSSL by Michal Zalewski (Google).
8013 * DoS via reachable assert in SSLv2 servers fix
8015 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8016 servers that both support SSLv2 and enable export cipher suites by sending
8017 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8019 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8020 (OpenSSL development team).
8025 * Use After Free following d2i_ECPrivatekey error fix
8027 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8028 could cause a use after free condition. This, in turn, could cause a double
8029 free in several private key parsing functions (such as d2i_PrivateKey
8030 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8031 for applications that receive EC private keys from untrusted
8032 sources. This scenario is considered rare.
8034 This issue was discovered by the BoringSSL project and fixed in their
8040 * X509_to_X509_REQ NULL pointer deref fix
8042 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8043 the certificate key is invalid. This function is rarely used in practice.
8045 This issue was discovered by Brian Carpenter.
8050 * Removed the export ciphers from the DEFAULT ciphers
8054 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8056 * Build fixes for the Windows and OpenVMS platforms
8058 *Matt Caswell and Richard Levitte*
8060 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8062 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8063 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8064 dereference. This could lead to a Denial Of Service attack. Thanks to
8065 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8070 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8071 dtls1_buffer_record function under certain conditions. In particular this
8072 could occur if an attacker sent repeated DTLS records with the same
8073 sequence number but for the next epoch. The memory leak could be exploited
8074 by an attacker in a Denial of Service attack through memory exhaustion.
8075 Thanks to Chris Mueller for reporting this issue.
8080 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8081 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8082 method would be set to NULL which could later result in a NULL pointer
8083 dereference. Thanks to Frank Schmirler for reporting this issue.
8088 * Abort handshake if server key exchange message is omitted for ephemeral
8091 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8092 reporting this issue.
8097 * Remove non-export ephemeral RSA code on client and server. This code
8098 violated the TLS standard by allowing the use of temporary RSA keys in
8099 non-export ciphersuites and could be used by a server to effectively
8100 downgrade the RSA key length used to a value smaller than the server
8101 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8102 INRIA or reporting this issue.
8107 * Fixed issue where DH client certificates are accepted without verification.
8108 An OpenSSL server will accept a DH certificate for client authentication
8109 without the certificate verify message. This effectively allows a client to
8110 authenticate without the use of a private key. This only affects servers
8111 which trust a client certificate authority which issues certificates
8112 containing DH keys: these are extremely rare and hardly ever encountered.
8113 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8119 * Ensure that the session ID context of an SSL is updated when its
8120 SSL_CTX is updated via SSL_set_SSL_CTX.
8122 The session ID context is typically set from the parent SSL_CTX,
8123 and can vary with the CTX.
8127 * Fix various certificate fingerprint issues.
8129 By using non-DER or invalid encodings outside the signed portion of a
8130 certificate the fingerprint can be changed without breaking the signature.
8131 Although no details of the signed portion of the certificate can be changed
8132 this can cause problems with some applications: e.g. those using the
8133 certificate fingerprint for blacklists.
8135 1. Reject signatures with non zero unused bits.
8137 If the BIT STRING containing the signature has non zero unused bits reject
8138 the signature. All current signature algorithms require zero unused bits.
8140 2. Check certificate algorithm consistency.
8142 Check the AlgorithmIdentifier inside TBS matches the one in the
8143 certificate signature. NB: this will result in signature failure
8144 errors for some broken certificates.
8146 Thanks to Konrad Kraszewski from Google for reporting this issue.
8148 3. Check DSA/ECDSA signatures use DER.
8150 Re-encode DSA/ECDSA signatures and compare with the original received
8151 signature. Return an error if there is a mismatch.
8153 This will reject various cases including garbage after signature
8154 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8155 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8156 (negative or with leading zeroes).
8158 Further analysis was conducted and fixes were developed by Stephen Henson
8159 of the OpenSSL core team.
8165 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8166 results on some platforms, including x86_64. This bug occurs at random
8167 with a very low probability, and is not known to be exploitable in any
8168 way, though its exact impact is difficult to determine. Thanks to Pieter
8169 Wuille (Blockstream) who reported this issue and also suggested an initial
8170 fix. Further analysis was conducted by the OpenSSL development team and
8171 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8172 the OpenSSL core team.
8177 * Do not resume sessions on the server if the negotiated protocol
8178 version does not match the session's version. Resuming with a different
8179 version, while not strictly forbidden by the RFC, is of questionable
8180 sanity and breaks all known clients.
8182 *David Benjamin, Emilia Käsper*
8184 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8185 early CCS messages during renegotiation. (Note that because
8186 renegotiation is encrypted, this early CCS was not exploitable.)
8190 * Tighten client-side session ticket handling during renegotiation:
8191 ensure that the client only accepts a session ticket if the server sends
8192 the extension anew in the ServerHello. Previously, a TLS client would
8193 reuse the old extension state and thus accept a session ticket if one was
8194 announced in the initial ServerHello.
8196 Similarly, ensure that the client requires a session ticket if one
8197 was advertised in the ServerHello. Previously, a TLS client would
8198 ignore a missing NewSessionTicket message.
8202 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8206 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8207 sends a carefully crafted handshake message, to cause OpenSSL to fail
8208 to free up to 64k of memory causing a memory leak. This could be
8209 exploited in a Denial Of Service attack. This issue affects OpenSSL
8210 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8211 whether SRTP is used or configured. Implementations of OpenSSL that
8212 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8214 The fix was developed by the OpenSSL team.
8219 * Session Ticket Memory Leak.
8221 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8222 integrity of that ticket is first verified. In the event of a session
8223 ticket integrity check failing, OpenSSL will fail to free memory
8224 causing a memory leak. By sending a large number of invalid session
8225 tickets an attacker could exploit this issue in a Denial Of Service
8231 * Build option no-ssl3 is incomplete.
8233 When OpenSSL is configured with "no-ssl3" as a build option, servers
8234 could accept and complete an SSL 3.0 handshake, and clients could be
8235 configured to send them.
8238 *Akamai and the OpenSSL team*
8240 * Add support for TLS_FALLBACK_SCSV.
8241 Client applications doing fallback retries should call
8242 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8245 *Adam Langley, Bodo Moeller*
8247 * Add additional DigestInfo checks.
8249 Re-encode DigestInto in DER and check against the original when
8250 verifying RSA signature: this will reject any improperly encoded
8251 DigestInfo structures.
8253 Note: this is a precautionary measure and no attacks are currently known.
8257 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8259 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8260 SRP code can be overrun an internal buffer. Add sanity check that
8261 g, A, B < N to SRP code.
8263 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8264 Group for discovering this issue.
8269 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8270 TLS 1.0 instead of higher protocol versions when the ClientHello message
8271 is badly fragmented. This allows a man-in-the-middle attacker to force a
8272 downgrade to TLS 1.0 even if both the server and the client support a
8273 higher protocol version, by modifying the client's TLS records.
8275 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8276 researching this issue.
8281 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8282 to a denial of service attack. A malicious server can crash the client
8283 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8284 ciphersuite and sending carefully crafted handshake messages.
8286 Thanks to Felix Gröbert (Google) for discovering and researching this
8292 * By sending carefully crafted DTLS packets an attacker could cause openssl
8293 to leak memory. This can be exploited through a Denial of Service attack.
8294 Thanks to Adam Langley for discovering and researching this issue.
8299 * An attacker can force openssl to consume large amounts of memory whilst
8300 processing DTLS handshake messages. This can be exploited through a
8301 Denial of Service attack.
8302 Thanks to Adam Langley for discovering and researching this issue.
8307 * An attacker can force an error condition which causes openssl to crash
8308 whilst processing DTLS packets due to memory being freed twice. This
8309 can be exploited through a Denial of Service attack.
8310 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8316 * If a multithreaded client connects to a malicious server using a resumed
8317 session and the server sends an ec point format extension it could write
8318 up to 255 bytes to freed memory.
8320 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8326 * A malicious server can crash an OpenSSL client with a null pointer
8327 dereference (read) by specifying an SRP ciphersuite even though it was not
8328 properly negotiated with the client. This can be exploited through a
8329 Denial of Service attack.
8331 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8332 discovering and researching this issue.
8337 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8338 X509_name_oneline, X509_name_print_ex et al. to leak some information
8339 from the stack. Applications may be affected if they echo pretty printing
8340 output to the attacker.
8342 Thanks to Ivan Fratric (Google) for discovering this issue.
8345 *Emilia Käsper, and Steve Henson*
8347 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8348 for corner cases. (Certain input points at infinity could lead to
8349 bogus results, with non-infinity inputs mapped to infinity too.)
8353 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8355 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8356 handshake can force the use of weak keying material in OpenSSL
8357 SSL/TLS clients and servers.
8359 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8360 researching this issue. ([CVE-2014-0224])
8362 *KIKUCHI Masashi, Steve Henson*
8364 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8365 OpenSSL DTLS client the code can be made to recurse eventually crashing
8368 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8371 *Imre Rad, Steve Henson*
8373 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8374 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8375 client or server. This is potentially exploitable to run arbitrary
8376 code on a vulnerable client or server.
8378 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8380 *Jüri Aedla, Steve Henson*
8382 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8383 are subject to a denial of service attack.
8385 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8386 this issue. ([CVE-2014-3470])
8388 *Felix Gröbert, Ivan Fratric, Steve Henson*
8390 * Harmonize version and its documentation. -f flag is used to display
8393 *mancha <mancha1@zoho.com>*
8395 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8396 in i2d_ECPrivateKey.
8398 *mancha <mancha1@zoho.com>*
8400 * Fix some double frees. These are not thought to be exploitable.
8402 *mancha <mancha1@zoho.com>*
8404 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8406 * A missing bounds check in the handling of the TLS heartbeat extension
8407 can be used to reveal up to 64k of memory to a connected client or
8410 Thanks for Neel Mehta of Google Security for discovering this bug and to
8411 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8412 preparing the fix ([CVE-2014-0160])
8414 *Adam Langley, Bodo Moeller*
8416 * Fix for the attack described in the paper "Recovering OpenSSL
8417 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8418 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8419 <http://eprint.iacr.org/2014/140>
8421 Thanks to Yuval Yarom and Naomi Benger for discovering this
8422 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8424 *Yuval Yarom and Naomi Benger*
8426 * TLS pad extension: draft-agl-tls-padding-03
8428 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8429 TLS client Hello record length value would otherwise be > 255 and
8430 less that 512 pad with a dummy extension containing zeroes so it
8431 is at least 512 bytes long.
8433 *Adam Langley, Steve Henson*
8435 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8437 * Fix for TLS record tampering bug. A carefully crafted invalid
8438 handshake could crash OpenSSL with a NULL pointer exception.
8439 Thanks to Anton Johansson for reporting this issues.
8442 * Keep original DTLS digest and encryption contexts in retransmission
8443 structures so we can use the previous session parameters if they need
8444 to be resent. ([CVE-2013-6450])
8448 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8449 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8450 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8451 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8452 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8453 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8455 *Rob Stradling, Adam Langley*
8457 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8459 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8460 supporting platforms or when small records were transferred.
8462 *Andy Polyakov, Steve Henson*
8464 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8466 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8468 This addresses the flaw in CBC record processing discovered by
8469 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8470 at: <http://www.isg.rhul.ac.uk/tls/>
8472 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8473 Security Group at Royal Holloway, University of London
8474 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8475 Emilia Käsper for the initial patch.
8478 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8480 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8481 ciphersuites which can be exploited in a denial of service attack.
8482 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8483 and detecting this bug and to Wolfgang Ettlinger
8484 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8489 * Return an error when checking OCSP signatures when key is NULL.
8490 This fixes a DoS attack. ([CVE-2013-0166])
8494 * Make openssl verify return errors.
8496 *Chris Palmer <palmer@google.com> and Ben Laurie*
8498 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8499 the right response is stapled. Also change SSL_get_certificate()
8500 so it returns the certificate actually sent.
8501 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8503 *Rob Stradling <rob.stradling@comodo.com>*
8505 * Fix possible deadlock when decoding public keys.
8509 * Don't use TLS 1.0 record version number in initial client hello
8514 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8516 * Sanity check record length before skipping explicit IV in TLS
8517 1.2, 1.1 and DTLS to fix DoS attack.
8519 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8520 fuzzing as a service testing platform.
8525 * Initialise tkeylen properly when encrypting CMS messages.
8526 Thanks to Solar Designer of Openwall for reporting this issue.
8530 * In FIPS mode don't try to use composite ciphers as they are not
8535 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8537 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8538 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8539 mean any application compiled against OpenSSL 1.0.0 headers setting
8540 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8541 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8542 0x10000000L Any application which was previously compiled against
8543 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8544 will need to be recompiled as a result. Letting be results in
8545 inability to disable specifically TLS 1.1 and in client context,
8546 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8550 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8551 disable just protocol X, but all protocols above X *if* there are
8552 protocols *below* X still enabled. In more practical terms it means
8553 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8554 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8555 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8560 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8562 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8563 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8564 in CRYPTO_realloc_clean.
8566 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8567 issue and to Adam Langley <agl@chromium.org> for fixing it.
8570 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8572 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8576 * Workarounds for some broken servers that "hang" if a client hello
8577 record length exceeds 255 bytes.
8579 1. Do not use record version number > TLS 1.0 in initial client
8580 hello: some (but not all) hanging servers will now work.
8581 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8582 the number of ciphers sent in the client hello. This should be
8583 set to an even number, such as 50, for example by passing:
8584 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8585 Most broken servers should now work.
8586 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8587 TLS 1.2 client support entirely.
8591 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8595 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8597 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8598 STRING form instead of a DigestInfo.
8602 * The format used for MDC2 RSA signatures is inconsistent between EVP
8603 and the RSA_sign/RSA_verify functions. This was made more apparent when
8604 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8605 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8606 the correct format in RSA_verify so both forms transparently work.
8610 * Some servers which support TLS 1.0 can choke if we initially indicate
8611 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8612 encrypted premaster secret. As a workaround use the maximum permitted
8613 client version in client hello, this should keep such servers happy
8614 and still work with previous versions of OpenSSL.
8618 * Add support for TLS/DTLS heartbeats.
8620 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8622 * Add support for SCTP.
8624 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8626 * Improved PRNG seeding for VOS.
8628 *Paul Green <Paul.Green@stratus.com>*
8630 * Extensive assembler packs updates, most notably:
8632 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8633 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8634 - x86_64: bit-sliced AES implementation;
8635 - ARM: NEON support, contemporary platforms optimizations;
8636 - s390x: z196 support;
8637 - `*`: GHASH and GF(2^m) multiplication implementations;
8641 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8642 (removal of unnecessary code)
8644 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8646 * Add TLS key material exporter from RFC 5705.
8650 * Add DTLS-SRTP negotiation from RFC 5764.
8654 * Add Next Protocol Negotiation,
8655 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8656 disabled with a no-npn flag to config or Configure. Code donated
8659 *Adam Langley <agl@google.com> and Ben Laurie*
8661 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8662 NIST-P256, NIST-P521, with constant-time single point multiplication on
8663 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8664 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8665 Code made available under Apache License version 2.0.
8667 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8668 line to include this in your build of OpenSSL, and run "make depend" (or
8669 "make update"). This enables the following EC_METHODs:
8671 EC_GFp_nistp224_method()
8672 EC_GFp_nistp256_method()
8673 EC_GFp_nistp521_method()
8675 EC_GROUP_new_by_curve_name() will automatically use these (while
8676 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8679 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8681 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8682 all platforms. Move ssize_t definition from e_os.h to the public
8683 header file e_os2.h as it now appears in public header file cms.h
8687 * New -sigopt option to the ca, req and x509 utilities. Additional
8688 signature parameters can be passed using this option and in
8693 * Add RSA PSS signing function. This will generate and set the
8694 appropriate AlgorithmIdentifiers for PSS based on those in the
8695 corresponding EVP_MD_CTX structure. No application support yet.
8699 * Support for companion algorithm specific ASN1 signing routines.
8700 New function ASN1_item_sign_ctx() signs a pre-initialised
8701 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8702 the appropriate parameters.
8706 * Add new algorithm specific ASN1 verification initialisation function
8707 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8708 handling will be the same no matter what EVP_PKEY_METHOD is used.
8709 Add a PSS handler to support verification of PSS signatures: checked
8710 against a number of sample certificates.
8714 * Add signature printing for PSS. Add PSS OIDs.
8716 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8718 * Add algorithm specific signature printing. An individual ASN1 method
8719 can now print out signatures instead of the standard hex dump.
8721 More complex signatures (e.g. PSS) can print out more meaningful
8722 information. Include DSA version that prints out the signature
8727 * Password based recipient info support for CMS library: implementing
8732 * Split password based encryption into PBES2 and PBKDF2 functions. This
8733 neatly separates the code into cipher and PBE sections and is required
8734 for some algorithms that split PBES2 into separate pieces (such as
8735 password based CMS).
8739 * Session-handling fixes:
8740 - Fix handling of connections that are resuming with a session ID,
8741 but also support Session Tickets.
8742 - Fix a bug that suppressed issuing of a new ticket if the client
8743 presented a ticket with an expired session.
8744 - Try to set the ticket lifetime hint to something reasonable.
8745 - Make tickets shorter by excluding irrelevant information.
8746 - On the client side, don't ignore renewed tickets.
8748 *Adam Langley, Bodo Moeller (Google)*
8750 * Fix PSK session representation.
8754 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8756 This work was sponsored by Intel.
8760 * Add GCM support to TLS library. Some custom code is needed to split
8761 the IV between the fixed (from PRF) and explicit (from TLS record)
8762 portions. This adds all GCM ciphersuites supported by RFC5288 and
8763 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8764 add a special AESGCM string for GCM only.
8768 * Expand range of ctrls for AES GCM. Permit setting invocation
8769 field on decrypt and retrieval of invocation field only on encrypt.
8773 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8774 As required by RFC5289 these ciphersuites cannot be used if for
8775 versions of TLS earlier than 1.2.
8779 * For FIPS capable OpenSSL interpret a NULL default public key method
8780 as unset and return the appropriate default but do *not* set the default.
8781 This means we can return the appropriate method in applications that
8782 switch between FIPS and non-FIPS modes.
8786 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8787 ENGINE is used then we cannot handle that in the FIPS module so we
8788 keep original code iff non-FIPS operations are allowed.
8792 * Add -attime option to openssl utilities.
8794 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8796 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8800 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8801 FIPS EC methods unconditionally for now.
8805 * New build option no-ec2m to disable characteristic 2 code.
8809 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8810 all cases can be covered as some introduce binary incompatibilities.
8814 * Redirect RSA operations to FIPS module including keygen,
8815 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8819 * Add similar low-level API blocking to ciphers.
8823 * low-level digest APIs are not approved in FIPS mode: any attempt
8824 to use these will cause a fatal error. Applications that *really* want
8825 to use them can use the `private_*` version instead.
8829 * Redirect cipher operations to FIPS module for FIPS builds.
8833 * Redirect digest operations to FIPS module for FIPS builds.
8837 * Update build system to add "fips" flag which will link in fipscanister.o
8838 for static and shared library builds embedding a signature if needed.
8842 * Output TLS supported curves in preference order instead of numerical
8843 order. This is currently hardcoded for the highest order curves first.
8844 This should be configurable so applications can judge speed vs strength.
8848 * Add TLS v1.2 server support for client authentication.
8852 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8857 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8858 FIPS modules versions.
8862 * Add TLS v1.2 client side support for client authentication. Keep cache
8863 of handshake records longer as we don't know the hash algorithm to use
8864 until after the certificate request message is received.
8868 * Initial TLS v1.2 client support. Add a default signature algorithms
8869 extension including all the algorithms we support. Parse new signature
8870 format in client key exchange. Relax some ECC signing restrictions for
8871 TLS v1.2 as indicated in RFC5246.
8875 * Add server support for TLS v1.2 signature algorithms extension. Switch
8876 to new signature format when needed using client digest preference.
8877 All server ciphersuites should now work correctly in TLS v1.2. No client
8878 support yet and no support for client certificates.
8882 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8883 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8884 ciphersuites. At present only RSA key exchange ciphersuites work with
8885 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8886 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8887 and version checking.
8891 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8892 with this defined it will not be affected by any changes to ssl internal
8893 structures. Add several utility functions to allow openssl application
8894 to work with OPENSSL_NO_SSL_INTERN defined.
8898 * A long standing patch to add support for SRP from EdelWeb (Peter
8899 Sylvester and Christophe Renou) was integrated.
8900 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8901 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8904 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8908 * Permit abbreviated handshakes when renegotiating using the function
8909 SSL_renegotiate_abbreviated().
8911 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8913 * Add call to ENGINE_register_all_complete() to
8914 ENGINE_load_builtin_engines(), so some implementations get used
8915 automatically instead of needing explicit application support.
8919 * Add support for TLS key exporter as described in RFC5705.
8921 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8923 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8924 a few changes are required:
8926 Add SSL_OP_NO_TLSv1_1 flag.
8927 Add TLSv1_1 methods.
8928 Update version checking logic to handle version 1.1.
8929 Add explicit IV handling (ported from DTLS code).
8930 Add command line options to s_client/s_server.
8937 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8939 * X509_ATTRIBUTE memory leak
8941 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8942 memory. This structure is used by the PKCS#7 and CMS routines so any
8943 application which reads PKCS#7 or CMS data from untrusted sources is
8944 affected. SSL/TLS is not affected.
8946 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8952 * Race condition handling PSK identify hint
8954 If PSK identity hints are received by a multi-threaded client then
8955 the values are wrongly updated in the parent SSL_CTX structure. This can
8956 result in a race condition potentially leading to a double free of the
8962 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8964 * Malformed ECParameters causes infinite loop
8966 When processing an ECParameters structure OpenSSL enters an infinite loop
8967 if the curve specified is over a specially malformed binary polynomial
8970 This can be used to perform denial of service against any
8971 system which processes public keys, certificate requests or
8972 certificates. This includes TLS clients and TLS servers with
8973 client authentication enabled.
8975 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8980 * Exploitable out-of-bounds read in X509_cmp_time
8982 X509_cmp_time does not properly check the length of the ASN1_TIME
8983 string and can read a few bytes out of bounds. In addition,
8984 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8987 An attacker can use this to craft malformed certificates and CRLs of
8988 various sizes and potentially cause a segmentation fault, resulting in
8989 a DoS on applications that verify certificates or CRLs. TLS clients
8990 that verify CRLs are affected. TLS clients and servers with client
8991 authentication enabled may be affected if they use custom verification
8994 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8995 independently by Hanno Böck.
9000 * PKCS7 crash with missing EnvelopedContent
9002 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9003 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9004 with missing content and trigger a NULL pointer dereference on parsing.
9006 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9007 structures from untrusted sources are affected. OpenSSL clients and
9008 servers are not affected.
9010 This issue was reported to OpenSSL by Michal Zalewski (Google).
9015 * CMS verify infinite loop with unknown hash function
9017 When verifying a signedData message the CMS code can enter an infinite loop
9018 if presented with an unknown hash function OID. This can be used to perform
9019 denial of service against any system which verifies signedData messages using
9021 This issue was reported to OpenSSL by Johannes Bauer.
9026 * Race condition handling NewSessionTicket
9028 If a NewSessionTicket is received by a multi-threaded client when attempting to
9029 reuse a previous ticket then a race condition can occur potentially leading to
9030 a double free of the ticket data.
9035 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9037 * Segmentation fault in ASN1_TYPE_cmp fix
9039 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9040 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9041 certificate signature algorithm consistency this can be used to crash any
9042 certificate verification operation and exploited in a DoS attack. Any
9043 application which performs certificate verification is vulnerable including
9044 OpenSSL clients and servers which enable client authentication.
9049 * ASN.1 structure reuse memory corruption fix
9051 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9052 memory corruption via an invalid write. Such reuse is and has been
9053 strongly discouraged and is believed to be rare.
9055 Applications that parse structures containing CHOICE or ANY DEFINED BY
9056 components may be affected. Certificate parsing (d2i_X509 and related
9057 functions) are however not affected. OpenSSL clients and servers are
9063 * PKCS7 NULL pointer dereferences fix
9065 The PKCS#7 parsing code does not handle missing outer ContentInfo
9066 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9067 missing content and trigger a NULL pointer dereference on parsing.
9069 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9070 otherwise parse PKCS#7 structures from untrusted sources are
9071 affected. OpenSSL clients and servers are not affected.
9073 This issue was reported to OpenSSL by Michal Zalewski (Google).
9078 * DoS via reachable assert in SSLv2 servers fix
9080 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9081 servers that both support SSLv2 and enable export cipher suites by sending
9082 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9084 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9085 (OpenSSL development team).
9090 * Use After Free following d2i_ECPrivatekey error fix
9092 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9093 could cause a use after free condition. This, in turn, could cause a double
9094 free in several private key parsing functions (such as d2i_PrivateKey
9095 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9096 for applications that receive EC private keys from untrusted
9097 sources. This scenario is considered rare.
9099 This issue was discovered by the BoringSSL project and fixed in their
9105 * X509_to_X509_REQ NULL pointer deref fix
9107 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9108 the certificate key is invalid. This function is rarely used in practice.
9110 This issue was discovered by Brian Carpenter.
9115 * Removed the export ciphers from the DEFAULT ciphers
9119 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9121 * Build fixes for the Windows and OpenVMS platforms
9123 *Matt Caswell and Richard Levitte*
9125 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9127 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9128 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9129 dereference. This could lead to a Denial Of Service attack. Thanks to
9130 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9135 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9136 dtls1_buffer_record function under certain conditions. In particular this
9137 could occur if an attacker sent repeated DTLS records with the same
9138 sequence number but for the next epoch. The memory leak could be exploited
9139 by an attacker in a Denial of Service attack through memory exhaustion.
9140 Thanks to Chris Mueller for reporting this issue.
9145 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9146 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9147 method would be set to NULL which could later result in a NULL pointer
9148 dereference. Thanks to Frank Schmirler for reporting this issue.
9153 * Abort handshake if server key exchange message is omitted for ephemeral
9156 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9157 reporting this issue.
9162 * Remove non-export ephemeral RSA code on client and server. This code
9163 violated the TLS standard by allowing the use of temporary RSA keys in
9164 non-export ciphersuites and could be used by a server to effectively
9165 downgrade the RSA key length used to a value smaller than the server
9166 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9167 INRIA or reporting this issue.
9172 * Fixed issue where DH client certificates are accepted without verification.
9173 An OpenSSL server will accept a DH certificate for client authentication
9174 without the certificate verify message. This effectively allows a client to
9175 authenticate without the use of a private key. This only affects servers
9176 which trust a client certificate authority which issues certificates
9177 containing DH keys: these are extremely rare and hardly ever encountered.
9178 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9184 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9185 results on some platforms, including x86_64. This bug occurs at random
9186 with a very low probability, and is not known to be exploitable in any
9187 way, though its exact impact is difficult to determine. Thanks to Pieter
9188 Wuille (Blockstream) who reported this issue and also suggested an initial
9189 fix. Further analysis was conducted by the OpenSSL development team and
9190 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9191 the OpenSSL core team.
9196 * Fix various certificate fingerprint issues.
9198 By using non-DER or invalid encodings outside the signed portion of a
9199 certificate the fingerprint can be changed without breaking the signature.
9200 Although no details of the signed portion of the certificate can be changed
9201 this can cause problems with some applications: e.g. those using the
9202 certificate fingerprint for blacklists.
9204 1. Reject signatures with non zero unused bits.
9206 If the BIT STRING containing the signature has non zero unused bits reject
9207 the signature. All current signature algorithms require zero unused bits.
9209 2. Check certificate algorithm consistency.
9211 Check the AlgorithmIdentifier inside TBS matches the one in the
9212 certificate signature. NB: this will result in signature failure
9213 errors for some broken certificates.
9215 Thanks to Konrad Kraszewski from Google for reporting this issue.
9217 3. Check DSA/ECDSA signatures use DER.
9219 Re-encode DSA/ECDSA signatures and compare with the original received
9220 signature. Return an error if there is a mismatch.
9222 This will reject various cases including garbage after signature
9223 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9224 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9225 (negative or with leading zeroes).
9227 Further analysis was conducted and fixes were developed by Stephen Henson
9228 of the OpenSSL core team.
9234 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9236 * Session Ticket Memory Leak.
9238 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9239 integrity of that ticket is first verified. In the event of a session
9240 ticket integrity check failing, OpenSSL will fail to free memory
9241 causing a memory leak. By sending a large number of invalid session
9242 tickets an attacker could exploit this issue in a Denial Of Service
9248 * Build option no-ssl3 is incomplete.
9250 When OpenSSL is configured with "no-ssl3" as a build option, servers
9251 could accept and complete an SSL 3.0 handshake, and clients could be
9252 configured to send them.
9255 *Akamai and the OpenSSL team*
9257 * Add support for TLS_FALLBACK_SCSV.
9258 Client applications doing fallback retries should call
9259 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9262 *Adam Langley, Bodo Moeller*
9264 * Add additional DigestInfo checks.
9266 Re-encode DigestInto in DER and check against the original when
9267 verifying RSA signature: this will reject any improperly encoded
9268 DigestInfo structures.
9270 Note: this is a precautionary measure and no attacks are currently known.
9274 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9276 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9277 to a denial of service attack. A malicious server can crash the client
9278 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9279 ciphersuite and sending carefully crafted handshake messages.
9281 Thanks to Felix Gröbert (Google) for discovering and researching this
9287 * By sending carefully crafted DTLS packets an attacker could cause openssl
9288 to leak memory. This can be exploited through a Denial of Service attack.
9289 Thanks to Adam Langley for discovering and researching this issue.
9294 * An attacker can force openssl to consume large amounts of memory whilst
9295 processing DTLS handshake messages. This can be exploited through a
9296 Denial of Service attack.
9297 Thanks to Adam Langley for discovering and researching this issue.
9302 * An attacker can force an error condition which causes openssl to crash
9303 whilst processing DTLS packets due to memory being freed twice. This
9304 can be exploited through a Denial of Service attack.
9305 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9311 * If a multithreaded client connects to a malicious server using a resumed
9312 session and the server sends an ec point format extension it could write
9313 up to 255 bytes to freed memory.
9315 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9321 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9322 X509_name_oneline, X509_name_print_ex et al. to leak some information
9323 from the stack. Applications may be affected if they echo pretty printing
9324 output to the attacker.
9326 Thanks to Ivan Fratric (Google) for discovering this issue.
9329 *Emilia Käsper, and Steve Henson*
9331 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9332 for corner cases. (Certain input points at infinity could lead to
9333 bogus results, with non-infinity inputs mapped to infinity too.)
9337 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9339 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9340 handshake can force the use of weak keying material in OpenSSL
9341 SSL/TLS clients and servers.
9343 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9344 researching this issue. ([CVE-2014-0224])
9346 *KIKUCHI Masashi, Steve Henson*
9348 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9349 OpenSSL DTLS client the code can be made to recurse eventually crashing
9352 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9355 *Imre Rad, Steve Henson*
9357 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9358 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9359 client or server. This is potentially exploitable to run arbitrary
9360 code on a vulnerable client or server.
9362 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9364 *Jüri Aedla, Steve Henson*
9366 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9367 are subject to a denial of service attack.
9369 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9370 this issue. ([CVE-2014-3470])
9372 *Felix Gröbert, Ivan Fratric, Steve Henson*
9374 * Harmonize version and its documentation. -f flag is used to display
9377 *mancha <mancha1@zoho.com>*
9379 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9380 in i2d_ECPrivateKey.
9382 *mancha <mancha1@zoho.com>*
9384 * Fix some double frees. These are not thought to be exploitable.
9386 *mancha <mancha1@zoho.com>*
9388 * Fix for the attack described in the paper "Recovering OpenSSL
9389 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9390 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9391 <http://eprint.iacr.org/2014/140>
9393 Thanks to Yuval Yarom and Naomi Benger for discovering this
9394 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9396 *Yuval Yarom and Naomi Benger*
9398 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9400 * Keep original DTLS digest and encryption contexts in retransmission
9401 structures so we can use the previous session parameters if they need
9402 to be resent. ([CVE-2013-6450])
9406 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9407 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9408 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9409 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9410 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9411 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9413 *Rob Stradling, Adam Langley*
9415 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9417 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9419 This addresses the flaw in CBC record processing discovered by
9420 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9421 at: <http://www.isg.rhul.ac.uk/tls/>
9423 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9424 Security Group at Royal Holloway, University of London
9425 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9426 Emilia Käsper for the initial patch.
9429 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9431 * Return an error when checking OCSP signatures when key is NULL.
9432 This fixes a DoS attack. ([CVE-2013-0166])
9436 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9437 the right response is stapled. Also change SSL_get_certificate()
9438 so it returns the certificate actually sent.
9439 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9440 (This is a backport)
9442 *Rob Stradling <rob.stradling@comodo.com>*
9444 * Fix possible deadlock when decoding public keys.
9448 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9450 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9453 * Sanity check record length before skipping explicit IV in DTLS
9456 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9457 fuzzing as a service testing platform.
9462 * Initialise tkeylen properly when encrypting CMS messages.
9463 Thanks to Solar Designer of Openwall for reporting this issue.
9467 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9469 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9470 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9471 in CRYPTO_realloc_clean.
9473 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9474 issue and to Adam Langley <agl@chromium.org> for fixing it.
9477 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9479 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9481 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9482 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9483 content decryption and always return the same error. Note: this attack
9484 needs on average 2^20 messages so it only affects automated senders. The
9485 old behaviour can be re-enabled in the CMS code by setting the
9486 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9487 an MMA defence is not necessary.
9488 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9489 this issue. ([CVE-2012-0884])
9493 * Fix CVE-2011-4619: make sure we really are receiving a
9494 client hello before rejecting multiple SGC restarts. Thanks to
9495 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9499 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9501 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9502 Thanks to Antonio Martin, Enterprise Secure Access Research and
9503 Development, Cisco Systems, Inc. for discovering this bug and
9504 preparing a fix. ([CVE-2012-0050])
9508 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9510 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9511 of the Vaudenay padding oracle attack on CBC mode encryption
9512 which enables an efficient plaintext recovery attack against
9513 the OpenSSL implementation of DTLS. Their attack exploits timing
9514 differences arising during decryption processing. A research
9515 paper describing this attack can be found at:
9516 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9517 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9518 Security Group at Royal Holloway, University of London
9519 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9520 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9521 for preparing the fix. ([CVE-2011-4108])
9523 *Robin Seggelmann, Michael Tuexen*
9525 * Clear bytes used for block padding of SSL 3.0 records.
9528 *Adam Langley (Google)*
9530 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9531 Kadianakis <desnacked@gmail.com> for discovering this issue and
9532 Adam Langley for preparing the fix. ([CVE-2011-4619])
9534 *Adam Langley (Google)*
9536 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9538 *Andrey Kulikov <amdeich@gmail.com>*
9540 * Prevent malformed RFC3779 data triggering an assertion failure.
9541 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9542 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9544 *Rob Austein <sra@hactrn.net>*
9546 * Improved PRNG seeding for VOS.
9548 *Paul Green <Paul.Green@stratus.com>*
9550 * Fix ssl_ciph.c set-up race.
9552 *Adam Langley (Google)*
9554 * Fix spurious failures in ecdsatest.c.
9556 *Emilia Käsper (Google)*
9558 * Fix the BIO_f_buffer() implementation (which was mixing different
9559 interpretations of the `..._len` fields).
9561 *Adam Langley (Google)*
9563 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9564 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9565 threads won't reuse the same blinding coefficients.
9567 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9568 lock to call BN_BLINDING_invert_ex, and avoids one use of
9569 BN_BLINDING_update for each BN_BLINDING structure (previously,
9570 the last update always remained unused).
9572 *Emilia Käsper (Google)*
9574 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9576 *Bob Buckholz (Google)*
9578 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9580 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9581 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9583 *Kaspar Brand <ossl@velox.ch>*
9585 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9586 for multi-threaded use of ECDH. ([CVE-2011-3210])
9588 *Adam Langley (Google)*
9590 * Fix x509_name_ex_d2i memory leak on bad inputs.
9594 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9595 signature public key algorithm by using OID xref utilities instead.
9596 Before this you could only use some ECC ciphersuites with SHA1 only.
9600 * Add protection against ECDSA timing attacks as mentioned in the paper
9601 by Billy Bob Brumley and Nicola Tuveri, see:
9602 <http://eprint.iacr.org/2011/232.pdf>
9604 *Billy Bob Brumley and Nicola Tuveri*
9606 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9608 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9610 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9612 * Fix bug in string printing code: if *any* escaping is enabled we must
9613 escape the escape character (backslash) or the resulting string is
9618 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9620 * Disable code workaround for ancient and obsolete Netscape browsers
9621 and servers: an attacker can use it in a ciphersuite downgrade attack.
9622 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9626 * Fixed J-PAKE implementation error, originally discovered by
9627 Sebastien Martini, further info and confirmation from Stefan
9628 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9632 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9634 * Fix extension code to avoid race conditions which can result in a buffer
9635 overrun vulnerability: resumed sessions must not be modified as they can
9636 be shared by multiple threads. CVE-2010-3864
9640 * Fix WIN32 build system to correctly link an ENGINE directory into
9645 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9647 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9650 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9652 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9654 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9655 context. The operation can be customised via the ctrl mechanism in
9656 case ENGINEs want to include additional functionality.
9660 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9664 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9665 output hashes compatible with older versions of OpenSSL.
9667 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9669 * Fix compression algorithm handling: if resuming a session use the
9670 compression algorithm of the resumed session instead of determining
9671 it from client hello again. Don't allow server to change algorithm.
9675 * Add load_crls() function to commands tidying load_certs() too. Add option
9676 to verify utility to allow additional CRLs to be included.
9680 * Update OCSP request code to permit adding custom headers to the request:
9681 some responders need this.
9685 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9688 *Julia Lawall <julia@diku.dk>*
9690 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9691 needlessly dereferenced structures, used obsolete functions and
9692 didn't handle all updated verify codes correctly.
9696 * Disable MD2 in the default configuration.
9700 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9701 indicate the initial BIO being pushed or popped. This makes it possible
9702 to determine whether the BIO is the one explicitly called or as a result
9703 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9704 it handles reference counts correctly and doesn't zero out the I/O bio
9705 when it is not being explicitly popped. WARNING: applications which
9706 included workarounds for the old buggy behaviour will need to be modified
9707 or they could free up already freed BIOs.
9711 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9712 renaming to all platforms (within the 0.9.8 branch, this was
9713 done conditionally on Netware platforms to avoid a name clash).
9715 *Guenter <lists@gknw.net>*
9717 * Add ECDHE and PSK support to DTLS.
9719 *Michael Tuexen <tuexen@fh-muenster.de>*
9721 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9726 * Add "missing" function EVP_MD_flags() (without this the only way to
9727 retrieve a digest flags is by accessing the structure directly. Update
9728 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9729 or cipher is registered as in the "from" argument. Print out all
9730 registered digests in the dgst usage message instead of manually
9731 attempting to work them out.
9735 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9736 this allows the use of compression and extensions. Change default cipher
9737 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9738 by default unless an application cipher string requests it.
9742 * Alter match criteria in PKCS12_parse(). It used to try to use local
9743 key ids to find matching certificates and keys but some PKCS#12 files
9744 don't follow the (somewhat unwritten) rules and this strategy fails.
9745 Now just gather all certificates together and the first private key
9746 then look for the first certificate that matches the key.
9750 * Support use of registered digest and cipher names for dgst and cipher
9751 commands instead of having to add each one as a special case. So now
9758 openssl dgst -sha256 foo
9760 and this works for ENGINE based algorithms too.
9764 * Update Gost ENGINE to support parameter files.
9766 *Victor B. Wagner <vitus@cryptocom.ru>*
9768 * Support GeneralizedTime in ca utility.
9770 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9772 * Enhance the hash format used for certificate directory links. The new
9773 form uses the canonical encoding (meaning equivalent names will work
9774 even if they aren't identical) and uses SHA1 instead of MD5. This form
9775 is incompatible with the older format and as a result c_rehash should
9776 be used to rebuild symbolic links.
9780 * Make PKCS#8 the default write format for private keys, replacing the
9781 traditional format. This form is standardised, more secure and doesn't
9782 include an implicit MD5 dependency.
9786 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9787 committed to OpenSSL should pass this lot as a minimum.
9791 * Add session ticket override functionality for use by EAP-FAST.
9793 *Jouni Malinen <j@w1.fi>*
9795 * Modify HMAC functions to return a value. Since these can be implemented
9796 in an ENGINE errors can occur.
9800 * Type-checked OBJ_bsearch_ex.
9804 * Type-checked OBJ_bsearch. Also some constification necessitated
9805 by type-checking. Still to come: TXT_DB, bsearch(?),
9806 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9811 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9812 seconds to a tm structure directly, instead of going through OS
9813 specific date routines. This avoids any issues with OS routines such
9814 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9815 and X509_time_adj_ex() to cover the extended range. The existing
9816 X509_time_adj() is still usable and will no longer have any date issues.
9820 * Delta CRL support. New use deltas option which will attempt to locate
9821 and search any appropriate delta CRLs available.
9823 This work was sponsored by Google.
9827 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9828 code and add additional score elements. Validate alternate CRL paths
9829 as part of the CRL checking and indicate a new error "CRL path validation
9830 error" in this case. Applications wanting additional details can use
9831 the verify callback and check the new "parent" field. If this is not
9832 NULL CRL path validation is taking place. Existing applications won't
9833 see this because it requires extended CRL support which is off by
9836 This work was sponsored by Google.
9840 * Support for freshest CRL extension.
9842 This work was sponsored by Google.
9846 * Initial indirect CRL support. Currently only supported in the CRLs
9847 passed directly and not via lookup. Process certificate issuer
9848 CRL entry extension and lookup CRL entries by bother issuer name
9849 and serial number. Check and process CRL issuer entry in IDP extension.
9851 This work was sponsored by Google.
9855 * Add support for distinct certificate and CRL paths. The CRL issuer
9856 certificate is validated separately in this case. Only enabled if
9857 an extended CRL support flag is set: this flag will enable additional
9858 CRL functionality in future.
9860 This work was sponsored by Google.
9864 * Add support for policy mappings extension.
9866 This work was sponsored by Google.
9870 * Fixes to pathlength constraint, self issued certificate handling,
9871 policy processing to align with RFC3280 and PKITS tests.
9873 This work was sponsored by Google.
9877 * Support for name constraints certificate extension. DN, email, DNS
9878 and URI types are currently supported.
9880 This work was sponsored by Google.
9884 * To cater for systems that provide a pointer-based thread ID rather
9885 than numeric, deprecate the current numeric thread ID mechanism and
9886 replace it with a structure and associated callback type. This
9887 mechanism allows a numeric "hash" to be extracted from a thread ID in
9888 either case, and on platforms where pointers are larger than 'long',
9889 mixing is done to help ensure the numeric 'hash' is usable even if it
9890 can't be guaranteed unique. The default mechanism is to use "&errno"
9891 as a pointer-based thread ID to distinguish between threads.
9893 Applications that want to provide their own thread IDs should now use
9894 CRYPTO_THREADID_set_callback() to register a callback that will call
9895 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9897 Note that ERR_remove_state() is now deprecated, because it is tied
9898 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9899 to free the current thread's error state should be replaced by
9900 ERR_remove_thread_state(NULL).
9902 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9903 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9904 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9905 application was previously providing a numeric thread callback that
9906 was inappropriate for distinguishing threads, then uniqueness might
9907 have been obtained with &errno that happened immediately in the
9908 intermediate development versions of OpenSSL; this is no longer the
9909 case, the numeric thread callback will now override the automatic use
9912 *Geoff Thorpe, with help from Bodo Moeller*
9914 * Initial support for different CRL issuing certificates. This covers a
9915 simple case where the self issued certificates in the chain exist and
9916 the real CRL issuer is higher in the existing chain.
9918 This work was sponsored by Google.
9922 * Removed effectively defunct crypto/store from the build.
9926 * Revamp of STACK to provide stronger type-checking. Still to come:
9927 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9928 ASN1_STRING, CONF_VALUE.
9932 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9933 RAM on SSL connections. This option can save about 34k per idle SSL.
9937 * Revamp of LHASH to provide stronger type-checking. Still to come:
9938 STACK, TXT_DB, bsearch, qsort.
9942 * Initial support for Cryptographic Message Syntax (aka CMS) based
9943 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9944 support for data, signedData, compressedData, digestedData and
9945 encryptedData, envelopedData types included. Scripts to check against
9946 RFC4134 examples draft and interop and consistency checks of many
9947 content types and variants.
9951 * Add options to enc utility to support use of zlib compression BIO.
9955 * Extend mk1mf to support importing of options and assembly language
9956 files from Configure script, currently only included in VC-WIN32.
9957 The assembly language rules can now optionally generate the source
9958 files from the associated perl scripts.
9962 * Implement remaining functionality needed to support GOST ciphersuites.
9963 Interop testing has been performed using CryptoPro implementations.
9965 *Victor B. Wagner <vitus@cryptocom.ru>*
9967 * s390x assembler pack.
9971 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9976 * Implement Opaque PRF Input TLS extension as specified in
9977 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9978 official specification yet and no extension type assignment by
9979 IANA exists, this extension (for now) will have to be explicitly
9980 enabled when building OpenSSL by providing the extension number
9981 to use. For example, specify an option
9983 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9985 to the "config" or "Configure" script to enable the extension,
9986 assuming extension number 0x9527 (which is a completely arbitrary
9987 and unofficial assignment based on the MD5 hash of the Internet
9988 Draft). Note that by doing so, you potentially lose
9989 interoperability with other TLS implementations since these might
9990 be using the same extension number for other purposes.
9992 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9993 opaque PRF input value to use in the handshake. This will create
9994 an internal copy of the length-'len' string at 'src', and will
9995 return non-zero for success.
9997 To get more control and flexibility, provide a callback function
10000 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10001 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10005 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10008 Callback function 'cb' will be called in handshakes, and is
10009 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10010 Argument 'arg' is for application purposes (the value as given to
10011 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10012 be provided to the callback function). The callback function
10013 has to return non-zero to report success: usually 1 to use opaque
10014 PRF input just if possible, or 2 to enforce use of the opaque PRF
10015 input. In the latter case, the library will abort the handshake
10016 if opaque PRF input is not successfully negotiated.
10018 Arguments 'peerinput' and 'len' given to the callback function
10019 will always be NULL and 0 in the case of a client. A server will
10020 see the client's opaque PRF input through these variables if
10021 available (NULL and 0 otherwise). Note that if the server
10022 provides an opaque PRF input, the length must be the same as the
10023 length of the client's opaque PRF input.
10025 Note that the callback function will only be called when creating
10026 a new session (session resumption can resume whatever was
10027 previously negotiated), and will not be called in SSL 2.0
10028 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10029 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10030 for applications that need to enforce opaque PRF input.
10034 * Update ssl code to support digests other than SHA1+MD5 for handshake
10037 *Victor B. Wagner <vitus@cryptocom.ru>*
10039 * Add RFC4507 support to OpenSSL. This includes the corrections in
10040 RFC4507bis. The encrypted ticket format is an encrypted encoded
10041 SSL_SESSION structure, that way new session features are automatically
10044 If a client application caches session in an SSL_SESSION structure
10045 support is transparent because tickets are now stored in the encoded
10048 The SSL_CTX structure automatically generates keys for ticket
10049 protection in servers so again support should be possible
10050 with no application modification.
10052 If a client or server wishes to disable RFC4507 support then the option
10053 SSL_OP_NO_TICKET can be set.
10055 Add a TLS extension debugging callback to allow the contents of any client
10056 or server extensions to be examined.
10058 This work was sponsored by Google.
10062 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10063 OpenSSL should now compile cleanly on gcc 4.2
10065 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10067 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10068 support including streaming MAC support: this is required for GOST
10069 ciphersuite support.
10071 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10073 * Add option -stream to use PKCS#7 streaming in smime utility. New
10074 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10075 to output in BER and PEM format.
10079 * Experimental support for use of HMAC via EVP_PKEY interface. This
10080 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10081 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10082 ENGINE support for HMAC keys which are unextractable. New -mac and
10083 -macopt options to dgst utility.
10087 * New option -sigopt to dgst utility. Update dgst to use
10088 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10089 alternative signing parameters such as X9.31 or PSS in the dgst
10094 * Change ssl_cipher_apply_rule(), the internal function that does
10095 the work each time a ciphersuite string requests enabling
10096 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10097 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10098 the order of disabled ciphersuites such that those ciphersuites
10099 that most recently went from enabled to disabled not only stay
10100 in order with respect to each other, but also have higher priority
10101 than other disabled ciphersuites the next time ciphersuites are
10104 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10105 the same ciphersuites as with "HIGH" alone, but in a specific
10106 order where the PSK ciphersuites come first (since they are the
10107 most recently disabled ciphersuites when "HIGH" is parsed).
10109 Also, change ssl_create_cipher_list() (using this new
10110 functionality) such that between otherwise identical
10111 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10116 * Change ssl_create_cipher_list() so that it automatically
10117 arranges the ciphersuites in reasonable order before starting
10118 to process the rule string. Thus, the definition for "DEFAULT"
10119 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10120 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10121 This makes it much easier to arrive at a reasonable default order
10122 in applications for which anonymous ciphers are OK (meaning
10123 that you can't actually use DEFAULT).
10125 *Bodo Moeller; suggested by Victor Duchovni*
10127 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10128 processing) into multiple integers instead of setting
10129 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10130 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10131 (These masks as well as the individual bit definitions are hidden
10132 away into the non-exported interface ssl/ssl_locl.h, so this
10133 change to the definition of the SSL_CIPHER structure shouldn't
10134 affect applications.) This give us more bits for each of these
10135 categories, so there is no longer a need to coagulate AES128 and
10136 AES256 into a single algorithm bit, and to coagulate Camellia128
10137 and Camellia256 into a single algorithm bit, which has led to all
10140 Thus, among other things, the kludge introduced in 0.9.7m and
10141 0.9.8e for masking out AES256 independently of AES128 or masking
10142 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10144 With the change, we also introduce new ciphersuite aliases that
10145 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10150 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10151 Use the leftmost N bytes of the signature input if the input is
10152 larger than the prime q (with N being the size in bytes of q).
10156 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10157 it yet and it is largely untested.
10161 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10165 * Initial incomplete changes to avoid need for function casts in OpenSSL
10166 some compilers (gcc 4.2 and later) reject their use. Safestack is
10167 reimplemented. Update ASN1 to avoid use of legacy functions.
10171 * Win32/64 targets are linked with Winsock2.
10175 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10176 to external functions. This can be used to increase CRL handling
10177 efficiency especially when CRLs are very large by (for example) storing
10178 the CRL revoked certificates in a database.
10182 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10183 new CRLs added to a directory can be used. New command line option
10184 -verify_return_error to s_client and s_server. This causes real errors
10185 to be returned by the verify callback instead of carrying on no matter
10186 what. This reflects the way a "real world" verify callback would behave.
10190 * GOST engine, supporting several GOST algorithms and public key formats.
10191 Kindly donated by Cryptocom.
10195 * Partial support for Issuing Distribution Point CRL extension. CRLs
10196 partitioned by DP are handled but no indirect CRL or reason partitioning
10197 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10198 selected via a scoring technique which handles IDP and AKID in CRLs.
10202 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10203 will ultimately be used for all verify operations: this will remove the
10204 X509_STORE dependency on certificate verification and allow alternative
10205 lookup methods. X509_STORE based implementations of these two callbacks.
10209 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10210 Modify get_crl() to find a valid (unexpired) CRL if possible.
10214 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10215 this would be called X509_CRL_cmp() but that name is already used by
10216 a function that just compares CRL issuer names. Cache several CRL
10217 extensions in X509_CRL structure and cache CRLDP in X509.
10221 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10222 this maps equivalent X509_NAME structures into a consistent structure.
10223 Name comparison can then be performed rapidly using memcmp().
10227 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10232 * Allow digests to supply their own micalg string for S/MIME type using
10233 the ctrl EVP_MD_CTRL_MICALG.
10237 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10238 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10239 ctrl. It can then customise the structure before and/or after signing
10244 * New function OBJ_add_sigid() to allow application defined signature OIDs
10245 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10246 to free up any added signature OIDs.
10250 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10251 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10252 digest and cipher tables. New options added to openssl utility:
10253 list-message-digest-algorithms and list-cipher-algorithms.
10257 * Change the array representation of binary polynomials: the list
10258 of degrees of non-zero coefficients is now terminated with -1.
10259 Previously it was terminated with 0, which was also part of the
10260 value; thus, the array representation was not applicable to
10261 polynomials where t^0 has coefficient zero. This change makes
10262 the array representation useful in a more general context.
10266 * Various modifications and fixes to SSL/TLS cipher string
10267 handling. For ECC, the code now distinguishes between fixed ECDH
10268 with RSA certificates on the one hand and with ECDSA certificates
10269 on the other hand, since these are separate ciphersuites. The
10270 unused code for Fortezza ciphersuites has been removed.
10272 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10273 (not "ECDHE"). For consistency with the code for DH
10274 certificates, use of ECDH certificates is now considered ECDH
10275 authentication, not RSA or ECDSA authentication (the latter is
10276 merely the CA's signing algorithm and not actively used in the
10279 The temporary ciphersuite alias "ECCdraft" is no longer
10280 available, and ECC ciphersuites are no longer excluded from "ALL"
10281 and "DEFAULT". The following aliases now exist for RFC 4492
10282 ciphersuites, most of these by analogy with the DH case:
10284 kECDHr - ECDH cert, signed with RSA
10285 kECDHe - ECDH cert, signed with ECDSA
10286 kECDH - ECDH cert (signed with either RSA or ECDSA)
10287 kEECDH - ephemeral ECDH
10288 ECDH - ECDH cert or ephemeral ECDH
10291 aECDSA - ECDSA cert
10294 AECDH - anonymous ECDH
10295 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10299 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10300 Use correct micalg parameters depending on digest(s) in signed message.
10304 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10305 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10309 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10310 an engine to register a method. Add ENGINE lookups for methods and
10311 functional reference processing.
10315 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10316 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10321 * New -resign option to smime utility. This adds one or more signers
10322 to an existing PKCS#7 signedData structure. Also -md option to use an
10323 alternative message digest algorithm for signing.
10327 * Tidy up PKCS#7 routines and add new functions to make it easier to
10328 create PKCS7 structures containing multiple signers. Update smime
10329 application to support multiple signers.
10333 * New -macalg option to pkcs12 utility to allow setting of an alternative
10338 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10339 Reorganize PBE internals to lookup from a static table using NIDs,
10340 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10341 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10342 PRF which will be automatically used with PBES2.
10346 * Replace the algorithm specific calls to generate keys in "req" with the
10351 * Update PKCS#7 enveloped data routines to use new API. This is now
10352 supported by any public key method supporting the encrypt operation. A
10353 ctrl is added to allow the public key algorithm to examine or modify
10354 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10359 * Add a ctrl to asn1 method to allow a public key algorithm to express
10360 a default digest type to use. In most cases this will be SHA1 but some
10361 algorithms (such as GOST) need to specify an alternative digest. The
10362 return value indicates how strong the preference is 1 means optional and
10363 2 is mandatory (that is it is the only supported type). Modify
10364 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10365 use the default md. Update openssl utilities to use the default digest
10366 type for signing if it is not explicitly indicated.
10370 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10371 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10372 signing method from the key type. This effectively removes the link
10373 between digests and public key types.
10377 * Add an OID cross reference table and utility functions. Its purpose is to
10378 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10379 rsaEncryption. This will allow some of the algorithm specific hackery
10380 needed to use the correct OID to be removed.
10384 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10385 structures for PKCS7_sign(). They are now set up by the relevant public
10390 * Add provisional EC pkey method with support for ECDSA and ECDH.
10394 * Add support for key derivation (agreement) in the API, DH method and
10399 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10400 public and private key formats. As a side effect these add additional
10401 command line functionality not previously available: DSA signatures can be
10402 generated and verified using pkeyutl and DH key support and generation in
10409 *Oliver Tappe <zooey@hirschkaefer.de>*
10411 * New make target "install_html_docs" installs HTML renditions of the
10414 *Oliver Tappe <zooey@hirschkaefer.de>*
10416 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10417 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10418 support key and parameter generation and add initial key generation
10419 functionality for RSA.
10423 * Add functions for main EVP_PKEY_method operations. The undocumented
10424 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10425 `EVP_PKEY_{encrypt,decrypt}_old`.
10429 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10430 key API, doesn't do much yet.
10434 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10435 public key algorithms. New option to openssl utility:
10436 "list-public-key-algorithms" to print out info.
10440 * Implement the Supported Elliptic Curves Extension for
10441 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10445 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10446 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10450 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10451 utilities such as rsa, dsa, dsaparam etc except they process any key
10456 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10457 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10458 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10463 * Initial support for pluggable public key ASN1.
10464 De-spaghettify the public key ASN1 handling. Move public and private
10465 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10466 algorithm specific handling to a single module within the relevant
10467 algorithm directory. Add functions to allow (near) opaque processing
10468 of public and private key structures.
10472 * Implement the Supported Point Formats Extension for
10473 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10477 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10478 for the psk identity [hint] and the psk callback functions to the
10479 SSL_SESSION, SSL and SSL_CTX structure.
10482 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10486 SSL_CTX_use_psk_identity_hint
10487 SSL_get_psk_identity_hint
10488 SSL_get_psk_identity
10489 SSL_use_psk_identity_hint
10491 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10493 * Add RFC 3161 compliant time stamp request creation, response generation
10494 and response verification functionality.
10496 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10498 * Add initial support for TLS extensions, specifically for the server_name
10499 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10500 have new members for a hostname. The SSL data structure has an
10501 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10502 stored in that context to allow for session resumption, even after the
10503 SSL has been switched to a new SSL_CTX in reaction to a client's
10504 server_name extension.
10506 New functions (subject to change):
10508 SSL_get_servername()
10509 SSL_get_servername_type()
10512 New CTRL codes and macros (subject to change):
10514 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10515 - SSL_CTX_set_tlsext_servername_callback()
10516 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10517 - SSL_CTX_set_tlsext_servername_arg()
10518 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10520 openssl s_client has a new '-servername ...' option.
10522 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10523 '-key2 ...', '-servername_fatal' (subject to change). This allows
10524 testing the HostName extension for a specific single hostname ('-cert'
10525 and '-key' remain fallbacks for handshakes without HostName
10526 negotiation). If the unrecognized_name alert has to be sent, this by
10527 default is a warning; it becomes fatal with the '-servername_fatal'
10530 *Peter Sylvester, Remy Allais, Christophe Renou*
10532 * Whirlpool hash implementation is added.
10536 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10537 bn(64,32). Because of instruction set limitations it doesn't have
10538 any negative impact on performance. This was done mostly in order
10539 to make it possible to share assembler modules, such as bn_mul_mont
10540 implementations, between 32- and 64-bit builds without hassle.
10544 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10545 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10550 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10551 dedicated Montgomery multiplication procedure, is introduced.
10552 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10553 "64-bit" performance on certain 32-bit targets.
10557 * New option SSL_OP_NO_COMP to disable use of compression selectively
10558 in SSL structures. New SSL ctrl to set maximum send fragment size.
10559 Save memory by setting the I/O buffer sizes dynamically instead of
10560 using the maximum available value.
10564 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10565 in addition to the text details.
10569 * Very, very preliminary EXPERIMENTAL support for printing of general
10570 ASN1 structures. This currently produces rather ugly output and doesn't
10571 handle several customised structures at all.
10575 * Integrated support for PVK file format and some related formats such
10576 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10577 these in the 'rsa' and 'dsa' utilities.
10581 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10585 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10586 place for the (very old) "NETSCAPE" format certificates which are now
10587 handled using new ASN1 code equivalents.
10591 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10592 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10593 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10597 * Modify CRL distribution points extension code to print out previously
10598 unsupported fields. Enhance extension setting code to allow setting of
10603 * Add print and set support for Issuing Distribution Point CRL extension.
10607 * Change 'Configure' script to enable Camellia by default.
10614 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10616 * When rejecting SSL/TLS records due to an incorrect version number, never
10617 update s->server with a new major version number. As of
10618 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10619 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10620 the previous behavior could result in a read attempt at NULL when
10621 receiving specific incorrect SSL/TLS records once record payload
10622 protection is active. ([CVE-2010-0740])
10624 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10626 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10627 could be crashed if the relevant tables were not present (e.g. chrooted).
10629 *Tomas Hoger <thoger@redhat.com>*
10631 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10633 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10635 *Martin Olsson, Neel Mehta*
10637 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10638 accommodate for stack sorting, always a write lock!).
10642 * On some versions of WIN32 Heap32Next is very slow. This can cause
10643 excessive delays in the RAND_poll(): over a minute. As a workaround
10644 include a time check in the inner Heap32Next loop too.
10648 * The code that handled flushing of data in SSL/TLS originally used the
10649 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10650 the problem outlined in PR#1949. The fix suggested there however can
10651 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10652 of Apache). So instead simplify the code to flush unconditionally.
10653 This should be fine since flushing with no data to flush is a no op.
10657 * Handle TLS versions 2.0 and later properly and correctly use the
10658 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10659 off ancient servers have a habit of sticking around for a while...
10663 * Modify compression code so it frees up structures without using the
10664 ex_data callbacks. This works around a problem where some applications
10665 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10666 restarting) then use compression (e.g. SSL with compression) later.
10667 This results in significant per-connection memory leaks and
10668 has caused some security issues including CVE-2008-1678 and
10673 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10674 change when encrypting or decrypting.
10678 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10679 connect and renegotiate with servers which do not support RI.
10680 Until RI is more widely deployed this option is enabled by default.
10684 * Add "missing" ssl ctrls to clear options and mode.
10688 * If client attempts to renegotiate and doesn't support RI respond with
10689 a no_renegotiation alert as required by RFC5746. Some renegotiating
10690 TLS clients will continue a connection gracefully when they receive
10691 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10692 waiting for a server hello which it will never receive. Now we treat a
10693 received no_renegotiation alert as a fatal error. This is because
10694 applications requesting a renegotiation might well expect it to succeed
10695 and would have no code in place to handle the server denying it so the
10696 only safe thing to do is to terminate the connection.
10700 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10701 peer supports secure renegotiation and 0 otherwise. Print out peer
10702 renegotiation support in s_client/s_server.
10706 * Replace the highly broken and deprecated SPKAC certification method with
10707 the updated NID creation version. This should correctly handle UTF8.
10711 * Implement RFC5746. Re-enable renegotiation but require the extension
10712 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10713 turns out to be a bad idea. It has been replaced by
10714 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10715 SSL_CTX_set_options(). This is really not recommended unless you
10716 know what you are doing.
10718 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10720 * Fixes to stateless session resumption handling. Use initial_ctx when
10721 issuing and attempting to decrypt tickets in case it has changed during
10722 servername handling. Use a non-zero length session ID when attempting
10723 stateless session resumption: this makes it possible to determine if
10724 a resumption has occurred immediately after receiving server hello
10725 (several places in OpenSSL subtly assume this) instead of later in
10730 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10731 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10732 fixes for a few places where the return code is not checked
10735 *Julia Lawall <julia@diku.dk>*
10737 * Add --strict-warnings option to Configure script to include devteam
10738 warnings in other configurations.
10742 * Add support for --libdir option and LIBDIR variable in makefiles. This
10743 makes it possible to install openssl libraries in locations which
10744 have names other than "lib", for example "/usr/lib64" which some
10747 *Steve Henson, based on patch from Jeremy Utley*
10749 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10750 X690 8.9.12 and can produce some misleading textual output of OIDs.
10752 *Steve Henson, reported by Dan Kaminsky*
10754 * Delete MD2 from algorithm tables. This follows the recommendation in
10755 several standards that it is not used in new applications due to
10756 several cryptographic weaknesses. For binary compatibility reasons
10757 the MD2 API is still compiled in by default.
10761 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10766 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10767 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10770 *Guenter <lists@gknw.net>*
10772 * Fix the server certificate chain building code to use X509_verify_cert(),
10773 it used to have an ad-hoc builder which was unable to cope with anything
10774 other than a simple chain.
10776 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10778 * Don't check self signed certificate signatures in X509_verify_cert()
10779 by default (a flag can override this): it just wastes time without
10780 adding any security. As a useful side effect self signed root CAs
10781 with non-FIPS digests are now usable in FIPS mode.
10785 * In dtls1_process_out_of_seq_message() the check if the current message
10786 is already buffered was missing. For every new message was memory
10787 allocated, allowing an attacker to perform an denial of service attack
10788 with sending out of seq handshake messages until there is no memory
10789 left. Additionally every future message was buffered, even if the
10790 sequence number made no sense and would be part of another handshake.
10791 So only messages with sequence numbers less than 10 in advance will be
10792 buffered. ([CVE-2009-1378])
10794 *Robin Seggelmann, discovered by Daniel Mentz*
10796 * Records are buffered if they arrive with a future epoch to be
10797 processed after finishing the corresponding handshake. There is
10798 currently no limitation to this buffer allowing an attacker to perform
10799 a DOS attack with sending records with future epochs until there is no
10800 memory left. This patch adds the pqueue_size() function to determine
10801 the size of a buffer and limits the record buffer to 100 entries.
10804 *Robin Seggelmann, discovered by Daniel Mentz*
10806 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10807 parent structure is freed. ([CVE-2009-1379])
10811 * Handle non-blocking I/O properly in SSL_shutdown() call.
10813 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10815 * Add `2.5.4.*` OIDs
10817 *Ilya O. <vrghost@gmail.com>*
10819 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10821 * Disable renegotiation completely - this fixes a severe security
10822 problem ([CVE-2009-3555]) at the cost of breaking all
10823 renegotiation. Renegotiation can be re-enabled by setting
10824 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10825 run-time. This is really not recommended unless you know what
10830 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10832 * Don't set val to NULL when freeing up structures, it is freed up by
10833 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10834 zeroing past the valid field. ([CVE-2009-0789])
10836 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10838 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10839 checked correctly. This would allow some invalid signed attributes to
10840 appear to verify correctly. ([CVE-2009-0591])
10842 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10844 * Reject UniversalString and BMPString types with invalid lengths. This
10845 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10846 a legal length. ([CVE-2009-0590])
10850 * Set S/MIME signing as the default purpose rather than setting it
10851 unconditionally. This allows applications to override it at the store
10856 * Permit restricted recursion of ASN1 strings. This is needed in practice
10857 to handle some structures.
10861 * Improve efficiency of mem_gets: don't search whole buffer each time
10864 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10866 * New -hex option for openssl rand.
10870 * Print out UTF8String and NumericString when parsing ASN1.
10874 * Support NumericString type for name components.
10878 * Allow CC in the environment to override the automatically chosen
10879 compiler. Note that nothing is done to ensure flags work with the
10884 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10886 * Properly check EVP_VerifyFinal() and similar return values
10889 *Ben Laurie, Bodo Moeller, Google Security Team*
10891 * Enable TLS extensions by default.
10895 * Allow the CHIL engine to be loaded, whether the application is
10896 multithreaded or not. (This does not release the developer from the
10897 obligation to set up the dynamic locking callbacks.)
10899 *Sander Temme <sander@temme.net>*
10901 * Use correct exit code if there is an error in dgst command.
10903 *Steve Henson; problem pointed out by Roland Dirlewanger*
10905 * Tweak Configure so that you need to say "experimental-jpake" to enable
10906 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10910 * Add experimental JPAKE support, including demo authentication in
10911 s_client and s_server.
10915 * Set the comparison function in v3_addr_canonize().
10917 *Rob Austein <sra@hactrn.net>*
10919 * Add support for XMPP STARTTLS in s_client.
10921 *Philip Paeps <philip@freebsd.org>*
10923 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10924 to ensure that even with this option, only ciphersuites in the
10925 server's preference list will be accepted. (Note that the option
10926 applies only when resuming a session, so the earlier behavior was
10927 just about the algorithm choice for symmetric cryptography.)
10931 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10933 * Fix NULL pointer dereference if a DTLS server received
10934 ChangeCipherSpec as first record ([CVE-2009-1386]).
10938 * Fix a state transition in s3_srvr.c and d1_srvr.c
10939 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10941 *Nagendra Modadugu*
10943 * The fix in 0.9.8c that supposedly got rid of unsafe
10944 double-checked locking was incomplete for RSA blinding,
10945 addressing just one layer of what turns out to have been
10946 doubly unsafe triple-checked locking.
10948 So now fix this for real by retiring the MONT_HELPER macro
10949 in crypto/rsa/rsa_eay.c.
10951 *Bodo Moeller; problem pointed out by Marius Schilder*
10953 * Various precautionary measures:
10955 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10957 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10958 (NB: This would require knowledge of the secret session ticket key
10959 to exploit, in which case you'd be SOL either way.)
10961 - Change bn_nist.c so that it will properly handle input BIGNUMs
10962 outside the expected range.
10964 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10967 *Neel Mehta, Bodo Moeller*
10969 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10970 the load fails. Useful for distros.
10972 *Ben Laurie and the FreeBSD team*
10974 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10978 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10982 * Expand ENGINE to support engine supplied SSL client certificate functions.
10984 This work was sponsored by Logica.
10988 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10989 keystores. Support for SSL/TLS client authentication too.
10990 Not compiled unless enable-capieng specified to Configure.
10992 This work was sponsored by Logica.
10996 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10997 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10998 attribute creation routines such as certificate requests and PKCS#12
11003 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11005 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11006 handshake which could lead to a client crash as found using the
11007 Codenomicon TLS test suite ([CVE-2008-1672])
11009 *Steve Henson, Mark Cox*
11011 * Fix double free in TLS server name extensions which could lead to
11012 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11016 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11018 Clear the error queue to ensure that error entries left from
11019 older function calls do not interfere with the correct operation.
11021 *Lutz Jaenicke, Erik de Castro Lopo*
11023 * Remove root CA certificates of commercial CAs:
11025 The OpenSSL project does not recommend any specific CA and does not
11026 have any policy with respect to including or excluding any CA.
11027 Therefore, it does not make any sense to ship an arbitrary selection
11028 of root CA certificates with the OpenSSL software.
11032 * RSA OAEP patches to fix two separate invalid memory reads.
11033 The first one involves inputs when 'lzero' is greater than
11034 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11035 before the beginning of from). The second one involves inputs where
11036 the 'db' section contains nothing but zeroes (there is a one-byte
11037 invalid read after the end of 'db').
11039 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11041 * Partial backport from 0.9.9-dev:
11043 Introduce bn_mul_mont (dedicated Montgomery multiplication
11044 procedure) as a candidate for BIGNUM assembler implementation.
11045 While 0.9.9-dev uses assembler for various architectures, only
11046 x86_64 is available by default here in the 0.9.8 branch, and
11047 32-bit x86 is available through a compile-time setting.
11049 To try the 32-bit x86 assembler implementation, use Configure
11050 option "enable-montasm" (which exists only for this backport).
11052 As "enable-montasm" for 32-bit x86 disclaims code stability
11053 anyway, in this constellation we activate additional code
11054 backported from 0.9.9-dev for further performance improvements,
11055 namely BN_from_montgomery_word. (To enable this otherwise,
11056 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11058 *Andy Polyakov (backport partially by Bodo Moeller)*
11060 * Add TLS session ticket callback. This allows an application to set
11061 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11062 values. This is useful for key rollover for example where several key
11063 sets may exist with different names.
11067 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11068 This was broken until now in 0.9.8 releases, such that the only way
11069 a registered ENGINE could be used (assuming it initialises
11070 successfully on the host) was to explicitly set it as the default
11071 for the relevant algorithms. This is in contradiction with 0.9.7
11072 behaviour and the documentation. With this fix, when an ENGINE is
11073 registered into a given algorithm's table of implementations, the
11074 'uptodate' flag is reset so that auto-discovery will be used next
11075 time a new context for that algorithm attempts to select an
11078 *Ian Lister (tweaked by Geoff Thorpe)*
11080 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11081 implementation in the following ways:
11083 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11086 Lack of BER streaming support means one pass streaming processing is
11087 only supported if data is detached: setting the streaming flag is
11088 ignored for embedded content.
11090 CMS support is disabled by default and must be explicitly enabled
11091 with the enable-cms configuration option.
11095 * Update the GMP engine glue to do direct copies between BIGNUM and
11096 mpz_t when openssl and GMP use the same limb size. Otherwise the
11097 existing "conversion via a text string export" trick is still used.
11099 *Paul Sheer <paulsheer@gmail.com>*
11101 * Zlib compression BIO. This is a filter BIO which compressed and
11102 uncompresses any data passed through it.
11106 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11107 RFC3394 compatible AES key wrapping.
11111 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11112 sets string data without copying. X509_ALGOR_set0() and
11113 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11114 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11115 from an X509_ATTRIBUTE structure optionally checking it occurs only
11116 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11121 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11122 to get the expected BN_FLG_CONSTTIME behavior.
11124 *Bodo Moeller (Google)*
11128 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11129 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11130 - added some more tests to do_tests.pl
11131 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11132 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11133 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11134 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11135 - various changes to netware.pl to enable gcc-cross builds on Win32
11137 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11138 - various changes to fix missing prototype warnings
11139 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11140 - added AES, WHIRLPOOL and CPUID assembler code to build files
11141 - added missing AES assembler make rules to mk1mf.pl
11142 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11144 *Guenter Knauf <eflash@gmx.net>*
11146 * Implement certificate status request TLS extension defined in RFC3546.
11147 A client can set the appropriate parameters and receive the encoded
11148 OCSP response via a callback. A server can query the supplied parameters
11149 and set the encoded OCSP response in the callback. Add simplified examples
11150 to s_client and s_server.
11154 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11156 * Fix various bugs:
11157 + Binary incompatibility of ssl_ctx_st structure
11158 + DTLS interoperation with non-compliant servers
11159 + Don't call get_session_cb() without proposed session
11160 + Fix ia64 assembler code
11162 *Andy Polyakov, Steve Henson*
11164 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11166 * DTLS Handshake overhaul. There were longstanding issues with
11167 OpenSSL DTLS implementation, which were making it impossible for
11168 RFC 4347 compliant client to communicate with OpenSSL server.
11169 Unfortunately just fixing these incompatibilities would "cut off"
11170 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11171 server keeps tolerating non RFC compliant syntax. The opposite is
11172 not true, 0.9.8f client can not communicate with earlier server.
11173 This update even addresses CVE-2007-4995.
11177 * Changes to avoid need for function casts in OpenSSL: some compilers
11178 (gcc 4.2 and later) reject their use.
11179 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11182 * Add RFC4507 support to OpenSSL. This includes the corrections in
11183 RFC4507bis. The encrypted ticket format is an encrypted encoded
11184 SSL_SESSION structure, that way new session features are automatically
11187 If a client application caches session in an SSL_SESSION structure
11188 support is transparent because tickets are now stored in the encoded
11191 The SSL_CTX structure automatically generates keys for ticket
11192 protection in servers so again support should be possible
11193 with no application modification.
11195 If a client or server wishes to disable RFC4507 support then the option
11196 SSL_OP_NO_TICKET can be set.
11198 Add a TLS extension debugging callback to allow the contents of any client
11199 or server extensions to be examined.
11201 This work was sponsored by Google.
11205 * Add initial support for TLS extensions, specifically for the server_name
11206 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11207 have new members for a hostname. The SSL data structure has an
11208 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11209 stored in that context to allow for session resumption, even after the
11210 SSL has been switched to a new SSL_CTX in reaction to a client's
11211 server_name extension.
11213 New functions (subject to change):
11215 SSL_get_servername()
11216 SSL_get_servername_type()
11219 New CTRL codes and macros (subject to change):
11221 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11222 - SSL_CTX_set_tlsext_servername_callback()
11223 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11224 - SSL_CTX_set_tlsext_servername_arg()
11225 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11227 openssl s_client has a new '-servername ...' option.
11229 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11230 '-key2 ...', '-servername_fatal' (subject to change). This allows
11231 testing the HostName extension for a specific single hostname ('-cert'
11232 and '-key' remain fallbacks for handshakes without HostName
11233 negotiation). If the unrecognized_name alert has to be sent, this by
11234 default is a warning; it becomes fatal with the '-servername_fatal'
11237 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11239 * Add AES and SSE2 assembly language support to VC++ build.
11243 * Mitigate attack on final subtraction in Montgomery reduction.
11247 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11248 (which previously caused an internal error).
11252 * Squeeze another 10% out of IGE mode when in != out.
11256 * AES IGE mode speedup.
11258 *Dean Gaudet (Google)*
11260 * Add the Korean symmetric 128-bit cipher SEED (see
11261 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11262 add SEED ciphersuites from RFC 4162:
11264 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11265 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11266 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11267 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11269 To minimize changes between patchlevels in the OpenSSL 0.9.8
11270 series, SEED remains excluded from compilation unless OpenSSL
11271 is configured with 'enable-seed'.
11273 *KISA, Bodo Moeller*
11275 * Mitigate branch prediction attacks, which can be practical if a
11276 single processor is shared, allowing a spy process to extract
11277 information. For detailed background information, see
11278 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11279 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11280 and Necessary Software Countermeasures"). The core of the change
11281 are new versions BN_div_no_branch() and
11282 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11283 respectively, which are slower, but avoid the security-relevant
11284 conditional branches. These are automatically called by BN_div()
11285 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11286 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11287 remove a conditional branch.
11289 BN_FLG_CONSTTIME is the new name for the previous
11290 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11291 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11292 in the exponent causes BN_mod_exp_mont() to use the alternative
11293 implementation in BN_mod_exp_mont_consttime().) The old name
11294 remains as a deprecated alias.
11296 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11297 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11298 constant-time implementations for more than just exponentiation.
11299 Here too the old name is kept as a deprecated alias.
11301 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11302 the BN_BLINDING structure gets an independent copy of the
11303 modulus. This means that the previous `BIGNUM *m` argument to
11304 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11305 essentially becomes `const BIGNUM *m`, although we can't actually
11306 change this in the header file before 0.9.9. It allows
11307 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11308 enable BN_FLG_CONSTTIME.
11310 *Matthew D Wood (Intel Corp)*
11312 * In the SSL/TLS server implementation, be strict about session ID
11313 context matching (which matters if an application uses a single
11314 external cache for different purposes). Previously,
11315 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11316 set. This did ensure strict client verification, but meant that,
11317 with applications using a single external cache for quite
11318 different requirements, clients could circumvent ciphersuite
11319 restrictions for a given session ID context by starting a session
11320 in a different context.
11324 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11325 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11326 authentication-only ciphersuites.
11330 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11331 not complete and could lead to a possible single byte overflow
11332 ([CVE-2007-5135]) [Ben Laurie]
11334 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11336 * Since AES128 and AES256 (and similarly Camellia128 and
11337 Camellia256) share a single mask bit in the logic of
11338 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11339 kludge to work properly if AES128 is available and AES256 isn't
11340 (or if Camellia128 is available and Camellia256 isn't).
11344 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11345 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11346 When a point or a seed is encoded in a BIT STRING, we need to
11347 prevent the removal of trailing zero bits to get the proper DER
11348 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11349 of a NamedBitList, for which trailing 0 bits need to be removed.)
11353 * Have SSL/TLS server implementation tolerate "mismatched" record
11354 protocol version while receiving ClientHello even if the
11355 ClientHello is fragmented. (The server can't insist on the
11356 particular protocol version it has chosen before the ServerHello
11357 message has informed the client about his choice.)
11361 * Add RFC 3779 support.
11363 *Rob Austein for ARIN, Ben Laurie*
11365 * Load error codes if they are not already present instead of using a
11366 static variable. This allows them to be cleanly unloaded and reloaded.
11367 Improve header file function name parsing.
11371 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11372 or CAPABILITY handshake as required by RFCs.
11374 *Goetz Babin-Ebell*
11376 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11378 * Introduce limits to prevent malicious keys being able to
11379 cause a denial of service. ([CVE-2006-2940])
11381 *Steve Henson, Bodo Moeller*
11383 * Fix ASN.1 parsing of certain invalid structures that can result
11384 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11386 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11387 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11389 * Fix SSL client code which could crash if connecting to a
11390 malicious SSLv2 server. ([CVE-2006-4343])
11392 *Tavis Ormandy and Will Drewry, Google Security Team*
11394 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11395 match only those. Before that, "AES256-SHA" would be interpreted
11396 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11397 the same strength classification in 0.9.7h) as we currently only
11398 have a single AES bit in the ciphersuite description bitmap.
11399 That change, however, also applied to ciphersuite strings such as
11400 "RC4-MD5" that intentionally matched multiple ciphersuites --
11401 namely, SSL 2.0 ciphersuites in addition to the more common ones
11402 from SSL 3.0/TLS 1.0.
11404 So we change the selection algorithm again: Naming an explicit
11405 ciphersuite selects this one ciphersuite, and any other similar
11406 ciphersuite (same bitmap) from *other* protocol versions.
11407 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11408 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11410 Since SSL 2.0 does not have any ciphersuites for which the
11411 128/256 bit distinction would be relevant, this works for now.
11412 The proper fix will be to use different bits for AES128 and
11413 AES256, which would have avoided the problems from the beginning;
11414 however, bits are scarce, so we can only do this in a new release
11415 (not just a patchlevel) when we can change the SSL_CIPHER
11416 definition to split the single 'unsigned long mask' bitmap into
11417 multiple values to extend the available space.
11421 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11423 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11424 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11426 * Add AES IGE and biIGE modes.
11430 * Change the Unix randomness entropy gathering to use poll() when
11431 possible instead of select(), since the latter has some
11432 undesirable limitations.
11434 *Darryl Miles via Richard Levitte and Bodo Moeller*
11436 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11437 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11438 cannot be implicitly activated as part of, e.g., the "AES" alias.
11439 However, please upgrade to OpenSSL 0.9.9[-dev] for
11440 non-experimental use of the ECC ciphersuites to get TLS extension
11441 support, which is required for curve and point format negotiation
11442 to avoid potential handshake problems.
11446 * Disable rogue ciphersuites:
11448 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11449 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11450 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11452 The latter two were purportedly from
11453 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11456 Also deactivate the remaining ciphersuites from
11457 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11458 unofficial, and the ID has long expired.
11462 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11463 dual-core machines) and other potential thread-safety issues.
11467 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11468 versions), which is now available for royalty-free use
11469 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11470 Also, add Camellia TLS ciphersuites from RFC 4132.
11472 To minimize changes between patchlevels in the OpenSSL 0.9.8
11473 series, Camellia remains excluded from compilation unless OpenSSL
11474 is configured with 'enable-camellia'.
11478 * Disable the padding bug check when compression is in use. The padding
11479 bug check assumes the first packet is of even length, this is not
11480 necessarily true if compression is enabled and can result in false
11481 positives causing handshake failure. The actual bug test is ancient
11482 code so it is hoped that implementations will either have fixed it by
11483 now or any which still have the bug do not support compression.
11487 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11489 * When applying a cipher rule check to see if string match is an explicit
11490 cipher suite and only match that one cipher suite if it is.
11494 * Link in manifests for VC++ if needed.
11496 *Austin Ziegler <halostatue@gmail.com>*
11498 * Update support for ECC-based TLS ciphersuites according to
11499 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11500 TLS extensions, which are supported starting with the 0.9.9
11501 branch, not in the OpenSSL 0.9.8 branch).
11505 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11506 opaque EVP_CIPHER_CTX handling.
11510 * Fixes and enhancements to zlib compression code. We now only use
11511 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11512 to conform with the standards mentioned here:
11513 <http://www.zlib.net/DLL_FAQ.txt>
11514 Static zlib linking now works on Windows and the new --with-zlib-include
11515 --with-zlib-lib options to Configure can be used to supply the location
11516 of the headers and library. Gracefully handle case where zlib library
11521 * Several fixes and enhancements to the OID generation code. The old code
11522 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11523 handle numbers larger than ULONG_MAX, truncated printing and had a
11524 non standard OBJ_obj2txt() behaviour.
11528 * Add support for building of engines under engine/ as shared libraries
11529 under VC++ build system.
11533 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11534 Hopefully, we will not see any false combination of paths any more.
11538 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11540 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11541 (part of SSL_OP_ALL). This option used to disable the
11542 countermeasure against man-in-the-middle protocol-version
11543 rollback in the SSL 2.0 server implementation, which is a bad
11544 idea. ([CVE-2005-2969])
11546 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11547 for Information Security, National Institute of Advanced Industrial
11548 Science and Technology [AIST], Japan)*
11550 * Add two function to clear and return the verify parameter flags.
11554 * Keep cipherlists sorted in the source instead of sorting them at
11555 runtime, thus removing the need for a lock.
11559 * Avoid some small subgroup attacks in Diffie-Hellman.
11561 *Nick Mathewson and Ben Laurie*
11563 * Add functions for well-known primes.
11567 * Extended Windows CE support.
11569 *Satoshi Nakamura and Andy Polyakov*
11571 * Initialize SSL_METHOD structures at compile time instead of during
11572 runtime, thus removing the need for a lock.
11576 * Make PKCS7_decrypt() work even if no certificate is supplied by
11577 attempting to decrypt each encrypted key in turn. Add support to
11582 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11584 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11587 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11591 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11592 key into the same file any more.
11596 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11600 * Add -utf8 command line and config file option to 'ca'.
11602 *Stefan <stf@udoma.org*
11604 * Removed the macro des_crypt(), as it seems to conflict with some
11605 libraries. Use DES_crypt().
11609 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11610 involves renaming the source and generated shared-libs for
11611 both. The engines will accept the corrected or legacy ids
11612 ('ncipher' and '4758_cca' respectively) when binding. NB,
11613 this only applies when building 'shared'.
11615 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11617 * Add attribute functions to EVP_PKEY structure. Modify
11618 PKCS12_create() to recognize a CSP name attribute and
11619 use it. Make -CSP option work again in pkcs12 utility.
11623 * Add new functionality to the bn blinding code:
11624 - automatic re-creation of the BN_BLINDING parameters after
11625 a fixed number of uses (currently 32)
11626 - add new function for parameter creation
11627 - introduce flags to control the update behaviour of the
11628 BN_BLINDING parameters
11629 - hide BN_BLINDING structure
11630 Add a second BN_BLINDING slot to the RSA structure to improve
11631 performance when a single RSA object is shared among several
11636 * Add support for DTLS.
11638 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11640 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11641 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11645 * Remove buggy and incomplete DH cert support from
11646 ssl/ssl_rsa.c and ssl/s3_both.c
11650 * Use SHA-1 instead of MD5 as the default digest algorithm for
11651 the `apps/openssl` commands.
11655 * Compile clean with "-Wall -Wmissing-prototypes
11656 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11657 DEBUG_SAFESTACK must also be set.
11661 * Change ./Configure so that certain algorithms can be disabled by default.
11662 The new counterpiece to "no-xxx" is "enable-xxx".
11664 The patented RC5 and MDC2 algorithms will now be disabled unless
11665 "enable-rc5" and "enable-mdc2", respectively, are specified.
11667 (IDEA remains enabled despite being patented. This is because IDEA
11668 is frequently required for interoperability, and there is no license
11669 fee for non-commercial use. As before, "no-idea" can be used to
11670 avoid this algorithm.)
11674 * Add processing of proxy certificates (see RFC 3820). This work was
11675 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11676 EGEE (Enabling Grids for E-science in Europe).
11680 * RC4 performance overhaul on modern architectures/implementations, such
11681 as Intel P4, IA-64 and AMD64.
11685 * New utility extract-section.pl. This can be used specify an alternative
11686 section number in a pod file instead of having to treat each file as
11687 a separate case in Makefile. This can be done by adding two lines to the
11690 =for comment openssl_section:XXX
11692 The blank line is mandatory.
11696 * New arguments -certform, -keyform and -pass for s_client and s_server
11697 to allow alternative format key and certificate files and passphrase
11702 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11703 update associated structures and add various utility functions.
11705 Add new policy related verify parameters, include policy checking in
11706 standard verify code. Enhance 'smime' application with extra parameters
11707 to support policy checking and print out.
11711 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11712 Nehemiah processors. These extensions support AES encryption in hardware
11713 as well as RNG (though RNG support is currently disabled).
11715 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11717 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11721 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11723 *Andy Polyakov and a number of other people*
11725 * Improved PowerPC platform support. Most notably BIGNUM assembler
11726 implementation contributed by IBM.
11728 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11730 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11731 exponent rather than 'unsigned long'. There is a corresponding change to
11732 the new 'rsa_keygen' element of the RSA_METHOD structure.
11734 *Jelte Jansen, Geoff Thorpe*
11736 * Functionality for creating the initial serial number file is now
11737 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11739 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11740 number file to 1, which is bound to cause problems. To avoid
11741 the problems while respecting compatibility between different 0.9.7
11742 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11743 CA.pl for serial number initialization. With the new release 0.9.8,
11744 we can fix the problem directly in the 'ca' utility.)
11748 * Reduced header interdependencies by declaring more opaque objects in
11749 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11750 give fewer recursive includes, which could break lazy source code - so
11751 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11752 developers should define this symbol when building and using openssl to
11753 ensure they track the recommended behaviour, interfaces, [etc], but
11754 backwards-compatible behaviour prevails when this isn't defined.
11758 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11762 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11763 This will generate a random key of the appropriate length based on the
11764 cipher context. The EVP_CIPHER can provide its own random key generation
11765 routine to support keys of a specific form. This is used in the des and
11766 3des routines to generate a key of the correct parity. Update S/MIME
11767 code to use new functions and hence generate correct parity DES keys.
11768 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11769 valid (weak or incorrect parity).
11773 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11774 as looking them up. This is useful when the verified structure may contain
11775 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11776 present unless the new PKCS7_NO_CRL flag is asserted.
11780 * Extend ASN1 oid configuration module. It now additionally accepts the
11783 shortName = some long name, 1.2.3.4
11787 * Reimplemented the BN_CTX implementation. There is now no more static
11788 limitation on the number of variables it can handle nor the depth of the
11789 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11790 information can now expand as required, and rather than having a single
11791 static array of bignums, BN_CTX now uses a linked-list of such arrays
11792 allowing it to expand on demand whilst maintaining the usefulness of
11793 BN_CTX's "bundling".
11797 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11798 to allow all RSA operations to function using a single BN_CTX.
11802 * Preliminary support for certificate policy evaluation and checking. This
11803 is initially intended to pass the tests outlined in "Conformance Testing
11804 of Relying Party Client Certificate Path Processing Logic" v1.07.
11808 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11809 remained unused and not that useful. A variety of other little bignum
11810 tweaks and fixes have also been made continuing on from the audit (see
11815 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11816 associated ASN1, EVP and SSL functions and old ASN1 macros.
11820 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11821 and this should never fail. So the return value from the use of
11822 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11823 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11827 * BN_CTX_get() should return zero-valued bignums, providing the same
11828 initialised value as BN_new().
11830 *Geoff Thorpe, suggested by Ulf Möller*
11832 * Support for inhibitAnyPolicy certificate extension.
11836 * An audit of the BIGNUM code is underway, for which debugging code is
11837 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11838 is considered valid when processing BIGNUMs, and causes execution to
11839 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11840 further steps are taken to deliberately pollute unused data in BIGNUM
11841 structures to try and expose faulty code further on. For now, openssl will
11842 (in its default mode of operation) continue to tolerate the inconsistent
11843 forms that it has tolerated in the past, but authors and packagers should
11844 consider trying openssl and their own applications when compiled with
11845 these debugging symbols defined. It will help highlight potential bugs in
11846 their own code, and will improve the test coverage for OpenSSL itself. At
11847 some point, these tighter rules will become openssl's default to improve
11848 maintainability, though the assert()s and other overheads will remain only
11849 in debugging configurations. See bn.h for more details.
11851 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11853 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11854 that can only be obtained through BN_CTX_new() (which implicitly
11855 initialises it). The presence of this function only made it possible
11856 to overwrite an existing structure (and cause memory leaks).
11860 * Because of the callback-based approach for implementing LHASH as a
11861 template type, lh_insert() adds opaque objects to hash-tables and
11862 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11863 to clean up those corresponding objects before destroying the hash table
11864 (and losing the object pointers). So some over-zealous constifications in
11865 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11866 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11867 prototyped to have "const" restrictions on the object pointers they are
11868 given (and so aren't required to cast them away any more).
11872 * The tmdiff.h API was so ugly and minimal that our own timing utility
11873 (speed) prefers to use its own implementation. The two implementations
11874 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11875 its object type properly exposed (MS_TM) instead of casting to/from
11876 `char *`. This may still change yet if someone realises MS_TM and
11878 aren't necessarily the greatest nomenclatures - but this is what was used
11879 internally to the implementation so I've used that for now.
11883 * Ensure that deprecated functions do not get compiled when
11884 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11885 the self-tests were still using deprecated key-generation functions so
11886 these have been updated also.
11890 * Reorganise PKCS#7 code to separate the digest location functionality
11891 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11892 New function PKCS7_set_digest() to set the digest type for PKCS#7
11893 digestedData type. Add additional code to correctly generate the
11894 digestedData type and add support for this type in PKCS7 initialization
11899 * New function PKCS7_set0_type_other() this initializes a PKCS7
11900 structure of type "other".
11904 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11905 sure the loop does correctly stop and breaking ("division by zero")
11906 modulus operations are not performed. The (pre-generated) prime
11907 table crypto/bn/bn_prime.h was already correct, but it could not be
11908 re-generated on some platforms because of the "division by zero"
11909 situation in the script.
11911 *Ralf S. Engelschall*
11913 * Update support for ECC-based TLS ciphersuites according to
11914 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11915 SHA-1 now is only used for "small" curves (where the
11916 representation of a field element takes up to 24 bytes); for
11917 larger curves, the field element resulting from ECDH is directly
11918 used as premaster secret.
11920 *Douglas Stebila (Sun Microsystems Laboratories)*
11922 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11923 curve secp160r1 to the tests.
11925 *Douglas Stebila (Sun Microsystems Laboratories)*
11927 * Add the possibility to load symbols globally with DSO.
11929 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11931 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11932 control of the error stack.
11936 * Add support for STORE in ENGINE.
11940 * Add the STORE type. The intention is to provide a common interface
11941 to certificate and key stores, be they simple file-based stores, or
11942 HSM-type store, or LDAP stores, or...
11943 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11947 * Add a generic structure called OPENSSL_ITEM. This can be used to
11948 pass a list of arguments to any function as well as provide a way
11949 for a function to pass data back to the caller.
11953 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11954 works like BUF_strdup() but can be used to duplicate a portion of
11955 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11960 * Add the function sk_find_ex() which works like sk_find(), but will
11961 return an index to an element even if an exact match couldn't be
11962 found. The index is guaranteed to point at the element where the
11963 searched-for key would be inserted to preserve sorting order.
11967 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11968 takes an extra flags argument for optional functionality. Currently,
11969 the following flags are defined:
11971 OBJ_BSEARCH_VALUE_ON_NOMATCH
11972 This one gets OBJ_bsearch_ex() to return a pointer to the first
11973 element where the comparing function returns a negative or zero
11976 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11977 This one gets OBJ_bsearch_ex() to return a pointer to the first
11978 element where the comparing function returns zero. This is useful
11979 if there are more than one element where the comparing function
11984 * Make it possible to create self-signed certificates with 'openssl ca'
11985 in such a way that the self-signed certificate becomes part of the
11986 CA database and uses the same mechanisms for serial number generation
11987 as all other certificate signing. The new flag '-selfsign' enables
11988 this functionality. Adapt CA.sh and CA.pl.in.
11992 * Add functionality to check the public key of a certificate request
11993 against a given private. This is useful to check that a certificate
11994 request can be signed by that key (self-signing).
11998 * Make it possible to have multiple active certificates with the same
11999 subject in the CA index file. This is done only if the keyword
12000 'unique_subject' is set to 'no' in the main CA section (default
12001 if 'CA_default') of the configuration file. The value is saved
12002 with the database itself in a separate index attribute file,
12003 named like the index file with '.attr' appended to the name.
12007 * Generate multi-valued AVAs using '+' notation in config files for
12012 * Support for nameConstraints certificate extension.
12016 * Support for policyConstraints certificate extension.
12020 * Support for policyMappings certificate extension.
12024 * Make sure the default DSA_METHOD implementation only uses its
12025 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12026 and change its own handlers to be NULL so as to remove unnecessary
12027 indirection. This lets alternative implementations fallback to the
12028 default implementation more easily.
12032 * Support for directoryName in GeneralName related extensions
12037 * Make it possible to link applications using Makefile.shared.
12038 Make that possible even when linking against static libraries!
12042 * Support for single pass processing for S/MIME signing. This now
12043 means that S/MIME signing can be done from a pipe, in addition
12044 cleartext signing (multipart/signed type) is effectively streaming
12045 and the signed data does not need to be all held in memory.
12047 This is done with a new flag PKCS7_STREAM. When this flag is set
12048 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12049 is done after the data is output (and digests calculated) in
12050 SMIME_write_PKCS7().
12054 * Add full support for -rpath/-R, both in shared libraries and
12055 applications, at least on the platforms where it's known how
12060 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12061 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12062 will now compute a table of multiples of the generator that
12063 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12064 faster (notably in the case of a single point multiplication,
12065 scalar * generator).
12067 *Nils Larsch, Bodo Moeller*
12069 * IPv6 support for certificate extensions. The various extensions
12070 which use the IP:a.b.c.d can now take IPv6 addresses using the
12071 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12076 * Added an ENGINE that implements RSA by performing private key
12077 exponentiations with the GMP library. The conversions to and from
12078 GMP's mpz_t format aren't optimised nor are any montgomery forms
12079 cached, and on x86 it appears OpenSSL's own performance has caught up.
12080 However there are likely to be other architectures where GMP could
12081 provide a boost. This ENGINE is not built in by default, but it can be
12082 specified at Configure time and should be accompanied by the necessary
12083 linker additions, eg;
12084 ./config -DOPENSSL_USE_GMP -lgmp
12088 * "openssl engine" will not display ENGINE/DSO load failure errors when
12089 testing availability of engines with "-t" - the old behaviour is
12090 produced by increasing the feature's verbosity with "-tt".
12094 * ECDSA routines: under certain error conditions uninitialized BN objects
12095 could be freed. Solution: make sure initialization is performed early
12096 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12101 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12102 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12103 software implementations. For DSA and DH, parameter generation can
12104 also be overridden by providing the appropriate method callbacks.
12108 * Change the "progress" mechanism used in key-generation and
12109 primality testing to functions that take a new BN_GENCB pointer in
12110 place of callback/argument pairs. The new API functions have `_ex`
12111 postfixes and the older functions are reimplemented as wrappers for
12112 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12113 declarations of the old functions to help (graceful) attempts to
12114 migrate to the new functions. Also, the new key-generation API
12115 functions operate on a caller-supplied key-structure and return
12116 success/failure rather than returning a key or NULL - this is to
12117 help make "keygen" another member function of RSA_METHOD etc.
12119 Example for using the new callback interface:
12121 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12122 void *my_arg = ...;
12125 BN_GENCB_set(&my_cb, my_callback, my_arg);
12127 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12128 /* For the meaning of a, b in calls to my_callback(), see the
12129 * documentation of the function that calls the callback.
12130 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12131 * my_callback should return 1 if it wants BN_is_prime_ex()
12132 * to continue, or 0 to stop.
12137 * Change the ZLIB compression method to be stateful, and make it
12138 available to TLS with the number defined in
12139 draft-ietf-tls-compression-04.txt.
12143 * Add the ASN.1 structures and functions for CertificatePair, which
12144 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12146 CertificatePair ::= SEQUENCE {
12147 forward [0] Certificate OPTIONAL,
12148 reverse [1] Certificate OPTIONAL,
12149 -- at least one of the pair shall be present -- }
12151 Also implement the PEM functions to read and write certificate
12152 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12154 This needed to be defined, mostly for the sake of the LDAP
12155 attribute crossCertificatePair, but may prove useful elsewhere as
12160 * Make it possible to inhibit symlinking of shared libraries in
12161 Makefile.shared, for Cygwin's sake.
12165 * Extend the BIGNUM API by creating a function
12166 void BN_set_negative(BIGNUM *a, int neg);
12167 and a macro that behave like
12168 int BN_is_negative(const BIGNUM *a);
12170 to avoid the need to access 'a->neg' directly in applications.
12174 * Implement fast modular reduction for pseudo-Mersenne primes
12175 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12176 EC_GROUP_new_curve_GFp() will now automatically use this
12179 *Nils Larsch <nla@trustcenter.de>*
12181 * Add new lock type (CRYPTO_LOCK_BN).
12185 * Change the ENGINE framework to automatically load engines
12186 dynamically from specific directories unless they could be
12187 found to already be built in or loaded. Move all the
12188 current engines except for the cryptodev one to a new
12189 directory engines/.
12190 The engines in engines/ are built as shared libraries if
12191 the "shared" options was given to ./Configure or ./config.
12192 Otherwise, they are inserted in libcrypto.a.
12193 /usr/local/ssl/engines is the default directory for dynamic
12194 engines, but that can be overridden at configure time through
12195 the usual use of --prefix and/or --openssldir, and at run
12196 time with the environment variable OPENSSL_ENGINES.
12198 *Geoff Thorpe and Richard Levitte*
12200 * Add Makefile.shared, a helper makefile to build shared
12201 libraries. Adapt Makefile.org.
12205 * Add version info to Win32 DLLs.
12207 *Peter 'Luna' Runestig" <peter@runestig.com>*
12209 * Add new 'medium level' PKCS#12 API. Certificates and keys
12210 can be added using this API to created arbitrary PKCS#12
12211 files while avoiding the low-level API.
12213 New options to PKCS12_create(), key or cert can be NULL and
12214 will then be omitted from the output file. The encryption
12215 algorithm NIDs can be set to -1 for no encryption, the mac
12216 iteration count can be set to 0 to omit the mac.
12218 Enhance pkcs12 utility by making the -nokeys and -nocerts
12219 options work when creating a PKCS#12 file. New option -nomac
12220 to omit the mac, NONE can be set for an encryption algorithm.
12221 New code is modified to use the enhanced PKCS12_create()
12222 instead of the low-level API.
12226 * Extend ASN1 encoder to support indefinite length constructed
12227 encoding. This can output sequences tags and octet strings in
12228 this form. Modify pk7_asn1.c to support indefinite length
12229 encoding. This is experimental and needs additional code to
12230 be useful, such as an ASN1 bio and some enhanced streaming
12233 Extend template encode functionality so that tagging is passed
12234 down to the template encoder.
12238 * Let 'openssl req' fail if an argument to '-newkey' is not
12239 recognized instead of using RSA as a default.
12243 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12244 As these are not official, they are not included in "ALL";
12245 the "ECCdraft" ciphersuite group alias can be used to select them.
12247 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12249 * Add ECDH engine support.
12251 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12253 * Add ECDH in new directory crypto/ecdh/.
12255 *Douglas Stebila (Sun Microsystems Laboratories)*
12257 * Let BN_rand_range() abort with an error after 100 iterations
12258 without success (which indicates a broken PRNG).
12262 * Change BN_mod_sqrt() so that it verifies that the input value
12263 is really the square of the return value. (Previously,
12264 BN_mod_sqrt would show GIGO behaviour.)
12268 * Add named elliptic curves over binary fields from X9.62, SECG,
12269 and WAP/WTLS; add OIDs that were still missing.
12271 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12273 * Extend the EC library for elliptic curves over binary fields
12274 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12277 EC_GF2m_simple_method
12281 EC_GROUP_new_curve_GF2m
12282 EC_GROUP_set_curve_GF2m
12283 EC_GROUP_get_curve_GF2m
12284 EC_POINT_set_affine_coordinates_GF2m
12285 EC_POINT_get_affine_coordinates_GF2m
12286 EC_POINT_set_compressed_coordinates_GF2m
12288 Point compression for binary fields is disabled by default for
12289 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12292 As binary polynomials are represented as BIGNUMs, various members
12293 of the EC_GROUP and EC_POINT data structures can be shared
12294 between the implementations for prime fields and binary fields;
12295 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12296 are essentially identical to their `..._GFp` counterparts.
12297 (For simplicity, the `..._GFp` prefix has been dropped from
12298 various internal method names.)
12300 An internal 'field_div' method (similar to 'field_mul' and
12301 'field_sqr') has been added; this is used only for binary fields.
12303 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12305 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12306 through methods ('mul', 'precompute_mult').
12308 The generic implementations (now internally called 'ec_wNAF_mul'
12309 and 'ec_wNAF_precomputed_mult') remain the default if these
12310 methods are undefined.
12312 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12314 * New function EC_GROUP_get_degree, which is defined through
12315 EC_METHOD. For curves over prime fields, this returns the bit
12316 length of the modulus.
12318 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12320 * New functions EC_GROUP_dup, EC_POINT_dup.
12321 (These simply call ..._new and ..._copy).
12323 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12325 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12326 Polynomials are represented as BIGNUMs (where the sign bit is not
12327 used) in the following functions [macros]:
12330 BN_GF2m_sub [= BN_GF2m_add]
12331 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12332 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12333 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12335 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12336 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12337 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12338 BN_GF2m_cmp [= BN_ucmp]
12340 (Note that only the 'mod' functions are actually for fields GF(2^m).
12341 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12343 For some functions, an the irreducible polynomial defining a
12344 field can be given as an 'unsigned int[]' with strictly
12345 decreasing elements giving the indices of those bits that are set;
12346 i.e., p[] represents the polynomial
12347 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12349 p[0] > p[1] > ... > p[k] = 0.
12350 This applies to the following functions:
12353 BN_GF2m_mod_mul_arr
12354 BN_GF2m_mod_sqr_arr
12355 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12356 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12357 BN_GF2m_mod_exp_arr
12358 BN_GF2m_mod_sqrt_arr
12359 BN_GF2m_mod_solve_quad_arr
12363 Conversion can be performed by the following functions:
12368 bntest.c has additional tests for binary polynomial arithmetic.
12370 Two implementations for BN_GF2m_mod_div() are available.
12371 The default algorithm simply uses BN_GF2m_mod_inv() and
12372 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12373 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12374 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12376 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12378 * Add new error code 'ERR_R_DISABLED' that can be used when some
12379 functionality is disabled at compile-time.
12381 *Douglas Stebila <douglas.stebila@sun.com>*
12383 * Change default behaviour of 'openssl asn1parse' so that more
12384 information is visible when viewing, e.g., a certificate:
12386 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12387 mode the content of non-printable OCTET STRINGs is output in a
12388 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12389 avoid the appearance of a printable string.
12391 *Nils Larsch <nla@trustcenter.de>*
12393 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12395 EC_GROUP_set_asn1_flag()
12396 EC_GROUP_get_asn1_flag()
12397 EC_GROUP_set_point_conversion_form()
12398 EC_GROUP_get_point_conversion_form()
12399 These control ASN1 encoding details:
12400 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12401 has been set to OPENSSL_EC_NAMED_CURVE.
12402 - Points are encoded in uncompressed form by default; options for
12403 asn1_for are as for point2oct, namely
12404 POINT_CONVERSION_COMPRESSED
12405 POINT_CONVERSION_UNCOMPRESSED
12406 POINT_CONVERSION_HYBRID
12408 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12410 EC_GROUP_set_seed()
12411 EC_GROUP_get0_seed()
12412 EC_GROUP_get_seed_len()
12413 This is used only for ASN1 purposes (so far).
12415 *Nils Larsch <nla@trustcenter.de>*
12417 * Add 'field_type' member to EC_METHOD, which holds the NID
12418 of the appropriate field type OID. The new function
12419 EC_METHOD_get_field_type() returns this value.
12421 *Nils Larsch <nla@trustcenter.de>*
12424 EC_POINT_point2bn()
12425 EC_POINT_bn2point()
12426 EC_POINT_point2hex()
12427 EC_POINT_hex2point()
12428 providing useful interfaces to EC_POINT_point2oct() and
12429 EC_POINT_oct2point().
12431 *Nils Larsch <nla@trustcenter.de>*
12433 * Change internals of the EC library so that the functions
12434 EC_GROUP_set_generator()
12435 EC_GROUP_get_generator()
12436 EC_GROUP_get_order()
12437 EC_GROUP_get_cofactor()
12438 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12439 to methods, which would lead to unnecessary code duplication when
12440 adding different types of curves.
12442 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12444 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12445 arithmetic, and such that modified wNAFs are generated
12446 (which avoid length expansion in many cases).
12450 * Add a function EC_GROUP_check_discriminant() (defined via
12451 EC_METHOD) that verifies that the curve discriminant is non-zero.
12453 Add a function EC_GROUP_check() that makes some sanity tests
12454 on a EC_GROUP, its generator and order. This includes
12455 EC_GROUP_check_discriminant().
12457 *Nils Larsch <nla@trustcenter.de>*
12459 * Add ECDSA in new directory crypto/ecdsa/.
12461 Add applications 'openssl ecparam' and 'openssl ecdsa'
12462 (these are based on 'openssl dsaparam' and 'openssl dsa').
12464 ECDSA support is also included in various other files across the
12465 library. Most notably,
12466 - 'openssl req' now has a '-newkey ecdsa:file' option;
12467 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12468 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12469 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12470 them suitable for ECDSA where domain parameters must be
12471 extracted before the specific public key;
12472 - ECDSA engine support has been added.
12474 *Nils Larsch <nla@trustcenter.de>*
12476 * Include some named elliptic curves, and add OIDs from X9.62,
12477 SECG, and WAP/WTLS. Each curve can be obtained from the new
12479 EC_GROUP_new_by_curve_name(),
12480 and the list of available named curves can be obtained with
12481 EC_get_builtin_curves().
12482 Also add a 'curve_name' member to EC_GROUP objects, which can be
12484 EC_GROUP_set_curve_name()
12485 EC_GROUP_get_curve_name()
12487 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12489 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12490 was actually never needed) and in BN_mul(). The removal in BN_mul()
12491 required a small change in bn_mul_part_recursive() and the addition
12492 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12493 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12494 bn_sub_words() and bn_add_words() except they take arrays with
12499 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12501 * Cleanse PEM buffers before freeing them since they may contain
12504 *Benjamin Bennett <ben@psc.edu>*
12506 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12507 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12508 authentication-only ciphersuites.
12512 * Since AES128 and AES256 share a single mask bit in the logic of
12513 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12514 kludge to work properly if AES128 is available and AES256 isn't.
12518 * Expand security boundary to match 1.1.1 module.
12522 * Remove redundant features: hash file source, editing of test vectors
12523 modify fipsld to use external fips_premain.c signature.
12527 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12528 run algorithm test programs.
12532 * Make algorithm test programs more tolerant of whitespace.
12536 * Have SSL/TLS server implementation tolerate "mismatched" record
12537 protocol version while receiving ClientHello even if the
12538 ClientHello is fragmented. (The server can't insist on the
12539 particular protocol version it has chosen before the ServerHello
12540 message has informed the client about his choice.)
12544 * Load error codes if they are not already present instead of using a
12545 static variable. This allows them to be cleanly unloaded and reloaded.
12549 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12551 * Introduce limits to prevent malicious keys being able to
12552 cause a denial of service. ([CVE-2006-2940])
12554 *Steve Henson, Bodo Moeller*
12556 * Fix ASN.1 parsing of certain invalid structures that can result
12557 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12559 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12560 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12562 * Fix SSL client code which could crash if connecting to a
12563 malicious SSLv2 server. ([CVE-2006-4343])
12565 *Tavis Ormandy and Will Drewry, Google Security Team*
12567 * Change ciphersuite string processing so that an explicit
12568 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12569 will no longer include "AES128-SHA"), and any other similar
12570 ciphersuite (same bitmap) from *other* protocol versions (so that
12571 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12572 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12573 changes from 0.9.8b and 0.9.8d.
12577 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12579 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12580 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12582 * Change the Unix randomness entropy gathering to use poll() when
12583 possible instead of select(), since the latter has some
12584 undesirable limitations.
12586 *Darryl Miles via Richard Levitte and Bodo Moeller*
12588 * Disable rogue ciphersuites:
12590 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12591 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12592 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12594 The latter two were purportedly from
12595 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12598 Also deactivate the remaining ciphersuites from
12599 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12600 unofficial, and the ID has long expired.
12604 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12605 dual-core machines) and other potential thread-safety issues.
12609 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12611 * Adapt fipsld and the build system to link against the validated FIPS
12612 module in FIPS mode.
12616 * Fixes for VC++ 2005 build under Windows.
12620 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12621 from a Windows bash shell such as MSYS. It is autodetected from the
12622 "config" script when run from a VC++ environment. Modify standard VC++
12623 build to use fipscanister.o from the GNU make build.
12627 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12629 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12630 The value now differs depending on if you build for FIPS or not.
12631 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12632 safely run with a non-FIPSed libcrypto, as it may crash because of
12633 the difference induced by this change.
12637 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12639 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12640 (part of SSL_OP_ALL). This option used to disable the
12641 countermeasure against man-in-the-middle protocol-version
12642 rollback in the SSL 2.0 server implementation, which is a bad
12643 idea. ([CVE-2005-2969])
12645 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12646 for Information Security, National Institute of Advanced Industrial
12647 Science and Technology [AIST, Japan)]*
12649 * Minimal support for X9.31 signatures and PSS padding modes. This is
12650 mainly for FIPS compliance and not fully integrated at this stage.
12654 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12655 the exponentiation using a fixed-length exponent. (Otherwise,
12656 the information leaked through timing could expose the secret key
12657 after many signatures; cf. Bleichenbacher's attack on DSA with
12662 * Make a new fixed-window mod_exp implementation the default for
12663 RSA, DSA, and DH private-key operations so that the sequence of
12664 squares and multiplies and the memory access pattern are
12665 independent of the particular secret key. This will mitigate
12666 cache-timing and potential related attacks.
12668 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12669 and this is automatically used by BN_mod_exp_mont() if the new flag
12670 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12671 will use this BN flag for private exponents unless the flag
12672 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12673 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12675 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12677 * Change the client implementation for SSLv23_method() and
12678 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12679 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12680 (Previously, the SSL 2.0 backwards compatible Client Hello
12681 message format would be used even with SSL_OP_NO_SSLv2.)
12685 * Add support for smime-type MIME parameter in S/MIME messages which some
12690 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12691 a threadsafe manner. Modify rsa code to use new function and add calls
12692 to dsa and dh code (which had race conditions before).
12696 * Include the fixed error library code in the C error file definitions
12697 instead of fixing them up at runtime. This keeps the error code
12698 structures constant.
12702 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12704 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12707 * Fixes for newer kerberos headers. NB: the casts are needed because
12708 the 'length' field is signed on one version and unsigned on another
12709 with no (?) obvious way to tell the difference, without these VC++
12710 complains. Also the "definition" of FAR (blank) is no longer included
12711 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12712 some needed definitions.
12716 * Undo Cygwin change.
12720 * Added support for proxy certificates according to RFC 3820.
12721 Because they may be a security thread to unaware applications,
12722 they must be explicitly allowed in run-time. See
12723 docs/HOWTO/proxy_certificates.txt for further information.
12727 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12729 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12730 server and client random values. Previously
12731 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12732 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12734 This change has negligible security impact because:
12736 1. Server and client random values still have 24 bytes of pseudo random
12739 2. Server and client random values are sent in the clear in the initial
12742 3. The master secret is derived using the premaster secret (48 bytes in
12743 size for static RSA ciphersuites) as well as client server and random
12746 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12749 *Stephen Henson, reported by UK NISCC*
12751 * Use Windows randomness collection on Cygwin.
12755 * Fix hang in EGD/PRNGD query when communication socket is closed
12756 prematurely by EGD/PRNGD.
12758 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12760 * Prompt for pass phrases when appropriate for PKCS12 input format.
12764 * Back-port of selected performance improvements from development
12765 branch, as well as improved support for PowerPC platforms.
12769 * Add lots of checks for memory allocation failure, error codes to indicate
12770 failure and freeing up memory if a failure occurs.
12772 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12774 * Add new -passin argument to dgst.
12778 * Perform some character comparisons of different types in X509_NAME_cmp:
12779 this is needed for some certificates that re-encode DNs into UTF8Strings
12780 (in violation of RFC3280) and can't or won't issue name rollover
12785 * Make an explicit check during certificate validation to see that
12786 the CA setting in each certificate on the chain is correct. As a
12787 side effect always do the following basic checks on extensions,
12788 not just when there's an associated purpose to the check:
12790 - if there is an unhandled critical extension (unless the user
12791 has chosen to ignore this fault)
12792 - if the path length has been exceeded (if one is set at all)
12793 - that certain extensions fit the associated purpose (if one has
12798 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12800 * Avoid a race condition when CRLs are checked in a multi threaded
12801 environment. This would happen due to the reordering of the revoked
12802 entries during signature checking and serial number lookup. Now the
12803 encoding is cached and the serial number sort performed under a lock.
12804 Add new STACK function sk_is_sorted().
12808 * Add Delta CRL to the extension code.
12812 * Various fixes to s3_pkt.c so alerts are sent properly.
12814 *David Holmes <d.holmes@f5.com>*
12816 * Reduce the chances of duplicate issuer name and serial numbers (in
12817 violation of RFC3280) using the OpenSSL certificate creation utilities.
12818 This is done by creating a random 64 bit value for the initial serial
12819 number when a serial number file is created or when a self signed
12820 certificate is created using 'openssl req -x509'. The initial serial
12821 number file is created using 'openssl x509 -next_serial' in CA.pl
12822 rather than being initialized to 1.
12826 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12828 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12829 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12831 *Joe Orton, Steve Henson*
12833 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12836 *Joe Orton, Steve Henson*
12838 * Make it possible to have multiple active certificates with the same
12839 subject in the CA index file. This is done only if the keyword
12840 'unique_subject' is set to 'no' in the main CA section (default
12841 if 'CA_default') of the configuration file. The value is saved
12842 with the database itself in a separate index attribute file,
12843 named like the index file with '.attr' appended to the name.
12847 * X509 verify fixes. Disable broken certificate workarounds when
12848 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12849 keyUsage extension present. Don't accept CRLs with unhandled critical
12850 extensions: since verify currently doesn't process CRL extensions this
12851 rejects a CRL with *any* critical extensions. Add new verify error codes
12856 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12857 A clarification of RFC2560 will require the use of OCTET STRINGs and
12858 some implementations cannot handle the current raw format. Since OpenSSL
12859 copies and compares OCSP nonces as opaque blobs without any attempt at
12860 parsing them this should not create any compatibility issues.
12864 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12865 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12866 this HMAC (and other) operations are several times slower than OpenSSL
12871 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12873 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12875 * Use the correct content when signing type "other".
12879 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12881 * Fix various bugs revealed by running the NISCC test suite:
12883 Stop out of bounds reads in the ASN1 code when presented with
12884 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12886 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12888 If verify callback ignores invalid public key errors don't try to check
12889 certificate signature with the NULL public key.
12893 * New -ignore_err option in ocsp application to stop the server
12894 exiting on the first error in a request.
12898 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12899 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12904 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12905 extra data after the compression methods not only for TLS 1.0
12906 but also for SSL 3.0 (as required by the specification).
12908 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12910 * Change X509_certificate_type() to mark the key as exported/exportable
12911 when it's 512 *bits* long, not 512 bytes.
12915 * Change AES_cbc_encrypt() so it outputs exact multiple of
12916 blocks during encryption.
12920 * Various fixes to base64 BIO and non blocking I/O. On write
12921 flushes were not handled properly if the BIO retried. On read
12922 data was not being buffered properly and had various logic bugs.
12923 This also affects blocking I/O when the data being decoded is a
12928 * Various S/MIME bugfixes and compatibility changes:
12929 output correct application/pkcs7 MIME type if
12930 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12931 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12932 of files as .eml work). Correctly handle very long lines in MIME
12937 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12939 * Countermeasure against the Klima-Pokorny-Rosa extension of
12940 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12941 a protocol version number mismatch like a decryption error
12942 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12946 * Turn on RSA blinding by default in the default implementation
12947 to avoid a timing attack. Applications that don't want it can call
12948 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12949 They would be ill-advised to do so in most cases.
12951 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12953 * Change RSA blinding code so that it works when the PRNG is not
12954 seeded (in this case, the secret RSA exponent is abused as
12955 an unpredictable seed -- if it is not unpredictable, there
12956 is no point in blinding anyway). Make RSA blinding thread-safe
12957 by remembering the creator's thread ID in rsa->blinding and
12958 having all other threads use local one-time blinding factors
12959 (this requires more computation than sharing rsa->blinding, but
12960 avoids excessive locking; and if an RSA object is not shared
12961 between threads, blinding will still be very fast).
12965 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12966 ENGINE as defaults for all supported algorithms irrespective of
12967 the 'flags' parameter. 'flags' is now honoured, so applications
12968 should make sure they are passing it correctly.
12972 * Target "mingw" now allows native Windows code to be generated in
12973 the Cygwin environment as well as with the MinGW compiler.
12977 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12979 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12980 via timing by performing a MAC computation even if incorrect
12981 block cipher padding has been found. This is a countermeasure
12982 against active attacks where the attacker has to distinguish
12983 between bad padding and a MAC verification error. ([CVE-2003-0078])
12985 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12986 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12987 Martin Vuagnoux (EPFL, Ilion)*
12989 * Make the no-err option work as intended. The intention with no-err
12990 is not to have the whole error stack handling routines removed from
12991 libcrypto, it's only intended to remove all the function name and
12992 reason texts, thereby removing some of the footprint that may not
12993 be interesting if those errors aren't displayed anyway.
12995 NOTE: it's still possible for any application or module to have its
12996 own set of error texts inserted. The routines are there, just not
12997 used by default when no-err is given.
13001 * Add support for FreeBSD on IA64.
13003 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13005 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13006 Kerberos function mit_des_cbc_cksum(). Before this change,
13007 the value returned by DES_cbc_cksum() was like the one from
13008 mit_des_cbc_cksum(), except the bytes were swapped.
13010 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13012 * Allow an application to disable the automatic SSL chain building.
13013 Before this a rather primitive chain build was always performed in
13014 ssl3_output_cert_chain(): an application had no way to send the
13015 correct chain if the automatic operation produced an incorrect result.
13017 Now the chain builder is disabled if either:
13019 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13021 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13023 The reasoning behind this is that an application would not want the
13024 auto chain building to take place if extra chain certificates are
13025 present and it might also want a means of sending no additional
13026 certificates (for example the chain has two certificates and the
13031 * Add the possibility to build without the ENGINE framework.
13033 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13035 * Under Win32 gmtime() can return NULL: check return value in
13036 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13040 * DSA routines: under certain error conditions uninitialized BN objects
13041 could be freed. Solution: make sure initialization is performed early
13042 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13043 Nils Larsch <nla@trustcenter.de> via PR#459)
13047 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13048 checked on reconnect on the client side, therefore session resumption
13049 could still fail with a "ssl session id is different" error. This
13050 behaviour is masked when SSL_OP_ALL is used due to
13051 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13052 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13053 followup to PR #377.
13057 * IA-32 assembler support enhancements: unified ELF targets, support
13058 for SCO/Caldera platforms, fix for Cygwin shared build.
13062 * Add support for FreeBSD on sparc64. As a consequence, support for
13063 FreeBSD on non-x86 processors is separate from x86 processors on
13064 the config script, much like the NetBSD support.
13066 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13068 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13070 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13073 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13074 code (06) was taken as the first octet of the session ID and the last
13075 octet was ignored consequently. As a result SSLv2 client side session
13076 caching could not have worked due to the session ID mismatch between
13078 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13083 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13084 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13089 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13090 seems that in spite of existing for more than a year, many application
13091 author have done nothing to provide the necessary callbacks, which
13092 means that this particular engine will not work properly anywhere.
13093 This is a very unfortunate situation which forces us, in the name
13094 of usability, to give the hw_ncipher.c a static lock, which is part
13096 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13097 appear in 0.9.8 or later. We EXPECT application authors to have
13098 dealt properly with this when 0.9.8 is released (unless we actually
13099 make such changes in the libcrypto locking code that changes will
13100 have to be made anyway).
13104 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13105 octets have been read, EOF or an error occurs. Without this change
13106 some truncated ASN1 structures will not produce an error.
13110 * Disable Heimdal support, since it hasn't been fully implemented.
13111 Still give the possibility to force the use of Heimdal, but with
13112 warnings and a request that patches get sent to openssl-dev.
13116 * Add the VC-CE target, introduce the WINCE sysname, and add
13117 INSTALL.WCE and appropriate conditionals to make it build.
13119 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13121 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13122 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13123 edit numbers of the version.
13125 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13127 * Introduce safe string copy and catenation functions
13128 (BUF_strlcpy() and BUF_strlcat()).
13130 *Ben Laurie (CHATS) and Richard Levitte*
13132 * Avoid using fixed-size buffers for one-line DNs.
13134 *Ben Laurie (CHATS)*
13136 * Add BUF_MEM_grow_clean() to avoid information leakage when
13137 resizing buffers containing secrets, and use where appropriate.
13139 *Ben Laurie (CHATS)*
13141 * Avoid using fixed size buffers for configuration file location.
13143 *Ben Laurie (CHATS)*
13145 * Avoid filename truncation for various CA files.
13147 *Ben Laurie (CHATS)*
13149 * Use sizeof in preference to magic numbers.
13151 *Ben Laurie (CHATS)*
13153 * Avoid filename truncation in cert requests.
13155 *Ben Laurie (CHATS)*
13157 * Add assertions to check for (supposedly impossible) buffer
13160 *Ben Laurie (CHATS)*
13162 * Don't cache truncated DNS entries in the local cache (this could
13163 potentially lead to a spoofing attack).
13165 *Ben Laurie (CHATS)*
13167 * Fix various buffers to be large enough for hex/decimal
13168 representations in a platform independent manner.
13170 *Ben Laurie (CHATS)*
13172 * Add CRYPTO_realloc_clean() to avoid information leakage when
13173 resizing buffers containing secrets, and use where appropriate.
13175 *Ben Laurie (CHATS)*
13177 * Add BIO_indent() to avoid much slightly worrying code to do
13180 *Ben Laurie (CHATS)*
13182 * Convert sprintf()/BIO_puts() to BIO_printf().
13184 *Ben Laurie (CHATS)*
13186 * buffer_gets() could terminate with the buffer only half
13189 *Ben Laurie (CHATS)*
13191 * Add assertions to prevent user-supplied crypto functions from
13192 overflowing internal buffers by having large block sizes, etc.
13194 *Ben Laurie (CHATS)*
13196 * New OPENSSL_assert() macro (similar to assert(), but enabled
13199 *Ben Laurie (CHATS)*
13201 * Eliminate unused copy of key in RC4.
13203 *Ben Laurie (CHATS)*
13205 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13207 *Ben Laurie (CHATS)*
13209 * Fix off-by-one error in EGD path.
13211 *Ben Laurie (CHATS)*
13213 * If RANDFILE path is too long, ignore instead of truncating.
13215 *Ben Laurie (CHATS)*
13217 * Eliminate unused and incorrectly sized X.509 structure
13220 *Ben Laurie (CHATS)*
13222 * Eliminate unused and dangerous function knumber().
13224 *Ben Laurie (CHATS)*
13226 * Eliminate unused and dangerous structure, KSSL_ERR.
13228 *Ben Laurie (CHATS)*
13230 * Protect against overlong session ID context length in an encoded
13231 session object. Since these are local, this does not appear to be
13234 *Ben Laurie (CHATS)*
13236 * Change from security patch (see 0.9.6e below) that did not affect
13237 the 0.9.6 release series:
13239 Remote buffer overflow in SSL3 protocol - an attacker could
13240 supply an oversized master key in Kerberos-enabled versions.
13243 *Ben Laurie (CHATS)*
13245 * Change the SSL kerb5 codes to match RFC 2712.
13249 * Make -nameopt work fully for req and add -reqopt switch.
13251 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13253 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13255 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13257 * Make sure tests can be performed even if the corresponding algorithms
13258 have been removed entirely. This was also the last step to make
13259 OpenSSL compilable with DJGPP under all reasonable conditions.
13261 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13263 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13264 to allow version independent disabling of normally unselected ciphers,
13265 which may be activated as a side-effect of selecting a single cipher.
13267 (E.g., cipher list string "RSA" enables ciphersuites that are left
13268 out of "ALL" because they do not provide symmetric encryption.
13269 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13271 *Lutz Jaenicke, Bodo Moeller*
13273 * Add appropriate support for separate platform-dependent build
13274 directories. The recommended way to make a platform-dependent
13275 build directory is the following (tested on Linux), maybe with
13278 # Place yourself outside of the OpenSSL source tree. In
13279 # this example, the environment variable OPENSSL_SOURCE
13280 # is assumed to contain the absolute OpenSSL source directory.
13281 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13282 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13283 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13284 mkdir -p `dirname $F`
13285 ln -s $OPENSSL_SOURCE/$F $F
13288 To be absolutely sure not to disturb the source tree, a "make clean"
13289 is a good thing. If it isn't successful, don't worry about it,
13290 it probably means the source directory is very clean.
13294 * Make sure any ENGINE control commands make local copies of string
13295 pointers passed to them whenever necessary. Otherwise it is possible
13296 the caller may have overwritten (or deallocated) the original string
13297 data when a later ENGINE operation tries to use the stored values.
13299 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13301 * Improve diagnostics in file reading and command-line digests.
13303 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13305 * Add AES modes CFB and OFB to the object database. Correct an
13306 error in AES-CFB decryption.
13310 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13311 allows existing EVP_CIPHER_CTX structures to be reused after
13312 calling `EVP_*Final()`. This behaviour is used by encryption
13313 BIOs and some applications. This has the side effect that
13314 applications must explicitly clean up cipher contexts with
13315 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13319 * Check the values of dna and dnb in bn_mul_recursive before calling
13320 bn_mul_comba (a non zero value means the a or b arrays do not contain
13321 n2 elements) and fallback to bn_mul_normal if either is not zero.
13325 * Fix escaping of non-ASCII characters when using the -subj option
13326 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13330 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13331 form for "surname", serialNumber has no short form.
13332 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13333 therefore remove "mail" short name for "internet 7".
13334 The OID for unique identifiers in X509 certificates is
13335 x500UniqueIdentifier, not uniqueIdentifier.
13336 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13340 * Add an "init" command to the ENGINE config module and auto initialize
13341 ENGINEs. Without any "init" command the ENGINE will be initialized
13342 after all ctrl commands have been executed on it. If init=1 the
13343 ENGINE is initialized at that point (ctrls before that point are run
13344 on the uninitialized ENGINE and after on the initialized one). If
13345 init=0 then the ENGINE will not be initialized at all.
13349 * Fix the 'app_verify_callback' interface so that the user-defined
13350 argument is actually passed to the callback: In the
13351 SSL_CTX_set_cert_verify_callback() prototype, the callback
13352 declaration has been changed from
13355 int (*cb)(X509_STORE_CTX *,void *);
13356 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13357 i=s->ctx->app_verify_callback(&ctx)
13358 has been changed into
13359 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13361 To update applications using SSL_CTX_set_cert_verify_callback(),
13362 a dummy argument can be added to their callback functions.
13364 *D. K. Smetters <smetters@parc.xerox.com>*
13366 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13368 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13370 * Add and OPENSSL_LOAD_CONF define which will cause
13371 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13372 This allows older applications to transparently support certain
13373 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13374 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13375 load the config file and OPENSSL_add_all_algorithms_conf() which will
13376 always load it have also been added.
13380 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13381 Adjust NIDs and EVP layer.
13383 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13385 * Config modules support in openssl utility.
13387 Most commands now load modules from the config file,
13388 though in a few (such as version) this isn't done
13389 because it couldn't be used for anything.
13391 In the case of ca and req the config file used is
13392 the same as the utility itself: that is the -config
13393 command line option can be used to specify an
13398 * Move default behaviour from OPENSSL_config(). If appname is NULL
13399 use "openssl_conf" if filename is NULL use default openssl config file.
13403 * Add an argument to OPENSSL_config() to allow the use of an alternative
13404 config section name. Add a new flag to tolerate a missing config file
13405 and move code to CONF_modules_load_file().
13409 * Support for crypto accelerator cards from Accelerated Encryption
13410 Processing, www.aep.ie. (Use engine 'aep')
13411 The support was copied from 0.9.6c [engine] and adapted/corrected
13412 to work with the new engine framework.
13414 *AEP Inc. and Richard Levitte*
13416 * Support for SureWare crypto accelerator cards from Baltimore
13417 Technologies. (Use engine 'sureware')
13418 The support was copied from 0.9.6c [engine] and adapted
13419 to work with the new engine framework.
13423 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13424 make the newer ENGINE framework commands for the CHIL engine work.
13426 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13428 * Make it possible to produce shared libraries on ReliantUNIX.
13430 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13432 * Add the configuration target debug-linux-ppro.
13433 Make 'openssl rsa' use the general key loading routines
13434 implemented in `apps.c`, and make those routines able to
13435 handle the key format FORMAT_NETSCAPE and the variant
13438 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13440 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13442 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13444 * Add -keyform to rsautl, and document -engine.
13446 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13448 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13449 BIO_R_NO_SUCH_FILE error code rather than the generic
13450 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13454 * Add new functions
13455 ERR_peek_last_error
13456 ERR_peek_last_error_line
13457 ERR_peek_last_error_line_data.
13458 These are similar to
13460 ERR_peek_error_line
13461 ERR_peek_error_line_data,
13462 but report on the latest error recorded rather than the first one
13463 still in the error queue.
13465 *Ben Laurie, Bodo Moeller*
13467 * default_algorithms option in ENGINE config module. This allows things
13469 default_algorithms = ALL
13470 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13474 * Preliminary ENGINE config module.
13478 * New experimental application configuration code.
13482 * Change the AES code to follow the same name structure as all other
13483 symmetric ciphers, and behave the same way. Move everything to
13484 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13486 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13488 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13490 *Ben Laurie and Theo de Raadt*
13492 * Add option to output public keys in req command.
13494 *Massimiliano Pala madwolf@openca.org*
13496 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13497 (up to about 10% better than before for P-192 and P-224).
13501 * New functions/macros
13503 SSL_CTX_set_msg_callback(ctx, cb)
13504 SSL_CTX_set_msg_callback_arg(ctx, arg)
13505 SSL_set_msg_callback(ssl, cb)
13506 SSL_set_msg_callback_arg(ssl, arg)
13508 to request calling a callback function
13510 void cb(int write_p, int version, int content_type,
13511 const void *buf, size_t len, SSL *ssl, void *arg)
13513 whenever a protocol message has been completely received
13514 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13515 protocol version according to which the SSL library interprets
13516 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13517 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13518 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13519 specification (change_cipher_spec(20), alert(21), handshake(22)).
13520 'buf' and 'len' point to the actual message, 'ssl' to the
13521 SSL object, and 'arg' is the application-defined value set by
13522 SSL[_CTX]_set_msg_callback_arg().
13524 'openssl s_client' and 'openssl s_server' have new '-msg' options
13525 to enable a callback that displays all protocol messages.
13529 * Change the shared library support so shared libraries are built as
13530 soon as the corresponding static library is finished, and thereby get
13531 openssl and the test programs linked against the shared library.
13532 This still only happens when the keyword "shard" has been given to
13533 the configuration scripts.
13535 NOTE: shared library support is still an experimental thing, and
13536 backward binary compatibility is still not guaranteed.
13538 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13540 * Add support for Subject Information Access extension.
13542 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13544 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13545 additional bytes when new memory had to be allocated, not just
13546 when reusing an existing buffer.
13550 * New command line and configuration option 'utf8' for the req command.
13551 This allows field values to be specified as UTF8 strings.
13555 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13556 runs for the former and machine-readable output for the latter.
13560 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13561 of the e-mail address in the DN (i.e., it will go into a certificate
13562 extension only). The new configuration file option 'email_in_dn = no'
13563 has the same effect.
13565 *Massimiliano Pala madwolf@openca.org*
13567 * Change all functions with names starting with `des_` to be starting
13568 with `DES_` instead. Add wrappers that are compatible with libdes,
13569 but are named `_ossl_old_des_*`. Finally, add macros that map the
13570 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13571 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13572 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13575 Since we provide two compatibility mappings, the user needs to
13576 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13577 compatibility is desired. The default (i.e., when that macro
13578 isn't defined) is OpenSSL 0.9.6c compatibility.
13580 There are also macros that enable and disable the support of old
13581 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13582 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13583 are defined, the default will apply: to support the old des routines.
13585 In either case, one must include openssl/des.h to get the correct
13586 definitions. Do not try to just include openssl/des_old.h, that
13589 NOTE: This is a major break of an old API into a new one. Software
13590 authors are encouraged to switch to the `DES_` style functions. Some
13591 time in the future, des_old.h and the libdes compatibility functions
13592 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13593 default), and then completely removed.
13597 * Test for certificates which contain unsupported critical extensions.
13598 If such a certificate is found during a verify operation it is
13599 rejected by default: this behaviour can be overridden by either
13600 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13601 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13602 X509_supported_extension() has also been added which returns 1 if a
13603 particular extension is supported.
13607 * Modify the behaviour of EVP cipher functions in similar way to digests
13608 to retain compatibility with existing code.
13612 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13613 compatibility with existing code. In particular the 'ctx' parameter does
13614 not have to be to be initialized before the call to EVP_DigestInit() and
13615 it is tidied up after a call to EVP_DigestFinal(). New function
13616 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13617 EVP_MD_CTX_copy() changed to not require the destination to be
13618 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13619 requires the destination to be valid.
13621 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13622 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13626 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13627 so that complete 'Handshake' protocol structures are kept in memory
13628 instead of overwriting 'msg_type' and 'length' with 'body' data.
13632 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13634 *Massimo Santin via Richard Levitte*
13636 * Major restructuring to the underlying ENGINE code. This includes
13637 reduction of linker bloat, separation of pure "ENGINE" manipulation
13638 (initialisation, etc) from functionality dealing with implementations
13639 of specific crypto interfaces. This change also introduces integrated
13640 support for symmetric ciphers and digest implementations - so ENGINEs
13641 can now accelerate these by providing EVP_CIPHER and EVP_MD
13642 implementations of their own. This is detailed in
13643 [crypto/engine/README.md](crypto/engine/README.md)
13644 as it couldn't be adequately described here. However, there are a few
13645 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13646 were changed in the original introduction of ENGINE code have now
13647 reverted back - the hooking from this code to ENGINE is now a good
13648 deal more passive and at run-time, operations deal directly with
13649 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13650 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13651 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13652 they were not being used by the framework as there is no concept of a
13653 BIGNUM_METHOD and they could not be generalised to the new
13654 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13655 ENGINE_cpy() has been removed as it cannot be consistently defined in
13660 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13664 * Change mkdef.pl to sort symbols that get the same entry number,
13665 and make sure the automatically generated functions `ERR_load_*`
13666 become part of libeay.num as well.
13670 * New function SSL_renegotiate_pending(). This returns true once
13671 renegotiation has been requested (either SSL_renegotiate() call
13672 or HelloRequest/ClientHello received from the peer) and becomes
13673 false once a handshake has been completed.
13674 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13675 sends a HelloRequest, but does not ensure that a handshake takes
13676 place. SSL_renegotiate_pending() is useful for checking if the
13677 client has followed the request.)
13681 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13682 By default, clients may request session resumption even during
13683 renegotiation (if session ID contexts permit); with this option,
13684 session resumption is possible only in the first handshake.
13686 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13687 more bits available for options that should not be part of
13688 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13692 * Add some demos for certificate and certificate request creation.
13696 * Make maximum certificate chain size accepted from the peer application
13697 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13698 "Douglas E. Engert" <deengert@anl.gov>.
13702 * Add support for shared libraries for Unixware-7
13703 (Boyd Lynn Gerber <gerberb@zenez.com>).
13707 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13708 be done prior to destruction. Use this to unload error strings from
13709 ENGINEs that load their own error strings. NB: This adds two new API
13710 functions to "get" and "set" this destroy handler in an ENGINE.
13714 * Alter all existing ENGINE implementations (except "openssl" and
13715 "openbsd") to dynamically instantiate their own error strings. This
13716 makes them more flexible to be built both as statically-linked ENGINEs
13717 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13718 Also, add stub code to each that makes building them as self-contained
13719 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13723 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13724 implementations into applications that are completely implemented in
13725 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13726 commands that can be used to configure what shared-library to load and
13727 to control aspects of the way it is handled. Also, made an update to
13728 the [README-Engine.md](README-Engine.md) file
13729 that brings its information up-to-date and
13730 provides some information and instructions on the "dynamic" ENGINE
13731 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13735 * Make it possible to unload ranges of ERR strings with a new
13736 "ERR_unload_strings" function.
13740 * Add a copy() function to EVP_MD.
13744 * Make EVP_MD routines take a context pointer instead of just the
13745 md_data void pointer.
13749 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13750 that the digest can only process a single chunk of data
13751 (typically because it is provided by a piece of
13752 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13753 is only going to provide a single chunk of data, and hence the
13754 framework needn't accumulate the data for oneshot drivers.
13758 * As with "ERR", make it possible to replace the underlying "ex_data"
13759 functions. This change also alters the storage and management of global
13760 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13761 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13762 index counters. The API functions that use this state have been changed
13763 to take a "class_index" rather than pointers to the class's local STACK
13764 and counter, and there is now an API function to dynamically create new
13765 classes. This centralisation allows us to (a) plug a lot of the
13766 thread-safety problems that existed, and (b) makes it possible to clean
13767 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13768 such data would previously have always leaked in application code and
13769 workarounds were in place to make the memory debugging turn a blind eye
13770 to it. Application code that doesn't use this new function will still
13771 leak as before, but their memory debugging output will announce it now
13772 rather than letting it slide.
13774 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13775 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13776 has a return value to indicate success or failure.
13780 * Make it possible to replace the underlying "ERR" functions such that the
13781 global state (2 LHASH tables and 2 locks) is only used by the "default"
13782 implementation. This change also adds two functions to "get" and "set"
13783 the implementation prior to it being automatically set the first time
13784 any other ERR function takes place. Ie. an application can call "get",
13785 pass the return value to a module it has just loaded, and that module
13786 can call its own "set" function using that value. This means the
13787 module's "ERR" operations will use (and modify) the error state in the
13788 application and not in its own statically linked copy of OpenSSL code.
13792 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13793 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13794 the operation, and provides a more encapsulated way for external code
13795 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13796 to use these functions rather than manually incrementing the counts.
13798 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13802 * Add EVP test program.
13806 * Add symmetric cipher support to ENGINE. Expect the API to change!
13810 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13811 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13812 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13813 These allow a CRL to be built without having to access X509_CRL fields
13814 directly. Modify 'ca' application to use new functions.
13818 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13819 bug workarounds. Rollback attack detection is a security feature.
13820 The problem will only arise on OpenSSL servers when TLSv1 is not
13821 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13822 Software authors not wanting to support TLSv1 will have special reasons
13823 for their choice and can explicitly enable this option.
13825 *Bodo Moeller, Lutz Jaenicke*
13827 * Rationalise EVP so it can be extended: don't include a union of
13828 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13829 (similar to those existing for EVP_CIPHER_CTX).
13834 EVP_MD_CTX_init(&md); /* new function call */
13835 EVP_DigestInit(&md, EVP_sha1());
13836 EVP_DigestUpdate(&md, in, len);
13837 EVP_DigestFinal(&md, out, NULL);
13838 EVP_MD_CTX_cleanup(&md); /* new function call */
13842 * Make DES key schedule conform to the usual scheme, as well as
13843 correcting its structure. This means that calls to DES functions
13844 now have to pass a pointer to a des_key_schedule instead of a
13845 plain des_key_schedule (which was actually always a pointer
13848 des_key_schedule ks;
13850 des_set_key_checked(..., &ks);
13851 des_ncbc_encrypt(..., &ks, ...);
13853 (Note that a later change renames 'des_...' into 'DES_...'.)
13857 * Initial reduction of linker bloat: the use of some functions, such as
13858 PEM causes large amounts of unused functions to be linked in due to
13859 poor organisation. For example pem_all.c contains every PEM function
13860 which has a knock on effect of linking in large amounts of (unused)
13861 ASN1 code. Grouping together similar functions and splitting unrelated
13862 functions prevents this.
13866 * Cleanup of EVP macros.
13870 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13871 correct `_ecb suffix`.
13875 * Add initial OCSP responder support to ocsp application. The
13876 revocation information is handled using the text based index
13877 use by the ca application. The responder can either handle
13878 requests generated internally, supplied in files (for example
13879 via a CGI script) or using an internal minimal server.
13883 * Add configuration choices to get zlib compression for TLS.
13887 * Changes to Kerberos SSL for RFC 2712 compliance:
13888 1. Implemented real KerberosWrapper, instead of just using
13889 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13890 2. Implemented optional authenticator field of KerberosWrapper.
13892 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13893 and authenticator structs; see crypto/krb5/.
13895 Generalized Kerberos calls to support multiple Kerberos libraries.
13896 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13897 via Richard Levitte*
13899 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13900 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13901 values for each of the key sizes rather than having just
13902 parameters (and 'speed' generating keys each time).
13906 * Speed up EVP routines.
13909 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13910 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13911 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13912 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13914 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13915 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13916 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13919 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13921 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13925 * Added the OS2-EMX target.
13927 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13929 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13930 New functions to support `NCONF` routines in extension code.
13931 New function `CONF_set_nconf()`
13932 to allow functions which take an `NCONF` to also handle the old `LHASH`
13933 structure: this means that the old `CONF` compatible routines can be
13934 retained (in particular w.rt. extensions) without having to duplicate the
13935 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13939 * Enhance the general user interface with mechanisms for inner control
13940 and with possibilities to have yes/no kind of prompts.
13944 * Change all calls to low-level digest routines in the library and
13945 applications to use EVP. Add missing calls to HMAC_cleanup() and
13946 don't assume HMAC_CTX can be copied using memcpy().
13948 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13950 * Add the possibility to control engines through control names but with
13951 arbitrary arguments instead of just a string.
13952 Change the key loaders to take a UI_METHOD instead of a callback
13953 function pointer. NOTE: this breaks binary compatibility with earlier
13954 versions of OpenSSL [engine].
13955 Adapt the nCipher code for these new conditions and add a card insertion
13960 * Enhance the general user interface with mechanisms to better support
13961 dialog box interfaces, application-defined prompts, the possibility
13962 to use defaults (for example default passwords from somewhere else)
13963 and interrupts/cancellations.
13967 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13968 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13972 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13973 tidy up some unnecessarily weird code in 'sk_new()').
13975 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13977 * Change the key loading routines for ENGINEs to use the same kind
13978 callback (pem_password_cb) as all other routines that need this
13983 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13984 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13985 than this minimum value is recommended.
13989 * New random seeder for OpenVMS, using the system process statistics
13990 that are easily reachable.
13994 * Windows apparently can't transparently handle global
13995 variables defined in DLLs. Initialisations such as:
13997 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13999 won't compile. This is used by the any applications that need to
14000 declare their own ASN1 modules. This was fixed by adding the option
14001 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14002 needed for static libraries under Win32.
14006 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14007 setting of purpose and trust fields. New X509_STORE trust and
14008 purpose functions and tidy up setting in other SSL functions.
14012 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14013 structure. These are inherited by X509_STORE_CTX when it is
14014 initialised. This allows various defaults to be set in the
14015 X509_STORE structure (such as flags for CRL checking and custom
14016 purpose or trust settings) for functions which only use X509_STORE_CTX
14017 internally such as S/MIME.
14019 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14020 trust settings if they are not set in X509_STORE. This allows X509_STORE
14021 purposes and trust (in S/MIME for example) to override any set by default.
14023 Add command line options for CRL checking to smime, s_client and s_server
14028 * Initial CRL based revocation checking. If the CRL checking flag(s)
14029 are set then the CRL is looked up in the X509_STORE structure and
14030 its validity and signature checked, then if the certificate is found
14031 in the CRL the verify fails with a revoked error.
14033 Various new CRL related callbacks added to X509_STORE_CTX structure.
14035 Command line options added to 'verify' application to support this.
14037 This needs some additional work, such as being able to handle multiple
14038 CRLs with different times, extension based lookup (rather than just
14039 by subject name) and ultimately more complete V2 CRL extension
14044 * Add a general user interface API (crypto/ui/). This is designed
14045 to replace things like des_read_password and friends (backward
14046 compatibility functions using this new API are provided).
14047 The purpose is to remove prompting functions from the DES code
14048 section as well as provide for prompting through dialog boxes in
14049 a window system and the like.
14053 * Add "ex_data" support to ENGINE so implementations can add state at a
14054 per-structure level rather than having to store it globally.
14058 * Make it possible for ENGINE structures to be copied when retrieved by
14059 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14060 This causes the "original" ENGINE structure to act like a template,
14061 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14062 operational state can be localised to each ENGINE structure, despite the
14063 fact they all share the same "methods". New ENGINE structures returned in
14064 this case have no functional references and the return value is the single
14065 structural reference. This matches the single structural reference returned
14066 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14071 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14072 needs to match any other type at all we need to manually clear the
14077 * Changes to the "openssl engine" utility to include;
14078 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14079 about an ENGINE's available control commands.
14080 - executing control commands from command line arguments using the
14081 '-pre' and '-post' switches. '-post' is only used if '-t' is
14082 specified and the ENGINE is successfully initialised. The syntax for
14083 the individual commands are colon-separated, for example;
14084 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14088 * New dynamic control command support for ENGINEs. ENGINEs can now
14089 declare their own commands (numbers), names (strings), descriptions,
14090 and input types for run-time discovery by calling applications. A
14091 subset of these commands are implicitly classed as "executable"
14092 depending on their input type, and only these can be invoked through
14093 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14094 can be based on user input, config files, etc). The distinction is
14095 that "executable" commands cannot return anything other than a boolean
14096 result and can only support numeric or string input, whereas some
14097 discoverable commands may only be for direct use through
14098 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14099 pointers, or other custom uses. The "executable" commands are to
14100 support parameterisations of ENGINE behaviour that can be
14101 unambiguously defined by ENGINEs and used consistently across any
14102 OpenSSL-based application. Commands have been added to all the
14103 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14104 control over shared-library paths without source code alterations.
14108 * Changed all ENGINE implementations to dynamically allocate their
14109 ENGINEs rather than declaring them statically. Apart from this being
14110 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14111 this also allows the implementations to compile without using the
14112 internal engine_int.h header.
14116 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14117 'const' value. Any code that should be able to modify a RAND_METHOD
14118 should already have non-const pointers to it (ie. they should only
14119 modify their own ones).
14123 * Made a variety of little tweaks to the ENGINE code.
14124 - "atalla" and "ubsec" string definitions were moved from header files
14125 to C code. "nuron" string definitions were placed in variables
14126 rather than hard-coded - allowing parameterisation of these values
14127 later on via ctrl() commands.
14128 - Removed unused "#if 0"'d code.
14129 - Fixed engine list iteration code so it uses ENGINE_free() to release
14130 structural references.
14131 - Constified the RAND_METHOD element of ENGINE structures.
14132 - Constified various get/set functions as appropriate and added
14133 missing functions (including a catch-all ENGINE_cpy that duplicates
14134 all ENGINE values onto a new ENGINE except reference counts/state).
14135 - Removed NULL parameter checks in get/set functions. Setting a method
14136 or function to NULL is a way of cancelling out a previously set
14137 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14138 and doesn't justify the extra error symbols and code.
14139 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14140 flags from engine_int.h to engine.h.
14141 - Changed prototypes for ENGINE handler functions (init(), finish(),
14142 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14146 * Implement binary inversion algorithm for BN_mod_inverse in addition
14147 to the algorithm using long division. The binary algorithm can be
14148 used only if the modulus is odd. On 32-bit systems, it is faster
14149 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14150 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14151 up to 450 bits. In 64-bit environments, the binary algorithm
14152 appears to be advantageous for much longer moduli; here we use it
14153 for moduli up to 2048 bits.
14157 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14158 could not support the combine flag in choice fields.
14162 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14163 extensions from a certificate request to the certificate.
14167 * Allow multiple 'certopt' and 'nameopt' options to be separated
14168 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14169 file: this allows the display of the certificate about to be
14170 signed to be customised, to allow certain fields to be included
14171 or excluded and extension details. The old system didn't display
14172 multicharacter strings properly, omitted fields not in the policy
14173 and couldn't display additional details such as extensions.
14177 * Function EC_POINTs_mul for multiple scalar multiplication
14178 of an arbitrary number of elliptic curve points
14179 \sum scalars[i]*points[i],
14180 optionally including the generator defined for the EC_GROUP:
14181 scalar*generator + \sum scalars[i]*points[i].
14183 EC_POINT_mul is a simple wrapper function for the typical case
14184 that the point list has just one item (besides the optional
14189 * First EC_METHODs for curves over GF(p):
14191 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14192 operations and provides various method functions that can also
14193 operate with faster implementations of modular arithmetic.
14195 EC_GFp_mont_method() reuses most functions that are part of
14196 EC_GFp_simple_method, but uses Montgomery arithmetic.
14198 *Bodo Moeller; point addition and point doubling
14199 implementation directly derived from source code provided by
14200 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14202 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14203 crypto/ec/ec_lib.c):
14205 Curves are EC_GROUP objects (with an optional group generator)
14206 based on EC_METHODs that are built into the library.
14208 Points are EC_POINT objects based on EC_GROUP objects.
14210 Most of the framework would be able to handle curves over arbitrary
14211 finite fields, but as there are no obvious types for fields other
14212 than GF(p), some functions are limited to that for now.
14216 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14217 that the file contains a complete HTTP response.
14221 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14222 change the def and num file printf format specifier from "%-40sXXX"
14223 to "%-39s XXX". The latter will always guarantee a space after the
14224 field while the former will cause them to run together if the field
14225 is 40 of more characters long.
14229 * Constify the cipher and digest 'method' functions and structures
14230 and modify related functions to take constant EVP_MD and EVP_CIPHER
14235 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14236 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14240 * Modify `EVP_Digest*()` routines so they now return values. Although the
14241 internal software routines can never fail additional hardware versions
14246 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14248 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14249 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14252 ERR_R_NESTED_ASN1_ERROR
14254 ERR_R_MISSING_ASN1_EOS
14255 were 4 .. 9, conflicting with
14256 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14258 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14259 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14261 Add new error code 'ERR_R_INTERNAL_ERROR'.
14265 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14270 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14271 sets the subject name for a new request or supersedes the
14272 subject name in a given request. Formats that can be parsed are
14273 'CN=Some Name, OU=myOU, C=IT'
14275 'CN=Some Name/OU=myOU/C=IT'.
14277 Add options '-batch' and '-verbose' to 'openssl req'.
14279 *Massimiliano Pala <madwolf@hackmasters.net>*
14281 * Introduce the possibility to access global variables through
14282 functions on platform were that's the best way to handle exporting
14283 global variables in shared libraries. To enable this functionality,
14284 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14285 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14286 is normally done by Configure or something similar).
14288 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14289 in the source file (foo.c) like this:
14291 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14292 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14294 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14295 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14297 OPENSSL_DECLARE_GLOBAL(int,foo);
14298 #define foo OPENSSL_GLOBAL_REF(foo)
14299 OPENSSL_DECLARE_GLOBAL(double,bar);
14300 #define bar OPENSSL_GLOBAL_REF(bar)
14302 The #defines are very important, and therefore so is including the
14303 header file everywhere where the defined globals are used.
14305 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14306 of ASN.1 items, but that structure is a bit different.
14308 The largest change is in util/mkdef.pl which has been enhanced with
14309 better and easier to understand logic to choose which symbols should
14310 go into the Windows .def files as well as a number of fixes and code
14311 cleanup (among others, algorithm keywords are now sorted
14312 lexicographically to avoid constant rewrites).
14316 * In BN_div() keep a copy of the sign of 'num' before writing the
14317 result to 'rm' because if rm==num the value will be overwritten
14318 and produce the wrong result if 'num' is negative: this caused
14319 problems with BN_mod() and BN_nnmod().
14323 * Function OCSP_request_verify(). This checks the signature on an
14324 OCSP request and verifies the signer certificate. The signer
14325 certificate is just checked for a generic purpose and OCSP request
14330 * Add OCSP_check_validity() function to check the validity of OCSP
14331 responses. OCSP responses are prepared in real time and may only
14332 be a few seconds old. Simply checking that the current time lies
14333 between thisUpdate and nextUpdate max reject otherwise valid responses
14334 caused by either OCSP responder or client clock inaccuracy. Instead
14335 we allow thisUpdate and nextUpdate to fall within a certain period of
14336 the current time. The age of the response can also optionally be
14337 checked. Two new options -validity_period and -status_age added to
14342 * If signature or public key algorithm is unrecognized print out its
14343 OID rather that just UNKNOWN.
14347 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14348 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14349 ID to be generated from the issuer certificate alone which can then be
14350 passed to OCSP_id_issuer_cmp().
14354 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14355 ASN1 modules to export functions returning ASN1_ITEM pointers
14356 instead of the ASN1_ITEM structures themselves. This adds several
14357 new macros which allow the underlying ASN1 function/structure to
14358 be accessed transparently. As a result code should not use ASN1_ITEM
14359 references directly (such as &X509_it) but instead use the relevant
14360 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14361 use of the new ASN1 code on platforms where exporting structures
14362 is problematical (for example in shared libraries) but exporting
14363 functions returning pointers to structures is not.
14367 * Add support for overriding the generation of SSL/TLS session IDs.
14368 These callbacks can be registered either in an SSL_CTX or per SSL.
14369 The purpose of this is to allow applications to control, if they wish,
14370 the arbitrary values chosen for use as session IDs, particularly as it
14371 can be useful for session caching in multiple-server environments. A
14372 command-line switch for testing this (and any client code that wishes
14373 to use such a feature) has been added to "s_server".
14375 *Geoff Thorpe, Lutz Jaenicke*
14377 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14378 of the form `#if defined(...) || defined(...) || ...` and
14379 `#if !defined(...) && !defined(...) && ...`. This also avoids
14380 the growing number of special cases it was previously handling.
14384 * Make all configuration macros available for application by making
14385 sure they are available in opensslconf.h, by giving them names starting
14386 with `OPENSSL_` to avoid conflicts with other packages and by making
14387 sure e_os2.h will cover all platform-specific cases together with
14389 Additionally, it is now possible to define configuration/platform-
14390 specific names (called "system identities"). In the C code, these
14391 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14392 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14393 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14398 * New option -set_serial to 'req' and 'x509' this allows the serial
14399 number to use to be specified on the command line. Previously self
14400 signed certificates were hard coded with serial number 0 and the
14401 CA options of 'x509' had to use a serial number in a file which was
14406 * New options to 'ca' utility to support V2 CRL entry extensions.
14407 Currently CRL reason, invalidity date and hold instruction are
14408 supported. Add new CRL extensions to V3 code and some new objects.
14412 * New function EVP_CIPHER_CTX_set_padding() this is used to
14413 disable standard block padding (aka PKCS#5 padding) in the EVP
14414 API, which was previously mandatory. This means that the data is
14415 not padded in any way and so the total length much be a multiple
14416 of the block size, otherwise an error occurs.
14420 * Initial (incomplete) OCSP SSL support.
14424 * New function OCSP_parse_url(). This splits up a URL into its host,
14425 port and path components: primarily to parse OCSP URLs. New -url
14426 option to ocsp utility.
14430 * New nonce behavior. The return value of OCSP_check_nonce() now
14431 reflects the various checks performed. Applications can decide
14432 whether to tolerate certain situations such as an absent nonce
14433 in a response when one was present in a request: the ocsp application
14434 just prints out a warning. New function OCSP_add1_basic_nonce()
14435 this is to allow responders to include a nonce in a response even if
14436 the request is nonce-less.
14440 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14441 skipped when using openssl x509 multiple times on a single input file,
14442 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14446 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14447 set string type: to handle setting ASN1_TIME structures. Fix ca
14448 utility to correctly initialize revocation date of CRLs.
14452 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14453 the clients preferred ciphersuites and rather use its own preferences.
14454 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14455 Internet Explorer by ensuring unchanged hash method during stepup.
14456 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14460 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14461 to aes and add a new 'exist' option to print out symbols that don't
14466 * Additional options to ocsp utility to allow flags to be set and
14467 additional certificates supplied.
14471 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14472 OCSP client a number of certificate to only verify the response
14477 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14478 handle the new API. Currently only ECB, CBC modes supported. Add new
14481 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14482 Encryption Standard (AES) Ciphersuites for Transport Layer
14483 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14484 not enabled by default and were not part of the "ALL" ciphersuite
14485 alias because they were not yet official; they could be
14486 explicitly requested by specifying the "AESdraft" ciphersuite
14487 group alias. In the final release of OpenSSL 0.9.7, the group
14488 alias is called "AES" and is part of "ALL".)
14490 *Ben Laurie, Steve Henson, Bodo Moeller*
14492 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14493 request to response.
14497 * Functions for OCSP responders. OCSP_request_onereq_count(),
14498 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14499 extract information from a certificate request. OCSP_response_create()
14500 creates a response and optionally adds a basic response structure.
14501 OCSP_basic_add1_status() adds a complete single response to a basic
14502 response and returns the OCSP_SINGLERESP structure just added (to allow
14503 extensions to be included for example). OCSP_basic_add1_cert() adds a
14504 certificate to a basic response and OCSP_basic_sign() signs a basic
14505 response with various flags. New helper functions ASN1_TIME_check()
14506 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14507 (converts ASN1_TIME to GeneralizedTime).
14511 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14512 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14513 structure from a certificate. X509_pubkey_digest() digests the public_key
14514 contents: this is used in various key identifiers.
14518 * Make sk_sort() tolerate a NULL argument.
14520 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14522 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14523 passed by the function are trusted implicitly. If any of them signed the
14524 response then it is assumed to be valid and is not verified.
14528 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14529 to data. This was previously part of the PKCS7 ASN1 code. This
14530 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14531 *Steve Henson, reported by Kenneth R. Robinette
14532 <support@securenetterm.com>*
14534 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14535 routines: without these tracing memory leaks is very painful.
14536 Fix leaks in PKCS12 and PKCS7 routines.
14540 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14541 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14542 effectively meant GeneralizedTime would never be used. Now it
14543 is initialised to -1 but X509_time_adj() now has to check the value
14544 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14545 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14546 *Steve Henson, reported by Kenneth R. Robinette
14547 <support@securenetterm.com>*
14549 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14550 result in a zero length in the ASN1_INTEGER structure which was
14551 not consistent with the structure when d2i_ASN1_INTEGER() was used
14552 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14553 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14554 where it did not print out a minus for negative ASN1_INTEGER.
14558 * Add summary printout to ocsp utility. The various functions which
14559 convert status values to strings have been renamed to:
14560 OCSP_response_status_str(), OCSP_cert_status_str() and
14561 OCSP_crl_reason_str() and are no longer static. New options
14562 to verify nonce values and to disable verification. OCSP response
14563 printout format cleaned up.
14567 * Add additional OCSP certificate checks. These are those specified
14568 in RFC2560. This consists of two separate checks: the CA of the
14569 certificate being checked must either be the OCSP signer certificate
14570 or the issuer of the OCSP signer certificate. In the latter case the
14571 OCSP signer certificate must contain the OCSP signing extended key
14572 usage. This check is performed by attempting to match the OCSP
14573 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14574 in the OCSP_CERTID structures of the response.
14578 * Initial OCSP certificate verification added to OCSP_basic_verify()
14579 and related routines. This uses the standard OpenSSL certificate
14580 verify routines to perform initial checks (just CA validity) and
14581 to obtain the certificate chain. Then additional checks will be
14582 performed on the chain. Currently the root CA is checked to see
14583 if it is explicitly trusted for OCSP signing. This is used to set
14584 a root CA as a global signing root: that is any certificate that
14585 chains to that CA is an acceptable OCSP signing certificate.
14589 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14590 extensions from a separate configuration file.
14591 As when reading extensions from the main configuration file,
14592 the '-extensions ...' option may be used for specifying the
14595 *Massimiliano Pala <madwolf@comune.modena.it>*
14597 * New OCSP utility. Allows OCSP requests to be generated or
14598 read. The request can be sent to a responder and the output
14599 parsed, outputted or printed in text form. Not complete yet:
14600 still needs to check the OCSP response validity.
14604 * New subcommands for 'openssl ca':
14605 `openssl ca -status <serial>` prints the status of the cert with
14606 the given serial number (according to the index file).
14607 `openssl ca -updatedb` updates the expiry status of certificates
14610 *Massimiliano Pala <madwolf@comune.modena.it>*
14612 * New '-newreq-nodes' command option to CA.pl. This is like
14613 '-newreq', but calls 'openssl req' with the '-nodes' option
14614 so that the resulting key is not encrypted.
14616 *Damien Miller <djm@mindrot.org>*
14618 * New configuration for the GNU Hurd.
14620 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14622 * Initial code to implement OCSP basic response verify. This
14623 is currently incomplete. Currently just finds the signer's
14624 certificate and verifies the signature on the response.
14628 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14629 value of OPENSSLDIR. This is available via the new '-d' option
14630 to 'openssl version', and is also included in 'openssl version -a'.
14634 * Allowing defining memory allocation callbacks that will be given
14635 file name and line number information in additional arguments
14636 (a `const char*` and an int). The basic functionality remains, as
14637 well as the original possibility to just replace malloc(),
14638 realloc() and free() by functions that do not know about these
14639 additional arguments. To register and find out the current
14640 settings for extended allocation functions, the following
14641 functions are provided:
14643 CRYPTO_set_mem_ex_functions
14644 CRYPTO_set_locked_mem_ex_functions
14645 CRYPTO_get_mem_ex_functions
14646 CRYPTO_get_locked_mem_ex_functions
14648 These work the same way as CRYPTO_set_mem_functions and friends.
14649 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14650 extended allocation function is enabled.
14651 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14652 a conventional allocation function is enabled.
14654 *Richard Levitte, Bodo Moeller*
14656 * Finish off removing the remaining LHASH function pointer casts.
14657 There should no longer be any prototype-casting required when using
14658 the LHASH abstraction, and any casts that remain are "bugs". See
14659 the callback types and macros at the head of lhash.h for details
14660 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14664 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14665 If /dev/[u]random devices are not available or do not return enough
14666 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14668 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14669 /etc/entropy will be queried once each in this sequence, querying stops
14670 when enough entropy was collected without querying more sockets.
14674 * Change the Unix RAND_poll() variant to be able to poll several
14675 random devices, as specified by DEVRANDOM, until a sufficient amount
14676 of data has been collected. We spend at most 10 ms on each file
14677 (select timeout) and read in non-blocking mode. DEVRANDOM now
14678 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14679 (previously it was just the string "/dev/urandom"), so on typical
14680 platforms the 10 ms delay will never occur.
14681 Also separate out the Unix variant to its own file, rand_unix.c.
14682 For VMS, there's a currently-empty rand_vms.c.
14686 * Move OCSP client related routines to ocsp_cl.c. These
14687 provide utility functions which an application needing
14688 to issue a request to an OCSP responder and analyse the
14689 response will typically need: as opposed to those which an
14690 OCSP responder itself would need which will be added later.
14692 OCSP_request_sign() signs an OCSP request with an API similar
14693 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14694 response. OCSP_response_get1_basic() extracts basic response
14695 from response. OCSP_resp_find_status(): finds and extracts status
14696 information from an OCSP_CERTID structure (which will be created
14697 when the request structure is built). These are built from lower
14698 level functions which work on OCSP_SINGLERESP structures but
14699 won't normally be used unless the application wishes to examine
14700 extensions in the OCSP response for example.
14702 Replace nonce routines with a pair of functions.
14703 OCSP_request_add1_nonce() adds a nonce value and optionally
14704 generates a random value. OCSP_check_nonce() checks the
14705 validity of the nonce in an OCSP response.
14709 * Change function OCSP_request_add() to OCSP_request_add0_id().
14710 This doesn't copy the supplied OCSP_CERTID and avoids the
14711 need to free up the newly created id. Change return type
14712 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14713 This can then be used to add extensions to the request.
14714 Deleted OCSP_request_new(), since most of its functionality
14715 is now in OCSP_REQUEST_new() (and the case insensitive name
14716 clash) apart from the ability to set the request name which
14717 will be added elsewhere.
14721 * Update OCSP API. Remove obsolete extensions argument from
14722 various functions. Extensions are now handled using the new
14723 OCSP extension code. New simple OCSP HTTP function which
14724 can be used to send requests and parse the response.
14728 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14729 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14730 uses the special reorder version of SET OF to sort the attributes
14731 and reorder them to match the encoded order. This resolves a long
14732 standing problem: a verify on a PKCS7 structure just after signing
14733 it used to fail because the attribute order did not match the
14734 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14735 it uses the received order. This is necessary to tolerate some broken
14736 software that does not order SET OF. This is handled by encoding
14737 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14738 to produce the required SET OF.
14742 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14743 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14744 files to get correct declarations of the ASN.1 item variables.
14748 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14749 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14750 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14751 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14752 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14753 ASN1_ITEM and no wrapper functions.
14757 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14758 replace the old function pointer based I/O routines. Change most of
14759 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14763 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14764 lines, recognize more "algorithms" that can be deselected, and make
14765 it complain about algorithm deselection that isn't recognised.
14769 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14770 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14771 to use new functions. Add NO_ASN1_OLD which can be set to remove
14772 some old style ASN1 functions: this can be used to determine if old
14773 code will still work when these eventually go away.
14777 * New extension functions for OCSP structures, these follow the
14778 same conventions as certificates and CRLs.
14782 * New function X509V3_add1_i2d(). This automatically encodes and
14783 adds an extension. Its behaviour can be customised with various
14784 flags to append, replace or delete. Various wrappers added for
14785 certificates and CRLs.
14789 * Fix to avoid calling the underlying ASN1 print routine when
14790 an extension cannot be parsed. Correct a typo in the
14791 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14795 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14796 entries for variables.
14800 * Add functionality to `apps/openssl.c` for detecting locking
14801 problems: As the program is single-threaded, all we have
14802 to do is register a locking callback using an array for
14803 storing which locks are currently held by the program.
14807 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14808 SSL_get_ex_data_X509_STORE_idx(), which is used in
14809 ssl_verify_cert_chain() and thus can be called at any time
14810 during TLS/SSL handshakes so that thread-safety is essential.
14811 Unfortunately, the ex_data design is not at all suited
14812 for multi-threaded use, so it probably should be abolished.
14816 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14818 *Broadcom, tweaked and integrated by Geoff Thorpe*
14820 * Move common extension printing code to new function
14821 X509V3_print_extensions(). Reorganise OCSP print routines and
14822 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14826 * New function X509_signature_print() to remove duplication in some
14831 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14832 set (this was treated exactly the same as SET OF previously). This
14833 is used to reorder the STACK representing the structure to match the
14834 encoding. This will be used to get round a problem where a PKCS7
14835 structure which was signed could not be verified because the STACK
14836 order did not reflect the encoded order.
14840 * Reimplement the OCSP ASN1 module using the new code.
14844 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14845 for its ASN1 operations. The old style function pointers still exist
14846 for now but they will eventually go away.
14850 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14851 completely replaces the old ASN1 functionality with a table driven
14852 encoder and decoder which interprets an ASN1_ITEM structure describing
14853 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14854 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14855 has also been converted to the new form.
14859 * Change BN_mod_exp_recp so that negative moduli are tolerated
14860 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14861 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14862 for negative moduli.
14866 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14867 of not touching the result's sign bit.
14871 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14876 * Changed the LHASH code to use prototypes for callbacks, and created
14877 macros to declare and implement thin (optionally static) functions
14878 that provide type-safety and avoid function pointer casting for the
14879 type-specific callbacks.
14883 * Added Kerberos Cipher Suites to be used with TLS, as written in
14885 *Veers Staats <staatsvr@asc.hpc.mil>,
14886 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14888 * Reformat the FAQ so the different questions and answers can be divided
14889 in sections depending on the subject.
14893 * Have the zlib compression code load ZLIB.DLL dynamically under
14898 * New function BN_mod_sqrt for computing square roots modulo a prime
14899 (using the probabilistic Tonelli-Shanks algorithm unless
14900 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14901 be handled deterministically).
14903 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14905 * Make BN_mod_inverse faster by explicitly handling small quotients
14906 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14907 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14911 * New function BN_kronecker.
14915 * Fix BN_gcd so that it works on negative inputs; the result is
14916 positive unless both parameters are zero.
14917 Previously something reasonably close to an infinite loop was
14918 possible because numbers could be growing instead of shrinking
14919 in the implementation of Euclid's algorithm.
14923 * Fix BN_is_word() and BN_is_one() macros to take into account the
14924 sign of the number in question.
14926 Fix BN_is_word(a,w) to work correctly for w == 0.
14928 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14929 because its test if the absolute value of 'a' equals 'w'.
14930 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14931 it exists mostly for use in the implementations of BN_is_zero(),
14932 BN_is_one(), and BN_is_word().
14936 * New function BN_swap.
14940 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14941 the exponentiation functions are more likely to produce reasonable
14942 results on negative inputs.
14946 * Change BN_mod_mul so that the result is always non-negative.
14947 Previously, it could be negative if one of the factors was negative;
14948 I don't think anyone really wanted that behaviour.
14952 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14953 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14954 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14955 and add new functions:
14964 BN_mod_lshift1_quick
14966 BN_mod_lshift_quick
14968 These functions always generate non-negative results.
14970 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14971 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14973 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14974 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14975 be reduced modulo `m`.
14977 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14980 The following entry accidentally appeared in the CHANGES file
14981 distributed with OpenSSL 0.9.7. The modifications described in
14982 it do *not* apply to OpenSSL 0.9.7.
14984 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14985 was actually never needed) and in BN_mul(). The removal in BN_mul()
14986 required a small change in bn_mul_part_recursive() and the addition
14987 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14988 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14989 bn_sub_words() and bn_add_words() except they take arrays with
14995 * In 'openssl passwd', verify passwords read from the terminal
14996 unless the '-salt' option is used (which usually means that
14997 verification would just waste user's time since the resulting
14998 hash is going to be compared with some given password hash)
14999 or the new '-noverify' option is used.
15001 This is an incompatible change, but it does not affect
15002 non-interactive use of 'openssl passwd' (passwords on the command
15003 line, '-stdin' option, '-in ...' option) and thus should not
15004 cause any problems.
15008 * Remove all references to RSAref, since there's no more need for it.
15012 * Make DSO load along a path given through an environment variable
15013 (SHLIB_PATH) with shl_load().
15017 * Constify the ENGINE code as a result of BIGNUM constification.
15018 Also constify the RSA code and most things related to it. In a
15019 few places, most notable in the depth of the ASN.1 code, ugly
15020 casts back to non-const were required (to be solved at a later
15025 * Make it so the openssl application has all engines loaded by default.
15029 * Constify the BIGNUM routines a little more.
15033 * Add the following functions:
15035 ENGINE_load_cswift()
15037 ENGINE_load_atalla()
15038 ENGINE_load_nuron()
15039 ENGINE_load_builtin_engines()
15041 That way, an application can itself choose if external engines that
15042 are built-in in OpenSSL shall ever be used or not. The benefit is
15043 that applications won't have to be linked with libdl or other dso
15044 libraries unless it's really needed.
15046 Changed 'openssl engine' to load all engines on demand.
15047 Changed the engine header files to avoid the duplication of some
15048 declarations (they differed!).
15052 * 'openssl engine' can now list capabilities.
15056 * Better error reporting in 'openssl engine'.
15060 * Never call load_dh_param(NULL) in s_server.
15064 * Add engine application. It can currently list engines by name and
15065 identity, and test if they are actually available.
15069 * Improve RPM specification file by forcing symbolic linking and making
15070 sure the installed documentation is also owned by root.root.
15072 *Damien Miller <djm@mindrot.org>*
15074 * Give the OpenSSL applications more possibilities to make use of
15075 keys (public as well as private) handled by engines.
15079 * Add OCSP code that comes from CertCo.
15083 * Add VMS support for the Rijndael code.
15087 * Added untested support for Nuron crypto accelerator.
15091 * Add support for external cryptographic devices. This code was
15092 previously distributed separately as the "engine" branch.
15094 *Geoff Thorpe, Richard Levitte*
15096 * Rework the filename-translation in the DSO code. It is now possible to
15097 have far greater control over how a "name" is turned into a filename
15098 depending on the operating environment and any oddities about the
15099 different shared library filenames on each system.
15103 * Support threads on FreeBSD-elf in Configure.
15107 * Fix for SHA1 assembly problem with MASM: it produces
15108 warnings about corrupt line number information when assembling
15109 with debugging information. This is caused by the overlapping
15112 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15115 NCONF_get_number() has no error checking at all. As a replacement,
15116 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15117 promoted strongly. The old NCONF_get_number is kept around for
15118 binary backward compatibility.
15119 Make it possible for methods to load from something other than a BIO,
15120 by providing a function pointer that is given a name instead of a BIO.
15121 For example, this could be used to load configuration data from an
15126 * Fix for non blocking accept BIOs. Added new I/O special reason
15127 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15128 with non blocking I/O was not possible because no retry code was
15129 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15134 * Added the beginnings of Rijndael support.
15138 * Fix for bug in DirectoryString mask setting. Add support for
15139 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15140 to allow certificate printing to more controllable, additional
15141 'certopt' option to 'x509' to allow new printing options to be
15146 * Clean old EAY MD5 hack from e_os.h.
15150 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15152 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15153 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15155 *Joe Orton, Steve Henson*
15157 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15159 * Fix additional bug revealed by the NISCC test suite:
15161 Stop bug triggering large recursion when presented with
15162 certain ASN.1 tags ([CVE-2003-0851])
15166 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15168 * Fix various bugs revealed by running the NISCC test suite:
15170 Stop out of bounds reads in the ASN1 code when presented with
15171 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15173 If verify callback ignores invalid public key errors don't try to check
15174 certificate signature with the NULL public key.
15178 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15179 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15184 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15185 extra data after the compression methods not only for TLS 1.0
15186 but also for SSL 3.0 (as required by the specification).
15188 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15190 * Change X509_certificate_type() to mark the key as exported/exportable
15191 when it's 512 *bits* long, not 512 bytes.
15195 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15197 * Countermeasure against the Klima-Pokorny-Rosa extension of
15198 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15199 a protocol version number mismatch like a decryption error
15200 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15204 * Turn on RSA blinding by default in the default implementation
15205 to avoid a timing attack. Applications that don't want it can call
15206 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15207 They would be ill-advised to do so in most cases.
15209 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15211 * Change RSA blinding code so that it works when the PRNG is not
15212 seeded (in this case, the secret RSA exponent is abused as
15213 an unpredictable seed -- if it is not unpredictable, there
15214 is no point in blinding anyway). Make RSA blinding thread-safe
15215 by remembering the creator's thread ID in rsa->blinding and
15216 having all other threads use local one-time blinding factors
15217 (this requires more computation than sharing rsa->blinding, but
15218 avoids excessive locking; and if an RSA object is not shared
15219 between threads, blinding will still be very fast).
15223 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15225 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15226 via timing by performing a MAC computation even if incorrect
15227 block cipher padding has been found. This is a countermeasure
15228 against active attacks where the attacker has to distinguish
15229 between bad padding and a MAC verification error. ([CVE-2003-0078])
15231 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15232 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15233 Martin Vuagnoux (EPFL, Ilion)*
15235 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15237 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15238 memory from its contents. This is done with a counter that will
15239 place alternating values in each byte. This can be used to solve
15240 two issues: 1) the removal of calls to memset() by highly optimizing
15241 compilers, and 2) cleansing with other values than 0, since those can
15242 be read through on certain media, for example a swap space on disk.
15246 * Bugfix: client side session caching did not work with external caching,
15247 because the session->cipher setting was not restored when reloading
15248 from the external cache. This problem was masked, when
15249 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15250 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15254 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15255 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15257 *Zeev Lieber <zeev-l@yahoo.com>*
15259 * Undo an undocumented change introduced in 0.9.6e which caused
15260 repeated calls to OpenSSL_add_all_ciphers() and
15261 OpenSSL_add_all_digests() to be ignored, even after calling
15266 * Change the default configuration reader to deal with last line not
15267 being properly terminated.
15271 * Change X509_NAME_cmp() so it applies the special rules on handling
15272 DN values that are of type PrintableString, as well as RDNs of type
15273 emailAddress where the value has the type ia5String.
15275 *stefank@valicert.com via Richard Levitte*
15277 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15278 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15279 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15280 the bitwise-OR of the two for use by the majority of applications
15281 wanting this behaviour, and update the docs. The documented
15282 behaviour and actual behaviour were inconsistent and had been
15283 changing anyway, so this is more a bug-fix than a behavioural
15286 *Geoff Thorpe, diagnosed by Nadav Har'El*
15288 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15289 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15293 * Fix initialization code race conditions in
15294 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15295 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15296 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15297 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15298 ssl2_get_cipher_by_char(),
15299 ssl3_get_cipher_by_char().
15301 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15303 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15304 the cached sessions are flushed, as the remove_cb() might use ex_data
15305 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15306 (see [openssl.org #212]).
15308 *Geoff Thorpe, Lutz Jaenicke*
15310 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15311 length, instead of the encoding length to d2i_ASN1_OBJECT.
15315 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15317 * [In 0.9.6g-engine release:]
15318 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15320 *Lynn Gazis <lgazis@rainbow.com>*
15322 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15324 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15325 and get fix the header length calculation.
15326 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15327 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15329 * Use proper error handling instead of 'assertions' in buffer
15330 overflow checks added in 0.9.6e. This prevents DoS (the
15331 assertions could call abort()).
15333 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15335 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15337 * Add various sanity checks to asn1_get_length() to reject
15338 the ASN1 length bytes if they exceed sizeof(long), will appear
15339 negative or the content length exceeds the length of the
15342 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15344 * Fix cipher selection routines: ciphers without encryption had no flags
15345 for the cipher strength set and where therefore not handled correctly
15346 by the selection routines (PR #130).
15350 * Fix EVP_dsa_sha macro.
15355 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15356 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15357 that was added in OpenSSL 0.9.6d.
15359 As the countermeasure turned out to be incompatible with some
15360 broken SSL implementations, the new option is part of SSL_OP_ALL.
15361 SSL_OP_ALL is usually employed when compatibility with weird SSL
15362 implementations is desired (e.g. '-bugs' option to 's_client' and
15363 's_server'), so the new option is automatically set in many
15368 * Changes in security patch:
15370 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15371 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15372 Air Force Materiel Command, USAF, under agreement number
15375 * Add various sanity checks to asn1_get_length() to reject
15376 the ASN1 length bytes if they exceed sizeof(long), will appear
15377 negative or the content length exceeds the length of the
15378 supplied buffer. ([CVE-2002-0659])
15380 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15382 * Assertions for various potential buffer overflows, not known to
15383 happen in practice.
15385 *Ben Laurie (CHATS)*
15387 * Various temporary buffers to hold ASCII versions of integers were
15388 too small for 64 bit platforms. ([CVE-2002-0655])
15389 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15391 * Remote buffer overflow in SSL3 protocol - an attacker could
15392 supply an oversized session ID to a client. ([CVE-2002-0656])
15394 *Ben Laurie (CHATS)*
15396 * Remote buffer overflow in SSL2 protocol - an attacker could
15397 supply an oversized client master key. ([CVE-2002-0656])
15399 *Ben Laurie (CHATS)*
15401 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15403 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15404 encoded as NULL) with id-dsa-with-sha1.
15406 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15408 * Check various `X509_...()` return values in `apps/req.c`.
15410 *Nils Larsch <nla@trustcenter.de>*
15412 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15413 an end-of-file condition would erroneously be flagged, when the CRLF
15414 was just at the end of a processed block. The bug was discovered when
15415 processing data through a buffering memory BIO handing the data to a
15416 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15417 <ptsekov@syntrex.com> and Nedelcho Stanev.
15421 * Implement a countermeasure against a vulnerability recently found
15422 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15423 before application data chunks to avoid the use of known IVs
15424 with data potentially chosen by the attacker.
15428 * Fix length checks in ssl3_get_client_hello().
15432 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15433 to prevent ssl3_read_internal() from incorrectly assuming that
15434 ssl3_read_bytes() found application data while handshake
15435 processing was enabled when in fact s->s3->in_read_app_data was
15436 merely automatically cleared during the initial handshake.
15438 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15440 * Fix object definitions for Private and Enterprise: they were not
15441 recognized in their shortname (=lowercase) representation. Extend
15442 obj_dat.pl to issue an error when using undefined keywords instead
15443 of silently ignoring the problem (Svenning Sorensen
15444 <sss@sss.dnsalias.net>).
15448 * Fix DH_generate_parameters() so that it works for 'non-standard'
15449 generators, i.e. generators other than 2 and 5. (Previously, the
15450 code did not properly initialise the 'add' and 'rem' values to
15451 BN_generate_prime().)
15453 In the new general case, we do not insist that 'generator' is
15454 actually a primitive root: This requirement is rather pointless;
15455 a generator of the order-q subgroup is just as good, if not
15460 * Map new X509 verification errors to alerts. Discovered and submitted by
15461 Tom Wu <tom@arcot.com>.
15465 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15466 returning non-zero before the data has been completely received
15467 when using non-blocking I/O.
15469 *Bodo Moeller; problem pointed out by John Hughes*
15471 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15473 *Ben Laurie, Lutz Jaenicke*
15475 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15476 Yoram Zahavi <YoramZ@gilian.com>).
15480 * Add information about CygWin 1.3 and on, and preserve proper
15481 configuration for the versions before that.
15483 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15485 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15486 check whether we deal with a copy of a session and do not delete from
15487 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15488 <izhar@checkpoint.com>.
15492 * Do not store session data into the internal session cache, if it
15493 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15494 flag is set). Proposed by Aslam <aslam@funk.com>.
15498 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15503 * [In 0.9.6d-engine release:]
15504 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15506 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15508 * Add the configuration target linux-s390x.
15510 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15512 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15513 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15514 variable as an indication that a ClientHello message has been
15515 received. As the flag value will be lost between multiple
15516 invocations of ssl3_accept when using non-blocking I/O, the
15517 function may not be aware that a handshake has actually taken
15518 place, thus preventing a new session from being added to the
15521 To avoid this problem, we now set s->new_session to 2 instead of
15522 using a local variable.
15524 *Lutz Jaenicke, Bodo Moeller*
15526 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15527 if the SSL_R_LENGTH_MISMATCH error is detected.
15529 *Geoff Thorpe, Bodo Moeller*
15531 * New 'shared_ldflag' column in Configure platform table.
15535 * Fix EVP_CIPHER_mode macro.
15537 *"Dan S. Camper" <dan@bti.net>*
15539 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15540 type, we must throw them away by setting rr->length to 0.
15542 *D P Chang <dpc@qualys.com>*
15544 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15546 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15547 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15548 worked incorrectly for those cases where range = `10..._2` and
15549 `3*range` is two bits longer than range.)
15553 * Only add signing time to PKCS7 structures if it is not already
15558 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15559 OBJ_ld_ce should be OBJ_id_ce.
15560 Also some ip-pda OIDs in crypto/objects/objects.txt were
15561 incorrect (cf. RFC 3039).
15563 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15565 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15566 returns early because it has nothing to do.
15568 *Andy Schneider <andy.schneider@bjss.co.uk>*
15570 * [In 0.9.6c-engine release:]
15571 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15573 *Andy Schneider <andy.schneider@bjss.co.uk>*
15575 * [In 0.9.6c-engine release:]
15576 Add support for Cryptographic Appliance's keyserver technology.
15577 (Use engine 'keyclient')
15579 *Cryptographic Appliances and Geoff Thorpe*
15581 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15582 is called via tools/c89.sh because arguments have to be
15583 rearranged (all '-L' options must appear before the first object
15586 *Richard Shapiro <rshapiro@abinitio.com>*
15588 * [In 0.9.6c-engine release:]
15589 Add support for Broadcom crypto accelerator cards, backported
15592 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15594 * [In 0.9.6c-engine release:]
15595 Add support for SureWare crypto accelerator cards from
15596 Baltimore Technologies. (Use engine 'sureware')
15598 *Baltimore Technologies and Mark Cox*
15600 * [In 0.9.6c-engine release:]
15601 Add support for crypto accelerator cards from Accelerated
15602 Encryption Processing, www.aep.ie. (Use engine 'aep')
15604 *AEP Inc. and Mark Cox*
15606 * Add a configuration entry for gcc on UnixWare.
15608 *Gary Benson <gbenson@redhat.com>*
15610 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15611 messages are stored in a single piece (fixed-length part and
15612 variable-length part combined) and fix various bugs found on the way.
15616 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15617 instead. BIO_gethostbyname() does not know what timeouts are
15618 appropriate, so entries would stay in cache even when they have
15620 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15622 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15623 faced with a pathologically small ClientHello fragment that does
15624 not contain client_version: Instead of aborting with an error,
15625 simply choose the highest available protocol version (i.e.,
15626 TLS 1.0 unless it is disabled). In practice, ClientHello
15627 messages are never sent like this, but this change gives us
15628 strictly correct behaviour at least for TLS.
15632 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15633 never resets s->method to s->ctx->method when called from within
15634 one of the SSL handshake functions.
15636 *Bodo Moeller; problem pointed out by Niko Baric*
15638 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15639 (sent using the client's version number) if client_version is
15640 smaller than the protocol version in use. Also change
15641 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15642 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15643 the client will at least see that alert.
15647 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15652 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15653 client receives HelloRequest while in a handshake.
15655 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15657 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15658 should end in 'break', not 'goto end' which circumvents various
15659 cleanups done in state SSL_ST_OK. But session related stuff
15660 must be disabled for SSL_ST_OK in the case that we just sent a
15663 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15664 before just sending a HelloRequest.
15666 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15668 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15669 reveal whether illegal block cipher padding was found or a MAC
15670 verification error occurred. (Neither SSLerr() codes nor alerts
15671 are directly visible to potential attackers, but the information
15672 may leak via logfiles.)
15674 Similar changes are not required for the SSL 2.0 implementation
15675 because the number of padding bytes is sent in clear for SSL 2.0,
15676 and the extra bytes are just ignored. However ssl/s2_pkt.c
15677 failed to verify that the purported number of padding bytes is in
15682 * Add OpenUNIX-8 support including shared libraries
15683 (Boyd Lynn Gerber <gerberb@zenez.com>).
15687 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15688 'wristwatch attack' using huge encoding parameters (cf.
15689 James H. Manger's CRYPTO 2001 paper). Note that the
15690 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15691 encoding parameters and hence was not vulnerable.
15695 * BN_sqr() bug fix.
15697 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15699 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15700 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15701 followed by modular reduction.
15703 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15705 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15706 equivalent based on BN_pseudo_rand() instead of BN_rand().
15710 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15711 This function was broken, as the check for a new client hello message
15712 to handle SGC did not allow these large messages.
15713 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15717 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15721 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15722 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15726 * Rework the configuration and shared library support for Tru64 Unix.
15727 The configuration part makes use of modern compiler features and
15728 still retains old compiler behavior for those that run older versions
15729 of the OS. The shared library support part includes a variant that
15730 uses the RPATH feature, and is available through the special
15731 configuration target "alpha-cc-rpath", which will never be selected
15734 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15736 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15737 with the same message size as in ssl3_get_certificate_request().
15738 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15739 messages might inadvertently be reject as too long.
15741 *Petr Lampa <lampa@fee.vutbr.cz>*
15743 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15747 * Modified SSL library such that the verify_callback that has been set
15748 specifically for an SSL object with SSL_set_verify() is actually being
15749 used. Before the change, a verify_callback set with this function was
15750 ignored and the verify_callback() set in the SSL_CTX at the time of
15751 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15752 to allow the necessary settings.
15756 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15757 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15758 done automatically (in contradiction to the requirements of the C
15759 standard). This made problems when used from OpenSSH.
15763 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15764 dh->length and always used
15766 BN_rand_range(priv_key, dh->p).
15768 BN_rand_range() is not necessary for Diffie-Hellman, and this
15769 specific range makes Diffie-Hellman unnecessarily inefficient if
15770 dh->length (recommended exponent length) is much smaller than the
15771 length of dh->p. We could use BN_rand_range() if the order of
15772 the subgroup was stored in the DH structure, but we only have
15777 BN_rand(priv_key, l, ...)
15779 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15786 RSA_eay_public_encrypt
15787 RSA_eay_private_decrypt
15788 RSA_eay_private_encrypt (signing)
15789 RSA_eay_public_decrypt (signature verification)
15791 (default implementations for RSA_public_encrypt,
15792 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15793 always reject numbers >= n.
15797 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15798 to synchronize access to 'locking_thread'. This is necessary on
15799 systems where access to 'locking_thread' (an 'unsigned long'
15800 variable) is not atomic.
15804 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15805 *before* setting the 'crypto_lock_rand' flag. The previous code had
15806 a race condition if 0 is a valid thread ID.
15808 *Travis Vitek <vitek@roguewave.com>*
15810 * Add support for shared libraries under Irix.
15812 *Albert Chin-A-Young <china@thewrittenword.com>*
15814 * Add configuration option to build on Linux on both big-endian and
15815 little-endian MIPS.
15817 *Ralf Baechle <ralf@uni-koblenz.de>*
15819 * Add the possibility to create shared libraries on HP-UX.
15823 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15825 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15826 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15827 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15828 PRNG state recovery was possible based on the output of
15829 one PRNG request appropriately sized to gain knowledge on
15830 'md' followed by enough consecutive 1-byte PRNG requests
15831 to traverse all of 'state'.
15833 1. When updating 'md_local' (the current thread's copy of 'md')
15834 during PRNG output generation, hash all of the previous
15835 'md_local' value, not just the half used for PRNG output.
15837 2. Make the number of bytes from 'state' included into the hash
15838 independent from the number of PRNG bytes requested.
15840 The first measure alone would be sufficient to avoid
15841 Markku-Juhani's attack. (Actually it had never occurred
15842 to me that the half of 'md_local' used for chaining was the
15843 half from which PRNG output bytes were taken -- I had always
15844 assumed that the secret half would be used.) The second
15845 measure makes sure that additional data from 'state' is never
15846 mixed into 'md_local' in small portions; this heuristically
15847 further strengthens the PRNG.
15851 * Fix crypto/bn/asm/mips3.s.
15855 * When only the key is given to "enc", the IV is undefined. Print out
15856 an error message in this case.
15860 * Handle special case when X509_NAME is empty in X509 printing routines.
15864 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15865 positive and less than q.
15869 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15870 used: it isn't thread safe and the add_lock_callback should handle
15873 *Paul Rose <Paul.Rose@bridge.com>*
15875 * Verify that incoming data obeys the block size in
15876 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15882 *Ulf Möller, Bodo Möller*
15884 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15885 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15886 when fixing the server behaviour for backwards-compatible 'client
15887 hello' messages. (Note that the attack is impractical against
15888 SSL 3.0 and TLS 1.0 anyway because length and version checking
15889 means that the probability of guessing a valid ciphertext is
15890 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15893 Before 0.9.5, the countermeasure (hide the error by generating a
15894 random 'decryption result') did not work properly because
15895 ERR_clear_error() was missing, meaning that SSL_get_error() would
15896 detect the supposedly ignored error.
15898 Both problems are now fixed.
15902 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15903 (previously it was 1024).
15907 * Fix for compatibility mode trust settings: ignore trust settings
15908 unless some valid trust or reject settings are present.
15912 * Fix for blowfish EVP: its a variable length cipher.
15916 * Fix various bugs related to DSA S/MIME verification. Handle missing
15917 parameters in DSA public key structures and return an error in the
15918 DSA routines if parameters are absent.
15922 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15923 in the current directory if neither $RANDFILE nor $HOME was set.
15924 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15925 caused some confusion to Windows users who haven't defined $HOME.
15926 Thus RAND_file_name() is changed again: e_os.h can define a
15927 DEFAULT_HOME, which will be used if $HOME is not set.
15928 For Windows, we use "C:"; on other platforms, we still require
15929 environment variables.
15931 * Move 'if (!initialized) RAND_poll()' into regions protected by
15932 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15933 having multiple threads call RAND_poll() concurrently.
15937 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15938 combination of a flag and a thread ID variable.
15939 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15940 flag), *other* threads can enter ssleay_add_bytes without obeying
15941 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15942 that they do not hold after the first thread unsets add_do_not_lock).
15946 * Change bctest again: '-x' expressions are not available in all
15947 versions of 'test'.
15951 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15953 * Fix a couple of memory leaks in PKCS7_dataDecode()
15955 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15957 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15958 the default extension for executables, if any. Also, make the perl
15959 scripts that use symlink() to test if it really exists and use "cp"
15960 if it doesn't. All this made OpenSSL compilable and installable in
15965 * Fix for asn1_GetSequence() for indefinite length constructed data.
15966 If SEQUENCE is length is indefinite just set c->slen to the total
15967 amount of data available.
15969 *Steve Henson, reported by shige@FreeBSD.org*
15971 *This change does not apply to 0.9.7.*
15973 * Change bctest to avoid here-documents inside command substitution
15974 (workaround for FreeBSD /bin/sh bug).
15975 For compatibility with Ultrix, avoid shell functions (introduced
15976 in the bctest version that searches along $PATH).
15980 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15981 with des_encrypt() defined on some operating systems, like Solaris
15986 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15987 On the Importance of Eliminating Errors in Cryptographic
15988 Computations, J. Cryptology 14 (2001) 2, 101-119,
15989 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15993 * MIPS assembler BIGNUM division bug fix.
15997 * Disabled incorrect Alpha assembler code.
16001 * Fix PKCS#7 decode routines so they correctly update the length
16002 after reading an EOC for the EXPLICIT tag.
16006 *This change does not apply to 0.9.7.*
16008 * Fix bug in PKCS#12 key generation routines. This was triggered
16009 if a 3DES key was generated with a 0 initial byte. Include
16010 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16011 (but broken) behaviour.
16015 * Enhance bctest to search for a working bc along $PATH and print
16018 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16020 * Fix memory leaks in err.c: free err_data string if necessary;
16021 don't write to the wrong index in ERR_set_error_data.
16025 * Implement ssl23_peek (analogous to ssl23_read), which previously
16030 * Replace rdtsc with `_emit` statements for VC++ version 5.
16032 *Jeremy Cooper <jeremy@baymoo.org>*
16034 * Make it possible to reuse SSLv2 sessions.
16038 * In copy_email() check for >= 0 as a return value for
16039 X509_NAME_get_index_by_NID() since 0 is a valid index.
16041 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16043 * Avoid coredump with unsupported or invalid public keys by checking if
16044 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16045 PKCS7_verify() fails with non detached data.
16049 * Don't use getenv in library functions when run as setuid/setgid.
16050 New function OPENSSL_issetugid().
16054 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16055 due to incorrect handling of multi-threading:
16057 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16059 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16061 3. Count how many times MemCheck_off() has been called so that
16062 nested use can be treated correctly. This also avoids
16063 inband-signalling in the previous code (which relied on the
16064 assumption that thread ID 0 is impossible).
16068 * Add "-rand" option also to s_client and s_server.
16072 * Fix CPU detection on Irix 6.x.
16073 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16074 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16076 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16081 *This change does not apply to 0.9.7.*
16083 * Use the cached encoding of an X509_NAME structure rather than
16084 copying it. This is apparently the reason for the libsafe "errors"
16085 but the code is actually correct.
16089 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16090 Bleichenbacher's DSA attack.
16091 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16092 to be set and top=0 forces the highest bit to be set; top=-1 is new
16093 and leaves the highest bit random.
16095 *Ulf Moeller, Bodo Moeller*
16097 * In the `NCONF_...`-based implementations for `CONF_...` queries
16098 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16099 a temporary CONF structure with the data component set to NULL
16100 (which gives segmentation faults in lh_retrieve).
16101 Instead, use NULL for the CONF pointer in CONF_get_string and
16102 CONF_get_number (which may use environment variables) and directly
16103 return NULL from CONF_get_section.
16107 * Fix potential buffer overrun for EBCDIC.
16111 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16112 keyUsage if basicConstraints absent for a CA.
16116 * Make SMIME_write_PKCS7() write mail header values with a format that
16117 is more generally accepted (no spaces before the semicolon), since
16118 some programs can't parse those values properly otherwise. Also make
16119 sure BIO's that break lines after each write do not create invalid
16124 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16125 macros previously used would not encode an empty SEQUENCE OF
16126 and break the signature.
16130 *This change does not apply to 0.9.7.*
16132 * Zero the premaster secret after deriving the master secret in
16137 * Add some EVP_add_digest_alias registrations (as found in
16138 OpenSSL_add_all_digests()) to SSL_library_init()
16139 aka OpenSSL_add_ssl_algorithms(). This provides improved
16140 compatibility with peers using X.509 certificates
16141 with unconventional AlgorithmIdentifier OIDs.
16145 * Fix for Irix with NO_ASM.
16147 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16149 * ./config script fixes.
16151 *Ulf Moeller, Richard Levitte*
16153 * Fix 'openssl passwd -1'.
16157 * Change PKCS12_key_gen_asc() so it can cope with non null
16158 terminated strings whose length is passed in the passlen
16159 parameter, for example from PEM callbacks. This was done
16160 by adding an extra length parameter to asc2uni().
16162 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16164 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16165 call failed, free the DSA structure.
16169 * Fix to uni2asc() to cope with zero length Unicode strings.
16170 These are present in some PKCS#12 files.
16174 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16175 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16176 when writing a 32767 byte record.
16178 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16180 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16181 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16183 (RSA objects have a reference count access to which is protected
16184 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16185 so they are meant to be shared between threads.)
16186 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16187 "Reddie, Steven" <Steven.Reddie@ca.com>*
16189 * Fix a deadlock in CRYPTO_mem_leaks().
16193 * Use better test patterns in bntest.
16197 * rand_win.c fix for Borland C.
16201 * BN_rshift bugfix for n == 0.
16205 * Add a 'bctest' script that checks for some known 'bc' bugs
16206 so that 'make test' does not abort just because 'bc' is broken.
16210 * Store verify_result within SSL_SESSION also for client side to
16211 avoid potential security hole. (Re-used sessions on the client side
16212 always resulted in verify_result==X509_V_OK, not using the original
16213 result of the server certificate verification.)
16217 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16218 SSL3_RT_APPLICATION_DATA, return 0.
16219 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16224 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16225 releases, have been re-implemented by renaming the previous
16226 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16227 and ssl3_read_internal, respectively, and adding 'peek' parameters
16228 to them. The new ssl[23]_{read,peek} functions are calls to
16229 ssl[23]_read_internal with the 'peek' flag set appropriately.
16230 A 'peek' parameter has also been added to ssl3_read_bytes, which
16231 does the actual work for ssl3_read_internal.
16235 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16236 the method-specific "init()" handler. Also clean up ex_data after
16237 calling the method-specific "finish()" handler. Previously, this was
16238 happening the other way round.
16242 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16243 The previous value, 12, was not always sufficient for BN_mod_exp().
16247 * Make sure that shared libraries get the internal name engine with
16248 the full version number and not just 0. This should mark the
16249 shared libraries as not backward compatible. Of course, this should
16250 be changed again when we can guarantee backward binary compatibility.
16254 * Fix typo in get_cert_by_subject() in by_dir.c
16256 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16258 * Rework the system to generate shared libraries:
16260 - Make note of the expected extension for the shared libraries and
16261 if there is a need for symbolic links from for example libcrypto.so.0
16262 to libcrypto.so.0.9.7. There is extended info in Configure for
16265 - Make as few rebuilds of the shared libraries as possible.
16267 - Still avoid linking the OpenSSL programs with the shared libraries.
16269 - When installing, install the shared libraries separately from the
16274 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16276 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16277 and not in SSL_clear because the latter is also used by the
16278 accept/connect functions; previously, the settings made by
16279 SSL_set_read_ahead would be lost during the handshake.
16281 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16283 * Correct util/mkdef.pl to be selective about disabled algorithms.
16284 Previously, it would create entries for disabled algorithms no
16289 * Added several new manual pages for SSL_* function.
16293 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16295 * In ssl23_get_client_hello, generate an error message when faced
16296 with an initial SSL 3.0/TLS record that is too small to contain the
16297 first two bytes of the ClientHello message, i.e. client_version.
16298 (Note that this is a pathologic case that probably has never happened
16299 in real life.) The previous approach was to use the version number
16300 from the record header as a substitute; but our protocol choice
16301 should not depend on that one because it is not authenticated
16302 by the Finished messages.
16306 * More robust randomness gathering functions for Windows.
16308 *Jeffrey Altman <jaltman@columbia.edu>*
16310 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16311 not set then we don't setup the error code for issuer check errors
16312 to avoid possibly overwriting other errors which the callback does
16313 handle. If an application does set the flag then we assume it knows
16314 what it is doing and can handle the new informational codes
16319 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16320 a general "ANY" type, as such it should be able to decode anything
16321 including tagged types. However it didn't check the class so it would
16322 wrongly interpret tagged types in the same way as their universal
16323 counterpart and unknown types were just rejected. Changed so that the
16324 tagged and unknown types are handled in the same way as a SEQUENCE:
16325 that is the encoding is stored intact. There is also a new type
16326 "V_ASN1_OTHER" which is used when the class is not universal, in this
16327 case we have no idea what the actual type is so we just lump them all
16332 * On VMS, stdout may very well lead to a file that is written to
16333 in a record-oriented fashion. That means that every write() will
16334 write a separate record, which will be read separately by the
16335 programs trying to read from it. This can be very confusing.
16337 The solution is to put a BIO filter in the way that will buffer
16338 text until a linefeed is reached, and then write everything a
16339 line at a time, so every record written will be an actual line,
16340 not chunks of lines and not (usually doesn't happen, but I've
16341 seen it once) several lines in one record. BIO_f_linebuffer() is
16344 Currently, it's a VMS-only method, because that's where it has
16345 been tested well enough.
16349 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16350 it can return incorrect results.
16351 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16352 but it was in 0.9.6-beta[12].)
16356 * Disable the check for content being present when verifying detached
16357 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16358 include zero length content when signing messages.
16362 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16363 BIO_ctrl (for BIO pairs).
16367 * Add DSO method for VMS.
16371 * Bug fix: Montgomery multiplication could produce results with the
16376 * Add RPM specification openssl.spec and modify it to build three
16377 packages. The default package contains applications, application
16378 documentation and run-time libraries. The devel package contains
16379 include files, static libraries and function documentation. The
16380 doc package contains the contents of the doc directory. The original
16381 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16385 * Add a large number of documentation files for many SSL routines.
16387 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16389 * Add a configuration entry for Sony News 4.
16391 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16393 * Don't set the two most significant bits to one when generating a
16394 random number < q in the DSA library.
16398 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16399 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16400 the underlying transport is blocking) if a handshake took place.
16401 (The default behaviour is needed by applications such as s_client
16402 and s_server that use select() to determine when to use SSL_read;
16403 but for applications that know in advance when to expect data, it
16404 just makes things more complicated.)
16408 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16413 * Add a few more EBCDIC conditionals that make `req` and `x509`
16414 work better on such systems.
16416 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16418 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16419 Update PKCS12_parse() so it copies the friendlyName and the
16420 keyid to the certificates aux info.
16424 * Fix bug in PKCS7_verify() which caused an infinite loop
16425 if there was more than one signature.
16427 *Sven Uszpelkat <su@celocom.de>*
16429 * Major change in util/mkdef.pl to include extra information
16430 about each symbol, as well as presenting variables as well
16431 as functions. This change means that there's n more need
16432 to rebuild the .num files when some algorithms are excluded.
16436 * Allow the verify time to be set by an application,
16437 rather than always using the current time.
16441 * Phase 2 verify code reorganisation. The certificate
16442 verify code now looks up an issuer certificate by a
16443 number of criteria: subject name, authority key id
16444 and key usage. It also verifies self signed certificates
16445 by the same criteria. The main comparison function is
16446 X509_check_issued() which performs these checks.
16448 Lot of changes were necessary in order to support this
16449 without completely rewriting the lookup code.
16451 Authority and subject key identifier are now cached.
16453 The LHASH 'certs' is X509_STORE has now been replaced
16454 by a STACK_OF(X509_OBJECT). This is mainly because an
16455 LHASH can't store or retrieve multiple objects with
16456 the same hash value.
16458 As a result various functions (which were all internal
16459 use only) have changed to handle the new X509_STORE
16460 structure. This will break anything that messed round
16461 with X509_STORE internally.
16463 The functions X509_STORE_add_cert() now checks for an
16464 exact match, rather than just subject name.
16466 The X509_STORE API doesn't directly support the retrieval
16467 of multiple certificates matching a given criteria, however
16468 this can be worked round by performing a lookup first
16469 (which will fill the cache with candidate certificates)
16470 and then examining the cache for matches. This is probably
16471 the best we can do without throwing out X509_LOOKUP
16472 entirely (maybe later...).
16474 The X509_VERIFY_CTX structure has been enhanced considerably.
16476 All certificate lookup operations now go via a get_issuer()
16477 callback. Although this currently uses an X509_STORE it
16478 can be replaced by custom lookups. This is a simple way
16479 to bypass the X509_STORE hackery necessary to make this
16480 work and makes it possible to use more efficient techniques
16481 in future. A very simple version which uses a simple
16482 STACK for its trusted certificate store is also provided
16483 using X509_STORE_CTX_trusted_stack().
16485 The verify_cb() and verify() callbacks now have equivalents
16486 in the X509_STORE_CTX structure.
16488 X509_STORE_CTX also has a 'flags' field which can be used
16489 to customise the verify behaviour.
16493 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16494 excludes S/MIME capabilities.
16498 * When a certificate request is read in keep a copy of the
16499 original encoding of the signed data and use it when outputting
16500 again. Signatures then use the original encoding rather than
16501 a decoded, encoded version which may cause problems if the
16502 request is improperly encoded.
16506 * For consistency with other BIO_puts implementations, call
16507 buffer_write(b, ...) directly in buffer_puts instead of calling
16510 In BIO_puts, increment b->num_write as in BIO_write.
16512 *Peter.Sylvester@EdelWeb.fr*
16514 * Fix BN_mul_word for the case where the word is 0. (We have to use
16515 BN_zero, we may not return a BIGNUM with an array consisting of
16516 words set to zero.)
16520 * Avoid calling abort() from within the library when problems are
16521 detected, except if preprocessor symbols have been defined
16522 (such as REF_CHECK, BN_DEBUG etc.).
16526 * New openssl application 'rsautl'. This utility can be
16527 used for low-level RSA operations. DER public key
16528 BIO/fp routines also added.
16532 * New Configure entry and patches for compiling on QNX 4.
16534 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16536 * A demo state-machine implementation was sponsored by
16537 Nuron (<http://www.nuron.com/>) and is now available in
16538 demos/state_machine.
16542 * New options added to the 'dgst' utility for signature
16543 generation and verification.
16547 * Unrecognized PKCS#7 content types are now handled via a
16548 catch all ASN1_TYPE structure. This allows unsupported
16549 types to be stored as a "blob" and an application can
16550 encode and decode it manually.
16554 * Fix various signed/unsigned issues to make a_strex.c
16555 compile under VC++.
16557 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16559 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16560 length if passed a buffer. ASN1_INTEGER_to_BN failed
16561 if passed a NULL BN and its argument was negative.
16563 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16565 * Modification to PKCS#7 encoding routines to output definite
16566 length encoding. Since currently the whole structures are in
16567 memory there's not real point in using indefinite length
16568 constructed encoding. However if OpenSSL is compiled with
16569 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16573 * Added BIO_vprintf() and BIO_vsnprintf().
16577 * Added more prefixes to parse for in the strings written
16578 through a logging bio, to cover all the levels that are available
16579 through syslog. The prefixes are now:
16581 PANIC, EMERG, EMR => LOG_EMERG
16582 ALERT, ALR => LOG_ALERT
16583 CRIT, CRI => LOG_CRIT
16584 ERROR, ERR => LOG_ERR
16585 WARNING, WARN, WAR => LOG_WARNING
16586 NOTICE, NOTE, NOT => LOG_NOTICE
16587 INFO, INF => LOG_INFO
16588 DEBUG, DBG => LOG_DEBUG
16590 and as before, if none of those prefixes are present at the
16591 beginning of the string, LOG_ERR is chosen.
16593 On Win32, the `LOG_*` levels are mapped according to this:
16595 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16596 LOG_WARNING => EVENTLOG_WARNING_TYPE
16597 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16601 * Made it possible to reconfigure with just the configuration
16602 argument "reconf" or "reconfigure". The command line arguments
16603 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16604 and are retrieved from there when reconfiguring.
16610 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16612 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16616 * The obj_dat.pl script was messing up the sorting of object
16617 names. The reason was that it compared the quoted version
16618 of strings as a result "OCSP" > "OCSP Signing" because
16619 " > SPACE. Changed script to store unquoted versions of
16620 names and add quotes on output. It was also omitting some
16621 names from the lookup table if they were given a default
16622 value (that is if SN is missing it is given the same
16623 value as LN and vice versa), these are now added on the
16624 grounds that if an object has a name we should be able to
16625 look it up. Finally added warning output when duplicate
16626 short or long names are found.
16630 * Changes needed for Tandem NSK.
16632 *Scott Uroff <scott@xypro.com>*
16634 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16635 RSA_padding_check_SSLv23(), special padding was never detected
16636 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16637 version rollback attacks was not effective.
16639 In s23_clnt.c, don't use special rollback-attack detection padding
16640 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16641 client; similarly, in s23_srvr.c, don't do the rollback check if
16642 SSL 2.0 is the only protocol enabled in the server.
16646 * Make it possible to get hexdumps of unprintable data with 'openssl
16647 asn1parse'. By implication, the functions ASN1_parse_dump() and
16648 BIO_dump_indent() are added.
16652 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16653 these print out strings and name structures based on various
16654 flags including RFC2253 support and proper handling of
16655 multibyte characters. Added options to the 'x509' utility
16656 to allow the various flags to be set.
16660 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16661 Also change the functions X509_cmp_current_time() and
16662 X509_gmtime_adj() work with an ASN1_TIME structure,
16663 this will enable certificates using GeneralizedTime in validity
16664 dates to be checked.
16668 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16669 negative public key encodings) on by default,
16670 NO_NEG_PUBKEY_BUG can be set to disable it.
16674 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16675 content octets. An i2c_ASN1_OBJECT is unnecessary because
16676 the encoding can be trivially obtained from the structure.
16680 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16681 not read locks (`CRYPTO_r_[un]lock`).
16685 * A first attempt at creating official support for shared
16686 libraries through configuration. I've kept it so the
16687 default is static libraries only, and the OpenSSL programs
16688 are always statically linked for now, but there are
16689 preparations for dynamic linking in place.
16690 This has been tested on Linux and Tru64.
16694 * Randomness polling function for Win9x, as described in:
16695 Peter Gutmann, Software Generation of Practically Strong
16700 * Fix so PRNG is seeded in req if using an already existing
16705 * New options to smime application. -inform and -outform
16706 allow alternative formats for the S/MIME message including
16707 PEM and DER. The -content option allows the content to be
16708 specified separately. This should allow things like Netscape
16709 form signing output easier to verify.
16713 * Fix the ASN1 encoding of tags using the 'long form'.
16717 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16718 STRING types. These convert content octets to and from the
16719 underlying type. The actual tag and length octets are
16720 already assumed to have been read in and checked. These
16721 are needed because all other string types have virtually
16722 identical handling apart from the tag. By having versions
16723 of the ASN1 functions that just operate on content octets
16724 IMPLICIT tagging can be handled properly. It also allows
16725 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16726 and ASN1_INTEGER are identical apart from the tag.
16730 * Change the handling of OID objects as follows:
16732 - New object identifiers are inserted in objects.txt, following
16733 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16734 - objects.pl is used to process obj_mac.num and create a new
16736 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16739 This is currently kind of a hack, and the perl code in objects.pl
16740 isn't very elegant, but it works as I intended. The simplest way
16741 to check that it worked correctly is to look in obj_dat.h and
16742 check the array nid_objs and make sure the objects haven't moved
16743 around (this is important!). Additions are OK, as well as
16744 consistent name changes.
16748 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16752 * Addition of the command line parameter '-rand file' to 'openssl req'.
16753 The given file adds to whatever has already been seeded into the
16754 random pool through the RANDFILE configuration file option or
16755 environment variable, or the default random state file.
16759 * mkstack.pl now sorts each macro group into lexical order.
16760 Previously the output order depended on the order the files
16761 appeared in the directory, resulting in needless rewriting
16766 * Patches to make OpenSSL compile under Win32 again. Mostly
16767 work arounds for the VC++ problem that it treats func() as
16768 func(void). Also stripped out the parts of mkdef.pl that
16769 added extra typesafe functions: these no longer exist.
16773 * Reorganisation of the stack code. The macros are now all
16774 collected in safestack.h . Each macro is defined in terms of
16775 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16776 DEBUG_SAFESTACK is now handled in terms of function casts,
16777 this has the advantage of retaining type safety without the
16778 use of additional functions. If DEBUG_SAFESTACK is not defined
16779 then the non typesafe macros are used instead. Also modified the
16780 mkstack.pl script to handle the new form. Needs testing to see
16781 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16782 the default if no major problems. Similar behaviour for ASN1_SET_OF
16783 and PKCS12_STACK_OF.
16787 * When some versions of IIS use the 'NET' form of private key the
16788 key derivation algorithm is different. Normally MD5(password) is
16789 used as a 128 bit RC4 key. In the modified case
16790 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16791 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16792 as the old Netscape_RSA functions except they have an additional
16793 'sgckey' parameter which uses the modified algorithm. Also added
16794 an -sgckey command line option to the rsa utility. Thanks to
16795 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16796 algorithm to openssl-dev.
16800 * The evp_local.h macros were using 'c.##kname' which resulted in
16801 invalid expansion on some systems (SCO 5.0.5 for example).
16802 Corrected to 'c.kname'.
16804 *Phillip Porch <root@theporch.com>*
16806 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16807 a STACK of email addresses from a certificate or request, these look
16808 in the subject name and the subject alternative name extensions and
16809 omit any duplicate addresses.
16813 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16814 This makes DSA verification about 2 % faster.
16818 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16819 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16820 plus overhead for 1024 bit moduli).
16821 This makes exponentiations about 0.5 % faster for 1024 bit
16822 exponents (as measured by "openssl speed rsa2048").
16826 * Rename memory handling macros to avoid conflicts with other
16828 Malloc => OPENSSL_malloc
16829 Malloc_locked => OPENSSL_malloc_locked
16830 Realloc => OPENSSL_realloc
16831 Free => OPENSSL_free
16835 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16836 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16840 * CygWin32 support.
16842 *John Jarvie <jjarvie@newsguy.com>*
16844 * The type-safe stack code has been rejigged. It is now only compiled
16845 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16846 by default all type-specific stack functions are "#define"d back to
16847 standard stack functions. This results in more streamlined output
16848 but retains the type-safety checking possibilities of the original
16853 * The STACK code has been cleaned up, and certain type declarations
16854 that didn't make a lot of sense have been brought in line. This has
16855 also involved a cleanup of sorts in safestack.h to more correctly
16856 map type-safe stack functions onto their plain stack counterparts.
16857 This work has also resulted in a variety of "const"ifications of
16858 lots of the code, especially `_cmp` operations which should normally
16859 be prototyped with "const" parameters anyway.
16863 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16864 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16865 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16866 where all of 'md' is used each time the PRNG is used, but 'state'
16867 is used only indexed by a cyclic counter. As entropy may not be
16868 well distributed from the beginning, 'md' is important as a
16869 chaining variable. However, the output function chains only half
16870 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16871 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16872 in all of 'state' being rewritten, with the new values depending
16873 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16877 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16878 the handshake is continued after ssl_verify_cert_chain();
16879 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16880 can lead to 'unexplainable' connection aborts later.
16882 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16884 * Major EVP API cipher revision.
16885 Add hooks for extra EVP features. This allows various cipher
16886 parameters to be set in the EVP interface. Support added for variable
16887 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16888 setting of RC2 and RC5 parameters.
16890 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16893 Remove lots of duplicated code from the EVP library. For example *every*
16894 cipher init() function handles the 'iv' in the same way according to the
16895 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16896 for CFB and OFB modes they zero ctx->num.
16898 New functionality allows removal of S/MIME code RC2 hack.
16900 Most of the routines have the same form and so can be declared in terms
16903 By shifting this to the top level EVP_CipherInit() it can be removed from
16904 all individual ciphers. If the cipher wants to handle IVs or keys
16905 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16908 Change lots of functions like EVP_EncryptUpdate() to now return a
16909 value: although software versions of the algorithms cannot fail
16910 any installed hardware versions can.
16914 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16915 this option is set, tolerate broken clients that send the negotiated
16916 protocol version number instead of the requested protocol version
16921 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16922 i.e. non-zero for export ciphersuites, zero otherwise.
16923 Previous versions had this flag inverted, inconsistent with
16924 rsa_tmp_cb (..._TMP_RSA_CB).
16926 *Bodo Moeller; problem reported by Amit Chopra*
16928 * Add missing DSA library text string. Work around for some IIS
16929 key files with invalid SEQUENCE encoding.
16933 * Add a document (doc/standards.txt) that list all kinds of standards
16934 and so on that are implemented in OpenSSL.
16938 * Enhance c_rehash script. Old version would mishandle certificates
16939 with the same subject name hash and wouldn't handle CRLs at all.
16940 Added -fingerprint option to crl utility, to support new c_rehash
16945 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16949 * Fix for SSL server purpose checking. Server checking was
16950 rejecting certificates which had extended key usage present
16951 but no ssl client purpose.
16953 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16955 * Make PKCS#12 code work with no password. The PKCS#12 spec
16956 is a little unclear about how a blank password is handled.
16957 Since the password in encoded as a BMPString with terminating
16958 double NULL a zero length password would end up as just the
16959 double NULL. However no password at all is different and is
16960 handled differently in the PKCS#12 key generation code. NS
16961 treats a blank password as zero length. MSIE treats it as no
16962 password on export: but it will try both on import. We now do
16963 the same: PKCS12_parse() tries zero length and no password if
16964 the password is set to "" or NULL (NULL is now a valid password:
16965 it wasn't before) as does the pkcs12 application.
16969 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16970 perror when PEM_read_bio_X509_REQ fails, the error message must
16971 be obtained from the error queue.
16975 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16976 it in ERR_remove_state if appropriate, and change ERR_get_state
16977 accordingly to avoid race conditions (this is necessary because
16978 thread_hash is no longer constant once set).
16982 * Bugfix for linux-elf makefile.one.
16986 * RSA_get_default_method() will now cause a default
16987 RSA_METHOD to be chosen if one doesn't exist already.
16988 Previously this was only set during a call to RSA_new()
16989 or RSA_new_method(NULL) meaning it was possible for
16990 RSA_get_default_method() to return NULL.
16994 * Added native name translation to the existing DSO code
16995 that will convert (if the flag to do so is set) filenames
16996 that are sufficiently small and have no path information
16997 into a canonical native form. Eg. "blah" converted to
16998 "libblah.so" or "blah.dll" etc.
17002 * New function ERR_error_string_n(e, buf, len) which is like
17003 ERR_error_string(e, buf), but writes at most 'len' bytes
17004 including the 0 terminator. For ERR_error_string_n, 'buf'
17007 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17009 * CONF library reworked to become more general. A new CONF
17010 configuration file reader "class" is implemented as well as a
17011 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17012 old `CONF_*` functions are still there, but are reimplemented to
17013 work in terms of the new functions. Also, a set of functions
17014 to handle the internal storage of the configuration data is
17015 provided to make it easier to write new configuration file
17016 reader "classes" (I can definitely see something reading a
17017 configuration file in XML format, for example), called `_CONF_*`,
17018 or "the configuration storage API"...
17020 The new configuration file reading functions are:
17022 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17023 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17025 NCONF_default, NCONF_WIN32
17027 NCONF_dump_fp, NCONF_dump_bio
17029 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17030 NCONF_new creates a new CONF object. This works in the same way
17031 as other interfaces in OpenSSL, like the BIO interface.
17032 `NCONF_dump_*` dump the internal storage of the configuration file,
17033 which is useful for debugging. All other functions take the same
17034 arguments as the old `CONF_*` functions with the exception of the
17035 first that must be a `CONF *` instead of a `LHASH *`.
17037 To make it easier to use the new classes with the old `CONF_*` functions,
17038 the function CONF_set_default_method is provided.
17042 * Add '-tls1' option to 'openssl ciphers', which was already
17043 mentioned in the documentation but had not been implemented.
17044 (This option is not yet really useful because even the additional
17045 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17049 * Initial DSO code added into libcrypto for letting OpenSSL (and
17050 OpenSSL-based applications) load shared libraries and bind to
17051 them in a portable way.
17053 *Geoff Thorpe, with contributions from Richard Levitte*
17055 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17057 * Make sure _lrotl and _lrotr are only used with MSVC.
17059 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17060 (the default implementation of RAND_status).
17062 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17063 to '-clrext' (= clear extensions), as intended and documented.
17064 *Bodo Moeller; inconsistency pointed out by Michael Attili
17065 <attili@amaxo.com>*
17067 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17068 was larger than the MD block size.
17070 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17072 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17073 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17074 using the passed key: if the passed key was a private key the result
17075 of X509_print(), for example, would be to print out all the private key
17080 * des_quad_cksum() byte order bug fix.
17081 *Ulf Möller, using the problem description in krb4-0.9.7, where
17082 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17084 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17087 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17089 * For easily testing in shell scripts whether some command
17090 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17091 returns with exit code 0 iff no command of the given name is available.
17092 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17093 the output goes to stdout and nothing is printed to stderr.
17094 Additional arguments are always ignored.
17096 Since for each cipher there is a command of the same name,
17097 the 'no-cipher' compilation switches can be tested this way.
17099 ('openssl no-XXX' is not able to detect pseudo-commands such
17100 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17104 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17108 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17109 is set; it will be thrown away anyway because each handshake creates
17111 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17112 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17113 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17114 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17118 * New s_client option -ign_eof: EOF at stdin is ignored, and
17119 'Q' and 'R' lose their special meanings (quit/renegotiate).
17120 This is part of what -quiet does; unlike -quiet, -ign_eof
17121 does not suppress any output.
17125 * Add compatibility options to the purpose and trust code. The
17126 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17127 accepts a certificate or CA, this was the previous behaviour,
17128 with all the associated security issues.
17130 X509_TRUST_COMPAT is the old trust behaviour: only and
17131 automatically trust self signed roots in certificate store. A
17132 new trust setting X509_TRUST_DEFAULT is used to specify that
17133 a purpose has no associated trust setting and it should instead
17134 use the value in the default purpose.
17138 * Fix the PKCS#8 DSA private key code so it decodes keys again
17139 and fix a memory leak.
17143 * In util/mkerr.pl (which implements 'make errors'), preserve
17144 reason strings from the previous version of the .c file, as
17145 the default to have only downcase letters (and digits) in
17146 automatically generated reasons codes is not always appropriate.
17150 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17151 using strerror. Previously, ERR_reason_error_string() returned
17152 library names as reason strings for SYSerr; but SYSerr is a special
17153 case where small numbers are errno values, not library numbers.
17157 * Add '-dsaparam' option to 'openssl dhparam' application. This
17158 converts DSA parameters into DH parameters. (When creating parameters,
17159 DSA_generate_parameters is used.)
17163 * Include 'length' (recommended exponent length) in C code generated
17164 by 'openssl dhparam -C'.
17168 * The second argument to set_label in perlasm was already being used
17169 so couldn't be used as a "file scope" flag. Moved to third argument
17174 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17175 instead of RAND_bytes for encryption IVs and salts.
17179 * Include RAND_status() into RAND_METHOD instead of implementing
17180 it only for md_rand.c Otherwise replacing the PRNG by calling
17181 RAND_set_rand_method would be impossible.
17185 * Don't let DSA_generate_key() enter an infinite loop if the random
17186 number generation fails.
17190 * New 'rand' application for creating pseudo-random output.
17194 * Added configuration support for Linux/IA64
17196 *Rolf Haberrecker <rolf@suse.de>*
17198 * Assembler module support for Mingw32.
17202 * Shared library support for HPUX (in shlib/).
17204 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17206 * Shared library support for Solaris gcc.
17208 *Lutz Behnke <behnke@trustcenter.de>*
17210 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17212 * PKCS7_encrypt() was adding text MIME headers twice because they
17213 were added manually and by SMIME_crlf_copy().
17217 * In bntest.c don't call BN_rand with zero bits argument.
17219 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17221 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17222 case was implemented. This caused BN_div_recp() to fail occasionally.
17226 * Add an optional second argument to the set_label() in the perl
17227 assembly language builder. If this argument exists and is set
17228 to 1 it signals that the assembler should use a symbol whose
17229 scope is the entire file, not just the current function. This
17230 is needed with MASM which uses the format label:: for this scope.
17232 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17234 * Change the ASN1 types so they are typedefs by default. Before
17235 almost all types were #define'd to ASN1_STRING which was causing
17236 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17241 * Change names of new functions to the new get1/get0 naming
17242 convention: After 'get1', the caller owns a reference count
17243 and has to call `..._free`; 'get0' returns a pointer to some
17244 data structure without incrementing reference counters.
17245 (Some of the existing 'get' functions increment a reference
17246 counter, some don't.)
17247 Similarly, 'set1' and 'add1' functions increase reference
17248 counters or duplicate objects.
17252 * Allow for the possibility of temp RSA key generation failure:
17253 the code used to assume it always worked and crashed on failure.
17257 * Fix potential buffer overrun problem in BIO_printf().
17258 *Ulf Möller, using public domain code by Patrick Powell; problem
17259 pointed out by David Sacerdote <das33@cornell.edu>*
17261 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17262 RAND_egd() and RAND_status(). In the command line application,
17263 the EGD socket can be specified like a seed file using RANDFILE
17268 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17269 Some CAs (e.g. Verisign) distribute certificates in this form.
17273 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17274 list to exclude them. This means that no special compilation option
17275 is needed to use anonymous DH: it just needs to be included in the
17280 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17281 EVP_MD_type. The old functionality is available in a new macro called
17282 EVP_MD_md(). Change code that uses it and update docs.
17286 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17287 where the `void *` argument is replaced by a function pointer argument.
17288 Previously `void *` was abused to point to functions, which works on
17289 many platforms, but is not correct. As these functions are usually
17290 called by macros defined in OpenSSL header files, most source code
17291 should work without changes.
17295 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17296 sections with information on -D... compiler switches used for
17297 compiling the library so that applications can see them. To enable
17298 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17299 must be defined. E.g.,
17300 #define OPENSSL_ALGORITHM_DEFINES
17301 #include <openssl/opensslconf.h>
17302 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17304 *Richard Levitte, Ulf and Bodo Möller*
17306 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17311 * Change the 'other' type in certificate aux info to a STACK_OF
17312 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17313 the required ASN1 format: arbitrary types determined by an OID.
17317 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17318 argument to 'req'. This is not because the function is newer or
17319 better than others it just uses the work 'NEW' in the certificate
17320 request header lines. Some software needs this.
17324 * Reorganise password command line arguments: now passwords can be
17325 obtained from various sources. Delete the PEM_cb function and make
17326 it the default behaviour: i.e. if the callback is NULL and the
17327 usrdata argument is not NULL interpret it as a null terminated pass
17328 phrase. If usrdata and the callback are NULL then the pass phrase
17329 is prompted for as usual.
17333 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17334 the support is automatically enabled. The resulting binaries will
17335 autodetect the card and use it if present.
17337 *Ben Laurie and Compaq Inc.*
17339 * Work around for Netscape hang bug. This sends certificate request
17340 and server done in one record. Since this is perfectly legal in the
17341 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17342 the bugs/SSLv3 entry for more info.
17346 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17350 * Add -rand argument to smime and pkcs12 applications and read/write
17355 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17359 * Add command line password options to the remaining applications.
17363 * Bug fix for BN_div_recp() for numerators with an even number of
17368 * More tests in bntest.c, and changed test_bn output.
17372 * ./config recognizes MacOS X now.
17376 * Bug fix for BN_div() when the first words of num and divisor are
17377 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17381 * Add support for various broken PKCS#8 formats, and command line
17382 options to produce them.
17386 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17387 get temporary BIGNUMs from a BN_CTX.
17391 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17396 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17397 include a #define from the old name to the new. The original intent
17398 was that statically linked binaries could for example just call
17399 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17400 link with digests. This never worked because SSLeay_add_all_digests()
17401 and SSLeay_add_all_ciphers() were in the same source file so calling
17402 one would link with the other. They are now in separate source files.
17406 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17410 * Use a less unusual form of the Miller-Rabin primality test (it used
17411 a binary algorithm for exponentiation integrated into the Miller-Rabin
17412 loop, our standard modexp algorithms are faster).
17416 * Support for the EBCDIC character set completed.
17418 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17420 * Source code cleanups: use const where appropriate, eliminate casts,
17421 use `void *` instead of `char *` in lhash.
17425 * Bugfix: ssl3_send_server_key_exchange was not restartable
17426 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17427 this the server could overwrite ephemeral keys that the client
17432 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17433 using 50 iterations of the Rabin-Miller test.
17435 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17436 iterations of the Rabin-Miller test as required by the appendix
17437 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17438 As BN_is_prime_fasttest includes trial division, DSA parameter
17439 generation becomes much faster.
17441 This implies a change for the callback functions in DSA_is_prime
17442 and DSA_generate_parameters: The callback function is called once
17443 for each positive witness in the Rabin-Miller test, not just
17444 occasionally in the inner loop; and the parameters to the
17445 callback function now provide an iteration count for the outer
17446 loop rather than for the current invocation of the inner loop.
17447 DSA_generate_parameters additionally can call the callback
17448 function with an 'iteration count' of -1, meaning that a
17449 candidate has passed the trial division test (when q is generated
17450 from an application-provided seed, trial division is skipped).
17454 * New function BN_is_prime_fasttest that optionally does trial
17455 division before starting the Rabin-Miller test and has
17456 an additional BN_CTX * argument (whereas BN_is_prime always
17457 has to allocate at least one BN_CTX).
17458 'callback(1, -1, cb_arg)' is called when a number has passed the
17459 trial division stage.
17463 * Fix for bug in CRL encoding. The validity dates weren't being handled
17468 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17472 * New function BN_pseudo_rand().
17476 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17477 bignum version of BN_from_montgomery() with the working code from
17478 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17483 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17484 made it impossible to use the same SSL_SESSION data structure in
17485 SSL2 clients in multiple threads.
17489 * The return value of RAND_load_file() no longer counts bytes obtained
17490 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17491 to seed the PRNG (previously an explicit byte count was required).
17493 *Ulf Möller, Bodo Möller*
17495 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17496 used `char *` instead of `void *` and had casts all over the place.
17500 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17504 * Retain source code compatibility for BN_prime_checks macro:
17505 BN_is_prime(..., BN_prime_checks, ...) now uses
17506 BN_prime_checks_for_size to determine the appropriate number of
17507 Rabin-Miller iterations.
17511 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17512 DH_CHECK_P_NOT_SAFE_PRIME.
17513 (Check if this is true? OpenPGP calls them "strong".)
17517 * Merge the functionality of "dh" and "gendh" programs into a new program
17518 "dhparam". The old programs are retained for now but will handle DH keys
17519 (instead of parameters) in future.
17523 * Make the ciphers, s_server and s_client programs check the return values
17524 when a new cipher list is set.
17528 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17529 ciphers. Before when the 56bit ciphers were enabled the sorting was
17532 The syntax for the cipher sorting has been extended to support sorting by
17533 cipher-strength (using the strength_bits hard coded in the tables).
17534 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17536 Fix a bug in the cipher-command parser: when supplying a cipher command
17537 string with an "undefined" symbol (neither command nor alphanumeric
17538 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17539 an error is flagged.
17541 Due to the strength-sorting extension, the code of the
17542 ssl_create_cipher_list() function was completely rearranged. I hope that
17543 the readability was also increased :-)
17545 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17547 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17548 for the first serial number and places 2 in the serial number file. This
17549 avoids problems when the root CA is created with serial number zero and
17550 the first user certificate has the same issuer name and serial number
17555 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17556 the new code. Add documentation for this stuff.
17560 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17561 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17562 structures and behave in an analogous way to the X509v3 functions:
17563 they shouldn't be called directly but wrapper functions should be used
17566 So we also now have some wrapper functions that call the X509at functions
17567 when passed certificate requests. (TO DO: similar things can be done with
17568 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17569 things. Some of these need some d2i or i2d and print functionality
17570 because they handle more complex structures.)
17574 * Add missing #ifndefs that caused missing symbols when building libssl
17575 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17576 NO_RSA in `ssl/s2*.c`.
17578 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17580 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17581 has a return value which indicates the quality of the random data
17582 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17583 error queue. New function RAND_pseudo_bytes() generates output that is
17584 guaranteed to be unique but not unpredictable. RAND_add is like
17585 RAND_seed, but takes an extra argument for an entropy estimate
17586 (RAND_seed always assumes full entropy).
17590 * Do more iterations of Rabin-Miller probable prime test (specifically,
17591 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17592 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17593 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17594 false-positive rate of at most 2^-80 for random input.
17598 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17602 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17603 in the 0.9.5 release), this returns the chain
17604 from an X509_CTX structure with a dup of the stack and all
17605 the X509 reference counts upped: so the stack will exist
17606 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17609 Also make SSL_SESSION_print() print out the verify return
17614 * Add manpage for the pkcs12 command. Also change the default
17615 behaviour so MAC iteration counts are used unless the new
17616 -nomaciter option is used. This improves file security and
17617 only older versions of MSIE (4.0 for example) need it.
17621 * Honor the no-xxx Configure options when creating .DEF files.
17625 * Add PKCS#10 attributes to field table: challengePassword,
17626 unstructuredName and unstructuredAddress. These are taken from
17627 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17628 international characters are used.
17630 More changes to X509_ATTRIBUTE code: allow the setting of types
17631 based on strings. Remove the 'loc' parameter when adding
17632 attributes because these will be a SET OF encoding which is sorted
17637 * Initial changes to the 'req' utility to allow request generation
17638 automation. This will allow an application to just generate a template
17639 file containing all the field values and have req construct the
17642 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17643 used all over the place including certificate requests and PKCS#7
17644 structures. They are currently handled manually where necessary with
17645 some primitive wrappers for PKCS#7. The new functions behave in a
17646 manner analogous to the X509 extension functions: they allow
17647 attributes to be looked up by NID and added.
17649 Later something similar to the X509V3 code would be desirable to
17650 automatically handle the encoding, decoding and printing of the
17651 more complex types. The string types like challengePassword can
17652 be handled by the string table functions.
17654 Also modified the multi byte string table handling. Now there is
17655 a 'global mask' which masks out certain types. The table itself
17656 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17657 is useful when for example there is only one permissible type
17658 (as in countryName) and using the mask might result in no valid
17663 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17664 SSL_get_peer_finished to allow applications to obtain the latest
17665 Finished messages sent to the peer or expected from the peer,
17666 respectively. (SSL_get_peer_finished is usually the Finished message
17667 actually received from the peer, otherwise the protocol will be aborted.)
17669 As the Finished message are message digests of the complete handshake
17670 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17671 be used for external authentication procedures when the authentication
17672 provided by SSL/TLS is not desired or is not enough.
17676 * Enhanced support for Alpha Linux is added. Now ./config checks if
17677 the host supports BWX extension and if Compaq C is present on the
17678 $PATH. Just exploiting of the BWX extension results in 20-30%
17679 performance kick for some algorithms, e.g. DES and RC4 to mention
17680 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17685 * Add support for MS "fast SGC". This is arguably a violation of the
17686 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17687 weak crypto and after checking the certificate is SGC a second one
17688 with strong crypto. MS SGC stops the first handshake after receiving
17689 the server certificate message and sends a second client hello. Since
17690 a server will typically do all the time consuming operations before
17691 expecting any further messages from the client (server key exchange
17692 is the most expensive) there is little difference between the two.
17694 To get OpenSSL to support MS SGC we have to permit a second client
17695 hello message after we have sent server done. In addition we have to
17696 reset the MAC if we do get this second client hello.
17700 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17701 if a DER encoded private key is RSA or DSA traditional format. Changed
17702 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17703 format DER encoded private key. Newer code should use PKCS#8 format which
17704 has the key type encoded in the ASN1 structure. Added DER private key
17705 support to pkcs8 application.
17709 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17710 ciphersuites has been selected (as required by the SSL 3/TLS 1
17711 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17712 is set, we interpret this as a request to violate the specification
17713 (the worst that can happen is a handshake failure, and 'correct'
17714 behaviour would result in a handshake failure anyway).
17718 * In SSL_CTX_add_session, take into account that there might be multiple
17719 SSL_SESSION structures with the same session ID (e.g. when two threads
17720 concurrently obtain them from an external cache).
17721 The internal cache can handle only one SSL_SESSION with a given ID,
17722 so if there's a conflict, we now throw out the old one to achieve
17727 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17728 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17729 some routines that use cipher OIDs: some ciphers do not have OIDs
17730 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17735 * Simplify the trust setting structure and code. Now we just have
17736 two sequences of OIDs for trusted and rejected settings. These will
17737 typically have values the same as the extended key usage extension
17738 and any application specific purposes.
17740 The trust checking code now has a default behaviour: it will just
17741 check for an object with the same NID as the passed id. Functions can
17742 be provided to override either the default behaviour or the behaviour
17743 for a given id. SSL client, server and email already have functions
17744 in place for compatibility: they check the NID and also return "trusted"
17745 if the certificate is self signed.
17749 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17750 traditional format into an EVP_PKEY structure.
17754 * Add a password callback function PEM_cb() which either prompts for
17755 a password if usr_data is NULL or otherwise assumes it is a null
17756 terminated password. Allow passwords to be passed on command line
17757 environment or config files in a few more utilities.
17761 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17762 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17763 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17764 Update documentation.
17768 * Support for ASN1 "NULL" type. This could be handled before by using
17769 ASN1_TYPE but there wasn't any function that would try to read a NULL
17770 and produce an error if it couldn't. For compatibility we also have
17771 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17772 don't allocate anything because they don't need to.
17776 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17779 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17781 * Rebuild of the memory allocation routines used by OpenSSL code and
17782 possibly others as well. The purpose is to make an interface that
17783 provide hooks so anyone can build a separate set of allocation and
17784 deallocation routines to be used by OpenSSL, for example memory
17785 pool implementations, or something else, which was previously hard
17786 since Malloc(), Realloc() and Free() were defined as macros having
17787 the values malloc, realloc and free, respectively (except for Win32
17788 compilations). The same is provided for memory debugging code.
17789 OpenSSL already comes with functionality to find memory leaks, but
17790 this gives people a chance to debug other memory problems.
17792 With these changes, a new set of functions and macros have appeared:
17794 CRYPTO_set_mem_debug_functions() [F]
17795 CRYPTO_get_mem_debug_functions() [F]
17796 CRYPTO_dbg_set_options() [F]
17797 CRYPTO_dbg_get_options() [F]
17798 CRYPTO_malloc_debug_init() [M]
17800 The memory debug functions are NULL by default, unless the library
17801 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17802 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17803 gives the standard debugging functions that come with OpenSSL) or
17804 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17805 provided by the library user) must be used. When the standard
17806 debugging functions are used, CRYPTO_dbg_set_options can be used to
17807 request additional information:
17808 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17809 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17811 Also, things like CRYPTO_set_mem_functions will always give the
17812 expected result (the new set of functions is used for allocation
17813 and deallocation) at all times, regardless of platform and compiler
17816 To finish it up, some functions that were never use in any other
17817 way than through macros have a new API and new semantic:
17819 CRYPTO_dbg_malloc()
17820 CRYPTO_dbg_realloc()
17823 All macros of value have retained their old syntax.
17825 *Richard Levitte and Bodo Moeller*
17827 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17828 ordering of SMIMECapabilities wasn't in "strength order" and there
17829 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17834 * Some ASN1 types with illegal zero length encoding (INTEGER,
17835 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17837 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17839 * Merge in my S/MIME library for OpenSSL. This provides a simple
17840 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17841 functionality to handle multipart/signed properly) and a utility
17842 called 'smime' to call all this stuff. This is based on code I
17843 originally wrote for Celo who have kindly allowed it to be
17844 included in OpenSSL.
17848 * Add variants des_set_key_checked and des_set_key_unchecked of
17849 des_set_key (aka des_key_sched). Global variable des_check_key
17850 decides which of these is called by des_set_key; this way
17851 des_check_key behaves as it always did, but applications and
17852 the library itself, which was buggy for des_check_key == 1,
17853 have a cleaner way to pick the version they need.
17857 * New function PKCS12_newpass() which changes the password of a
17862 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17863 dynamic mix. In both cases the ids can be used as an index into the
17864 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17865 functions so they accept a list of the field values and the
17866 application doesn't need to directly manipulate the X509_TRUST
17871 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17876 * Modify the way the V3 extension code looks up extensions. This now
17877 works in a similar way to the object code: we have some "standard"
17878 extensions in a static table which is searched with OBJ_bsearch()
17879 and the application can add dynamic ones if needed. The file
17880 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17881 updated whenever a new extension is added to the core code and kept
17882 in ext_nid order. There is a simple program 'tabtest.c' which checks
17883 this. New extensions are not added too often so this file can readily
17884 be maintained manually.
17886 There are two big advantages in doing things this way. The extensions
17887 can be looked up immediately and no longer need to be "added" using
17888 X509V3_add_standard_extensions(): this function now does nothing.
17889 Side note: I get *lots* of email saying the extension code doesn't
17890 work because people forget to call this function.
17891 Also no dynamic allocation is done unless new extensions are added:
17892 so if we don't add custom extensions there is no need to call
17893 X509V3_EXT_cleanup().
17897 * Modify enc utility's salting as follows: make salting the default. Add a
17898 magic header, so unsalted files fail gracefully instead of just decrypting
17899 to garbage. This is because not salting is a big security hole, so people
17900 should be discouraged from doing it.
17904 * Fixes and enhancements to the 'x509' utility. It allowed a message
17905 digest to be passed on the command line but it only used this
17906 parameter when signing a certificate. Modified so all relevant
17907 operations are affected by the digest parameter including the
17908 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17909 DSA key was used because it didn't fix the digest.
17913 * Initial certificate chain verify code. Currently tests the untrusted
17914 certificates for consistency with the verify purpose (which is set
17915 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17917 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17918 this is because it will reject chains with invalid extensions whereas
17919 every previous version of OpenSSL and SSLeay made no checks at all.
17921 Trust code: checks the root CA for the relevant trust settings. Trust
17922 settings have an initial value consistent with the verify purpose: e.g.
17923 if the verify purpose is for SSL client use it expects the CA to be
17924 trusted for SSL client use. However the default value can be changed to
17925 permit custom trust settings: one example of this would be to only trust
17926 certificates from a specific "secure" set of CAs.
17928 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17929 which should be used for version portability: especially since the
17930 verify structure is likely to change more often now.
17932 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17933 to set them. If not set then assume SSL clients will verify SSL servers
17936 Two new options to the verify program: -untrusted allows a set of
17937 untrusted certificates to be passed in and -purpose which sets the
17938 intended purpose of the certificate. If a purpose is set then the
17939 new chain verify code is used to check extension consistency.
17943 * Support for the authority information access extension.
17947 * Modify RSA and DSA PEM read routines to transparently handle
17948 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17949 public keys in a format compatible with certificate
17950 SubjectPublicKeyInfo structures. Unfortunately there were already
17951 functions called *_PublicKey_* which used various odd formats so
17952 these are retained for compatibility: however the DSA variants were
17953 never in a public release so they have been deleted. Changed dsa/rsa
17954 utilities to handle the new format: note no releases ever handled public
17955 keys so we should be OK.
17957 The primary motivation for this change is to avoid the same fiasco
17958 that dogs private keys: there are several incompatible private key
17959 formats some of which are standard and some OpenSSL specific and
17960 require various evil hacks to allow partial transparent handling and
17961 even then it doesn't work with DER formats. Given the option anything
17962 other than PKCS#8 should be dumped: but the other formats have to
17963 stay in the name of compatibility.
17965 With public keys and the benefit of hindsight one standard format
17966 is used which works with EVP_PKEY, RSA or DSA structures: though
17967 it clearly returns an error if you try to read the wrong kind of key.
17969 Added a -pubkey option to the 'x509' utility to output the public key.
17970 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17971 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17972 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17973 that do the same as the `EVP_PKEY_assign_*()` except they up the
17974 reference count of the added key (they don't "swallow" the
17979 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17980 CRLs would fail if the file contained no certificates or no CRLs:
17981 added a new function to read in both types and return the number
17982 read: this means that if none are read it will be an error. The
17983 DER versions of the certificate and CRL reader would always fail
17984 because it isn't possible to mix certificates and CRLs in DER format
17985 without choking one or the other routine. Changed this to just read
17986 a certificate: this is the best we can do. Also modified the code
17987 in `apps/verify.c` to take notice of return codes: it was previously
17988 attempting to read in certificates from NULL pointers and ignoring
17989 any errors: this is one reason why the cert and CRL reader seemed
17990 to work. It doesn't check return codes from the default certificate
17991 routines: these may well fail if the certificates aren't installed.
17995 * Code to support otherName option in GeneralName.
17999 * First update to verify code. Change the verify utility
18000 so it warns if it is passed a self signed certificate:
18001 for consistency with the normal behaviour. X509_verify
18002 has been modified to it will now verify a self signed
18003 certificate if *exactly* the same certificate appears
18004 in the store: it was previously impossible to trust a
18005 single self signed certificate. This means that:
18006 openssl verify ss.pem
18007 now gives a warning about a self signed certificate but
18008 openssl verify -CAfile ss.pem ss.pem
18013 * For servers, store verify_result in SSL_SESSION data structure
18014 (and add it to external session representation).
18015 This is needed when client certificate verifications fails,
18016 but an application-provided verification callback (set by
18017 SSL_CTX_set_cert_verify_callback) allows accepting the session
18018 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18019 but returns 1): When the session is reused, we have to set
18020 ssl->verify_result to the appropriate error code to avoid
18023 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18025 * Fix a bug in the new PKCS#7 code: it didn't consider the
18026 case in PKCS7_dataInit() where the signed PKCS7 structure
18027 didn't contain any existing data because it was being created.
18029 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18031 * Add a salt to the key derivation routines in enc.c. This
18032 forms the first 8 bytes of the encrypted file. Also add a
18033 -S option to allow a salt to be input on the command line.
18037 * New function X509_cmp(). Oddly enough there wasn't a function
18038 to compare two certificates. We do this by working out the SHA1
18039 hash and comparing that. X509_cmp() will be needed by the trust
18044 * SSL_get1_session() is like SSL_get_session(), but increments
18045 the reference count in the SSL_SESSION returned.
18047 *Geoff Thorpe <geoff@eu.c2.net>*
18049 * Fix for 'req': it was adding a null to request attributes.
18050 Also change the X509_LOOKUP and X509_INFO code to handle
18051 certificate auxiliary information.
18055 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18060 * Add the possibility to add extra information to the memory leak
18061 detecting output, to form tracebacks, showing from where each
18062 allocation was originated: CRYPTO_push_info("constant string") adds
18063 the string plus current file name and line number to a per-thread
18064 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18065 is like calling CYRPTO_pop_info() until the stack is empty.
18066 Also updated memory leak detection code to be multi-thread-safe.
18070 * Add options -text and -noout to pkcs7 utility and delete the
18071 encryption options which never did anything. Update docs.
18075 * Add options to some of the utilities to allow the pass phrase
18076 to be included on either the command line (not recommended on
18077 OSes like Unix) or read from the environment. Update the
18078 manpages and fix a few bugs.
18082 * Add a few manpages for some of the openssl commands.
18086 * Fix the -revoke option in ca. It was freeing up memory twice,
18087 leaking and not finding already revoked certificates.
18091 * Extensive changes to support certificate auxiliary information.
18092 This involves the use of X509_CERT_AUX structure and X509_AUX
18093 functions. An X509_AUX function such as PEM_read_X509_AUX()
18094 can still read in a certificate file in the usual way but it
18095 will also read in any additional "auxiliary information". By
18096 doing things this way a fair degree of compatibility can be
18097 retained: existing certificates can have this information added
18098 using the new 'x509' options.
18100 Current auxiliary information includes an "alias" and some trust
18101 settings. The trust settings will ultimately be used in enhanced
18102 certificate chain verification routines: currently a certificate
18103 can only be trusted if it is self signed and then it is trusted
18108 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18109 The problem was that one of the replacement routines had not been working
18110 since SSLeay releases. For now the offending routine has been replaced
18111 with non-optimised assembler. Even so, this now gives around 95%
18112 performance improvement for 1024 bit RSA signs.
18116 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18117 handling. Most clients have the effective key size in bits equal to
18118 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18119 A few however don't do this and instead use the size of the decrypted key
18120 to determine the RC2 key length and the AlgorithmIdentifier to determine
18121 the effective key length. In this case the effective key length can still
18122 be 40 bits but the key length can be 168 bits for example. This is fixed
18123 by manually forcing an RC2 key into the EVP_PKEY structure because the
18124 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18125 the key length and effective key length are equal.
18129 * Add a bunch of functions that should simplify the creation of
18130 X509_NAME structures. Now you should be able to do:
18131 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18132 and have it automatically work out the correct field type and fill in
18133 the structures. The more adventurous can try:
18134 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18135 and it will (hopefully) work out the correct multibyte encoding.
18139 * Change the 'req' utility to use the new field handling and multibyte
18140 copy routines. Before the DN field creation was handled in an ad hoc
18141 way in req, ca, and x509 which was rather broken and didn't support
18142 BMPStrings or UTF8Strings. Since some software doesn't implement
18143 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18144 using the dirstring_type option. See the new comment in the default
18145 openssl.cnf for more info.
18149 * Make crypto/rand/md_rand.c more robust:
18150 - Assure unique random numbers after fork().
18151 - Make sure that concurrent threads access the global counter and
18152 md serializably so that we never lose entropy in them
18153 or use exactly the same state in multiple threads.
18154 Access to the large state is not always serializable because
18155 the additional locking could be a performance killer, and
18156 md should be large enough anyway.
18160 * New file `apps/app_rand.c` with commonly needed functionality
18161 for handling the random seed file.
18163 Use the random seed file in some applications that previously did not:
18165 dsaparam -genkey (which also ignored its '-rand' option),
18168 x509 (when signing).
18169 Except on systems with /dev/urandom, it is crucial to have a random
18170 seed file at least for key creation, DSA signing, and for DH exchanges;
18171 for RSA signatures we could do without one.
18173 gendh and gendsa (unlike genrsa) used to read only the first byte
18174 of each file listed in the '-rand' option. The function as previously
18175 found in genrsa is now in app_rand.c and is used by all programs
18176 that support '-rand'.
18180 * In RAND_write_file, use mode 0600 for creating files;
18181 don't just chmod when it may be too late.
18185 * Report an error from X509_STORE_load_locations
18186 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18190 * New function ASN1_mbstring_copy() this copies a string in either
18191 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18192 into an ASN1_STRING type. A mask of permissible types is passed
18193 and it chooses the "minimal" type to use or an error if not type
18198 * Add function equivalents to the various macros in asn1.h. The old
18199 macros are retained with an `M_` prefix. Code inside the library can
18200 use the `M_` macros. External code (including the openssl utility)
18201 should *NOT* in order to be "shared library friendly".
18205 * Add various functions that can check a certificate's extensions
18206 to see if it usable for various purposes such as SSL client,
18207 server or S/MIME and CAs of these types. This is currently
18208 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18209 verification. Also added a -purpose flag to x509 utility to
18210 print out all the purposes.
18214 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18219 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18220 for, obtain and decode and extension and obtain its critical flag.
18221 This allows all the necessary extension code to be handled in a
18222 single function call.
18226 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18227 platforms. See crypto/rc4/rc4_enc.c for further details.
18231 * New -noout option to asn1parse. This causes no output to be produced
18232 its main use is when combined with -strparse and -out to extract data
18233 from a file (which may not be in ASN.1 format).
18237 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18238 when producing the local key id.
18240 *Richard Levitte <levitte@stacken.kth.se>*
18242 * New option -dhparam in s_server. This allows a DH parameter file to be
18243 stated explicitly. If it is not stated then it tries the first server
18244 certificate file. The previous behaviour hard coded the filename
18249 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18250 a public key to be input or output. For example:
18251 openssl rsa -in key.pem -pubout -out pubkey.pem
18252 Also added necessary DSA public key functions to handle this.
18256 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18257 in the message. This was handled by allowing
18258 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18260 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18262 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18263 to the end of the strings whereas this didn't. This would cause problems
18264 if strings read with d2i_ASN1_bytes() were later modified.
18266 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18268 * Fix for base64 decode bug. When a base64 bio reads only one line of
18269 data and it contains EOF it will end up returning an error. This is
18270 caused by input 46 bytes long. The cause is due to the way base64
18271 BIOs find the start of base64 encoded data. They do this by trying a
18272 trial decode on each line until they find one that works. When they
18273 do a flag is set and it starts again knowing it can pass all the
18274 data directly through the decoder. Unfortunately it doesn't reset
18275 the context it uses. This means that if EOF is reached an attempt
18276 is made to pass two EOFs through the context and this causes the
18277 resulting error. This can also cause other problems as well. As is
18278 usual with these problems it takes *ages* to find and the fix is
18279 trivial: move one line.
18281 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18283 * Ugly workaround to get s_client and s_server working under Windows. The
18284 old code wouldn't work because it needed to select() on sockets and the
18285 tty (for keypresses and to see if data could be written). Win32 only
18286 supports select() on sockets so we select() with a 1s timeout on the
18287 sockets and then see if any characters are waiting to be read, if none
18288 are present then we retry, we also assume we can always write data to
18289 the tty. This isn't nice because the code then blocks until we've
18290 received a complete line of data and it is effectively polling the
18291 keyboard at 1s intervals: however it's quite a bit better than not
18292 working at all :-) A dedicated Windows application might handle this
18293 with an event loop for example.
18297 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18298 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18299 will be called when RSA_sign() and RSA_verify() are used. This is useful
18300 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18301 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18302 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18303 This necessitated the support of an extra signature type NID_md5_sha1
18304 for SSL signatures and modifications to the SSL library to use it instead
18305 of calling RSA_public_decrypt() and RSA_private_encrypt().
18309 * Add new -verify -CAfile and -CApath options to the crl program, these
18310 will lookup a CRL issuers certificate and verify the signature in a
18311 similar way to the verify program. Tidy up the crl program so it
18312 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18313 less strict. It will now permit CRL extensions even if it is not
18314 a V2 CRL: this will allow it to tolerate some broken CRLs.
18318 * Initialize all non-automatic variables each time one of the openssl
18319 sub-programs is started (this is necessary as they may be started
18320 multiple times from the "OpenSSL>" prompt).
18322 *Lennart Bang, Bodo Moeller*
18324 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18325 removing all other RSA functionality (this is what NO_RSA does). This
18326 is so (for example) those in the US can disable those operations covered
18327 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18332 * Non-copying interface to BIO pairs.
18333 (still largely untested)
18337 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18338 ASCII string. This was handled independently in various places before.
18342 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18343 UTF8 strings a character at a time.
18347 * Use client_version from client hello to select the protocol
18348 (s23_srvr.c) and for RSA client key exchange verification
18349 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18353 * Add various utility functions to handle SPKACs, these were previously
18354 handled by poking round in the structure internals. Added new function
18355 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18356 print, verify and generate SPKACs. Based on an original idea from
18357 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18361 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18365 * Allow the config file extension section to be overwritten on the
18366 command line. Based on an original idea from Massimiliano Pala
18367 <madwolf@comune.modena.it>. The new option is called -extensions
18368 and can be applied to ca, req and x509. Also -reqexts to override
18369 the request extensions in req and -crlexts to override the crl extensions
18374 * Add new feature to the SPKAC handling in ca. Now you can include
18375 the same field multiple times by preceding it by "XXXX." for example:
18378 this is the same syntax as used in the req config file.
18382 * Allow certificate extensions to be added to certificate requests. These
18383 are specified in a 'req_extensions' option of the req section of the
18384 config file. They can be printed out with the -text option to req but
18385 are otherwise ignored at present.
18389 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18390 data read consists of only the final block it would not decrypted because
18391 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18392 A misplaced 'break' also meant the decrypted final block might not be
18393 copied until the next read.
18397 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18398 a few extra parameters to the DH structure: these will be useful if
18399 for example we want the value of 'q' or implement X9.42 DH.
18403 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18404 provides hooks that allow the default DSA functions or functions on a
18405 "per key" basis to be replaced. This allows hardware acceleration and
18406 hardware key storage to be handled without major modification to the
18407 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18408 associated functions.
18412 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18413 as "read only": it can't be written to and the buffer it points to will
18414 not be freed. Reading from a read only BIO is much more efficient than
18415 a normal memory BIO. This was added because there are several times when
18416 an area of memory needs to be read from a BIO. The previous method was
18417 to create a memory BIO and write the data to it, this results in two
18418 copies of the data and an O(n^2) reading algorithm. There is a new
18419 function BIO_new_mem_buf() which creates a read only memory BIO from
18420 an area of memory. Also modified the PKCS#7 routines to use read only
18425 * Bugfix: ssl23_get_client_hello did not work properly when called in
18426 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18427 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18428 but a retry condition occurred while trying to read the rest.
18432 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18433 NID_pkcs7_encrypted by default: this was wrong since this should almost
18434 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18435 the encrypted data type: this is a more sensible place to put it and it
18436 allows the PKCS#12 code to be tidied up that duplicated this
18441 * Changed obj_dat.pl script so it takes its input and output files on
18442 the command line. This should avoid shell escape redirection problems
18447 * Initial support for certificate extension requests, these are included
18448 in things like Xenroll certificate requests. Included functions to allow
18449 extensions to be obtained and added.
18453 * -crlf option to s_client and s_server for sending newlines as
18454 CRLF (as required by many protocols).
18458 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18460 * Install libRSAglue.a when OpenSSL is built with RSAref.
18462 *Ralf S. Engelschall*
18464 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18466 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18468 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18473 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18474 DH parameters/keys (q is lost during that conversion, but the resulting
18475 DH parameters contain its length).
18477 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18478 much faster than DH_generate_parameters (which creates parameters
18479 where `p = 2*q + 1`), and also the smaller q makes DH computations
18480 much more efficient (160-bit exponentiation instead of 1024-bit
18481 exponentiation); so this provides a convenient way to support DHE
18482 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18483 utter importance to use
18484 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18486 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18487 when such DH parameters are used, because otherwise small subgroup
18488 attacks may become possible!
18492 * Avoid memory leak in i2d_DHparams.
18496 * Allow the -k option to be used more than once in the enc program:
18497 this allows the same encrypted message to be read by multiple recipients.
18501 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18502 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18503 it will always use the numerical form of the OID, even if it has a short
18508 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18509 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18510 otherwise bn_mod_exp was called. In the case of hardware keys for example
18511 no private key components need be present and it might store extra data
18512 in the RSA structure, which cannot be accessed from bn_mod_exp.
18513 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18514 private key operations.
18518 * Added support for SPARC Linux.
18522 * pem_password_cb function type incompatibly changed from
18523 typedef int pem_password_cb(char *buf, int size, int rwflag);
18525 ....(char *buf, int size, int rwflag, void *userdata);
18526 so that applications can pass data to their callbacks:
18527 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18528 additional void * argument, which is just handed through whenever
18529 the password callback is called.
18531 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18533 New function SSL_CTX_set_default_passwd_cb_userdata.
18535 Compatibility note: As many C implementations push function arguments
18536 onto the stack in reverse order, the new library version is likely to
18537 interoperate with programs that have been compiled with the old
18538 pem_password_cb definition (PEM_whatever takes some data that
18539 happens to be on the stack as its last argument, and the callback
18540 just ignores this garbage); but there is no guarantee whatsoever that
18543 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18544 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18545 problems not only on Windows, but also on some Unix platforms.
18546 To avoid problematic command lines, these definitions are now in an
18547 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18548 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18552 * MIPS III/IV assembler module is reimplemented.
18556 * More DES library cleanups: remove references to srand/rand and
18557 delete an unused file.
18561 * Add support for the free Netwide assembler (NASM) under Win32,
18562 since not many people have MASM (ml) and it can be hard to obtain.
18563 This is currently experimental but it seems to work OK and pass all
18564 the tests. Check out INSTALL.W32 for info.
18568 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18569 without temporary keys kept an extra copy of the server key,
18570 and connections with temporary keys did not free everything in case
18575 * New function RSA_check_key and new openssl rsa option -check
18576 for verifying the consistency of RSA keys.
18578 *Ulf Moeller, Bodo Moeller*
18580 * Various changes to make Win32 compile work:
18581 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18582 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18583 comparison" warnings.
18584 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18588 * Add a debugging option to PKCS#5 v2 key generation function: when
18589 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18590 derived keys are printed to stderr.
18594 * Copy the flags in ASN1_STRING_dup().
18596 *Roman E. Pavlov <pre@mo.msk.ru>*
18598 * The x509 application mishandled signing requests containing DSA
18599 keys when the signing key was also DSA and the parameters didn't match.
18601 It was supposed to omit the parameters when they matched the signing key:
18602 the verifying software was then supposed to automatically use the CA's
18603 parameters if they were absent from the end user certificate.
18605 Omitting parameters is no longer recommended. The test was also
18606 the wrong way round! This was probably due to unusual behaviour in
18607 EVP_cmp_parameters() which returns 1 if the parameters match.
18608 This meant that parameters were omitted when they *didn't* match and
18609 the certificate was useless. Certificates signed with 'ca' didn't have
18612 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18614 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18615 The interface is as follows:
18616 Applications can use
18617 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18618 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18619 "off" is now the default.
18620 The library internally uses
18621 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18622 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18623 to disable memory-checking temporarily.
18625 Some inconsistent states that previously were possible (and were
18626 even the default) are now avoided.
18628 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18629 with each memory chunk allocated; this is occasionally more helpful
18630 than just having a counter.
18632 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18634 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18639 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18640 which largely parallels "options", but is for changing API behaviour,
18641 whereas "options" are about protocol behaviour.
18642 Initial "mode" flags are:
18644 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18645 a single record has been written.
18646 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18647 retries use the same buffer location.
18648 (But all of the contents must be
18653 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18656 * Fix problems with no-hmac etc.
18658 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18660 * New functions RSA_get_default_method(), RSA_set_method() and
18661 RSA_get_method(). These allows replacement of RSA_METHODs without having
18662 to mess around with the internals of an RSA structure.
18666 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18667 Also really enable memory leak checks in openssl.c and in some
18670 *Chad C. Mulligan, Bodo Moeller*
18672 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18673 up the length of negative integers. This has now been simplified to just
18674 store the length when it is first determined and use it later, rather
18675 than trying to keep track of where data is copied and updating it to
18677 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18679 * Add a new function PKCS7_signatureVerify. This allows the verification
18680 of a PKCS#7 signature but with the signing certificate passed to the
18681 function itself. This contrasts with PKCS7_dataVerify which assumes the
18682 certificate is present in the PKCS#7 structure. This isn't always the
18683 case: certificates can be omitted from a PKCS#7 structure and be
18684 distributed by "out of band" means (such as a certificate database).
18688 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18689 function prototypes in pem.h, also change util/mkdef.pl to add the
18690 necessary function names.
18694 * mk1mf.pl (used by Windows builds) did not properly read the
18695 options set by Configure in the top level Makefile, and Configure
18696 was not even able to write more than one option correctly.
18697 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18701 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18702 file to be loaded from a BIO or FILE pointer. The BIO version will
18703 for example allow memory BIOs to contain config info.
18707 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18708 Whoever hopes to achieve shared-library compatibility across versions
18709 must use this, not the compile-time macro.
18710 (Exercise 0.9.4: Which is the minimum library version required by
18712 Note: All this applies only to multi-threaded programs, others don't
18717 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18718 through a BIO pair triggered the default case, i.e.
18719 SSLerr(...,SSL_R_UNKNOWN_STATE).
18723 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18724 can use the SSL library even if none of the specific BIOs is
18729 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18730 for the encoded length.
18732 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18734 * Add initial documentation of the X509V3 functions.
18738 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18739 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18740 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18741 secure PKCS#8 private key format with a high iteration count.
18745 * Fix determination of Perl interpreter: A perl or perl5
18746 *directory* in $PATH was also accepted as the interpreter.
18748 *Ralf S. Engelschall*
18750 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18751 wrong with it but it was very old and did things like calling
18752 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18753 unusual formatting.
18757 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18758 to use the new extension code.
18762 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18763 with macros. This should make it easier to change their form, add extra
18764 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18769 * Add to configuration table a new entry that can specify an alternative
18770 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18771 according to Mark Crispin <MRC@Panda.COM>.
18775 * DES CBC did not update the IV. Weird.
18779 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18780 Changing the behaviour of the former might break existing programs --
18781 where IV updating is needed, des_ncbc_encrypt can be used.
18784 * When bntest is run from "make test" it drives bc to check its
18785 calculations, as well as internally checking them. If an internal check
18786 fails, it needs to cause bc to give a non-zero result or make test carries
18787 on without noticing the failure. Fixed.
18791 * DES library cleanups.
18795 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18796 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18797 ciphers. NOTE: although the key derivation function has been verified
18798 against some published test vectors it has not been extensively tested
18799 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18804 * Instead of "mkdir -p", which is not fully portable, use new
18805 Perl script "util/mkdir-p.pl".
18809 * Rewrite the way password based encryption (PBE) is handled. It used to
18810 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18811 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18812 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18813 the 'parameter' field of the AlgorithmIdentifier is passed to the
18814 underlying key generation function so it must do its own ASN1 parsing.
18815 This has also changed the EVP_PBE_CipherInit() function which now has a
18816 'parameter' argument instead of literal salt and iteration count values
18817 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18821 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18822 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18823 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18824 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18825 value was just used as a "magic string" and not used directly its
18826 value doesn't matter.
18830 * Introduce some semblance of const correctness to BN. Shame C doesn't
18835 * "linux-sparc64" configuration (ultrapenguin).
18837 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18838 "linux-sparc" configuration.
18840 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18842 * config now generates no-xxx options for missing ciphers.
18846 * Support the EBCDIC character set (work in progress).
18847 File ebcdic.c not yet included because it has a different license.
18849 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18851 * Support BS2000/OSD-POSIX.
18853 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18855 * Make callbacks for key generation use `void *` instead of `char *`.
18859 * Make S/MIME samples compile (not yet tested).
18863 * Additional typesafe stacks.
18867 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18871 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18873 * New configuration variant "sco5-gcc".
18875 * Updated some demos.
18877 *Sean O Riordain, Wade Scholine*
18879 * Add missing BIO_free at exit of pkcs12 application.
18883 * Fix memory leak in conf.c.
18887 * Updates for Win32 to assembler version of MD5.
18891 * Set #! path to perl in `apps/der_chop` to where we found it
18892 instead of using a fixed path.
18896 * SHA library changes for irix64-mips4-cc.
18900 * Improvements for VMS support.
18904 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18906 * Bignum library bug fix. IRIX 6 passes "make test" now!
18907 This also avoids the problems with SC4.2 and unpatched SC5.
18909 *Andy Polyakov <appro@fy.chalmers.se>*
18911 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18912 These are required because of the typesafe stack would otherwise break
18913 existing code. If old code used a structure member which used to be STACK
18914 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18915 sk_num or sk_value it would produce an error because the num, data members
18916 are not present in STACK_OF. Now it just produces a warning. sk_set
18917 replaces the old method of assigning a value to sk_value
18918 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18919 that does this will no longer work (and should use sk_set instead) but
18920 this could be regarded as a "questionable" behaviour anyway.
18924 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18925 correctly handle encrypted S/MIME data.
18929 * Change type of various DES function arguments from des_cblock
18930 (which means, in function argument declarations, pointer to char)
18931 to des_cblock * (meaning pointer to array with 8 char elements),
18932 which allows the compiler to do more typechecking; it was like
18933 that back in SSLeay, but with lots of ugly casts.
18935 Introduce new type const_des_cblock.
18939 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18940 problems: find RecipientInfo structure that matches recipient certificate
18941 and initialise the ASN1 structures properly based on passed cipher.
18945 * Belatedly make the BN tests actually check the results.
18949 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18950 to and from BNs: it was completely broken. New compilation option
18951 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18952 key elements as negative integers.
18956 * Reorganize and speed up MD5.
18958 *Andy Polyakov <appro@fy.chalmers.se>*
18962 *Richard Levitte <richard@levitte.org>*
18964 * New option -out to asn1parse to allow the parsed structure to be
18965 output to a file. This is most useful when combined with the -strparse
18966 option to examine the output of things like OCTET STRINGS.
18970 * Make SSL library a little more fool-proof by not requiring any longer
18971 that `SSL_set_{accept,connect}_state` be called before
18972 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18973 in many applications because usually everything *appeared* to work as
18974 intended anyway -- now it really works as intended).
18978 * Move openssl.cnf out of lib/.
18982 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18983 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18984 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18986 *Ralf S. Engelschall*
18988 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18989 handle PKCS#7 enveloped data properly.
18991 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18993 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18994 copying pointers. The cert_st handling is changed by this in
18995 various ways (and thus what used to be known as ctx->default_cert
18996 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18997 any longer when s->cert does not give us what we need).
18998 ssl_cert_instantiate becomes obsolete by this change.
18999 As soon as we've got the new code right (possibly it already is?),
19000 we have solved a couple of bugs of the earlier code where s->cert
19001 was used as if it could not have been shared with other SSL structures.
19003 Note that using the SSL API in certain dirty ways now will result
19004 in different behaviour than observed with earlier library versions:
19005 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19006 does not influence s as it used to.
19008 In order to clean up things more thoroughly, inside SSL_SESSION
19009 we don't use CERT any longer, but a new structure SESS_CERT
19010 that holds per-session data (if available); currently, this is
19011 the peer's certificate chain and, for clients, the server's certificate
19012 and temporary key. CERT holds only those values that can have
19013 meaningful defaults in an SSL_CTX.
19017 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19018 from the internal representation. Various PKCS#7 fixes: remove some
19019 evil casts and set the enc_dig_alg field properly based on the signing
19024 * Allow PKCS#12 password to be set from the command line or the
19025 environment. Let 'ca' get its config file name from the environment
19026 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19031 * Allow certificate policies extension to use an IA5STRING for the
19032 organization field. This is contrary to the PKIX definition but
19033 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19038 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19039 without disallowing inline assembler and the like for non-pedantic builds.
19043 * Support Borland C++ builder.
19045 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19051 * SHA-1 cleanups and performance enhancements.
19053 *Andy Polyakov <appro@fy.chalmers.se>*
19055 * Sparc v8plus assembler for the bignum library.
19057 *Andy Polyakov <appro@fy.chalmers.se>*
19059 * Accept any -xxx and +xxx compiler options in Configure.
19063 * Update HPUX configuration.
19067 * Add missing `sk_<type>_unshift()` function to safestack.h
19069 *Ralf S. Engelschall*
19071 * New function SSL_CTX_use_certificate_chain_file that sets the
19072 "extra_cert"s in addition to the certificate. (This makes sense
19073 only for "PEM" format files, as chains as a whole are not
19078 * Support verify_depth from the SSL API.
19079 x509_vfy.c had what can be considered an off-by-one-error:
19080 Its depth (which was not part of the external interface)
19081 was actually counting the number of certificates in a chain;
19082 now it really counts the depth.
19086 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19087 instead of X509err, which often resulted in confusing error
19088 messages since the error codes are not globally unique
19089 (e.g. an alleged error in ssl3_accept when a certificate
19090 didn't match the private key).
19092 * New function SSL_CTX_set_session_id_context that allows to set a default
19093 value (so that you don't need SSL_set_session_id_context for each
19094 connection using the SSL_CTX).
19098 * OAEP decoding bug fix.
19102 * Support INSTALL_PREFIX for package builders, as proposed by
19107 * New Configure options "threads" and "no-threads". For systems
19108 where the proper compiler options are known (currently Solaris
19109 and Linux), "threads" is the default.
19113 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19117 * Install various scripts to $(OPENSSLDIR)/misc, not to
19118 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19119 such as /usr/local/bin.
19123 * "make linux-shared" to build shared libraries.
19125 *Niels Poppe <niels@netbox.org>*
19127 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19131 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19132 extension adding in x509 utility.
19136 * Remove NOPROTO sections and error code comments.
19140 * Partial rewrite of the DEF file generator to now parse the ANSI
19145 * New Configure options --prefix=DIR and --openssldir=DIR.
19149 * Complete rewrite of the error code script(s). It is all now handled
19150 by one script at the top level which handles error code gathering,
19151 header rewriting and C source file generation. It should be much better
19152 than the old method: it now uses a modified version of Ulf's parser to
19153 read the ANSI prototypes in all header files (thus the old K&R definitions
19154 aren't needed for error creation any more) and do a better job of
19155 translating function codes into names. The old 'ASN1 error code embedded
19156 in a comment' is no longer necessary and it doesn't use .err files which
19157 have now been deleted. Also the error code call doesn't have to appear all
19158 on one line (which resulted in some large lines...).
19162 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19166 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19167 0 (which usually indicates a closed connection), but continue reading.
19171 * Fix some race conditions.
19175 * Add support for CRL distribution points extension. Add Certificate
19176 Policies and CRL distribution points documentation.
19180 * Move the autogenerated header file parts to crypto/opensslconf.h.
19184 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19185 8 of keying material. Merlin has also confirmed interop with this fix
19186 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19188 *Merlin Hughes <merlin@baltimore.ie>*
19190 * Fix lots of warnings.
19192 *Richard Levitte <levitte@stacken.kth.se>*
19194 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19195 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19197 *Richard Levitte <levitte@stacken.kth.se>*
19199 * Fix problems with sizeof(long) == 8.
19201 *Andy Polyakov <appro@fy.chalmers.se>*
19203 * Change functions to ANSI C.
19207 * Fix typos in error codes.
19209 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19211 * Remove defunct assembler files from Configure.
19215 * SPARC v8 assembler BIGNUM implementation.
19217 *Andy Polyakov <appro@fy.chalmers.se>*
19219 * Support for Certificate Policies extension: both print and set.
19220 Various additions to support the r2i method this uses.
19224 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19225 return a const string when you are expecting an allocated buffer.
19229 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19230 types DirectoryString and DisplayText.
19234 * Add code to allow r2i extensions to access the configuration database,
19235 add an LHASH database driver and add several ctx helper functions.
19239 * Fix an evil bug in bn_expand2() which caused various BN functions to
19240 fail when they extended the size of a BIGNUM.
19244 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19245 support typesafe stack.
19249 * Fix typo in SSL_[gs]et_options().
19251 *Nils Frostberg <nils@medcom.se>*
19253 * Delete various functions and files that belonged to the (now obsolete)
19254 old X509V3 handling code.
19258 * New Configure option "rsaref".
19262 * Don't auto-generate pem.h.
19266 * Introduce type-safe ASN.1 SETs.
19270 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19272 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19274 * Introduce type-safe STACKs. This will almost certainly break lots of code
19275 that links with OpenSSL (well at least cause lots of warnings), but fear
19276 not: the conversion is trivial, and it eliminates loads of evil casts. A
19277 few STACKed things have been converted already. Feel free to convert more.
19278 In the fullness of time, I'll do away with the STACK type altogether.
19282 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19283 specified in `<certfile>` by updating the entry in the index.txt file.
19284 This way one no longer has to edit the index.txt file manually for
19285 revoking a certificate. The -revoke option does the gory details now.
19287 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19289 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19290 `-text` option at all and this way the `-noout -text` combination was
19291 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19293 *Ralf S. Engelschall*
19295 * Make sure a corresponding plain text error message exists for the
19296 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19297 verify callback function determined that a certificate was revoked.
19299 *Ralf S. Engelschall*
19301 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19302 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19303 all available ciphers including rc5, which was forgotten until now.
19304 In order to let the testing shell script know which algorithms
19305 are available, a new (up to now undocumented) command
19306 `openssl list-cipher-commands` is used.
19310 * Bugfix: s_client occasionally would sleep in select() when
19311 it should have checked SSL_pending() first.
19315 * New functions DSA_do_sign and DSA_do_verify to provide access to
19316 the raw DSA values prior to ASN.1 encoding.
19320 * Tweaks to Configure
19322 *Niels Poppe <niels@netbox.org>*
19324 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19329 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19333 * New config option to avoid instructions that are illegal on the 80386.
19334 The default code is faster, but requires at least a 486.
19338 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19339 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19340 same as SSL2_VERSION anyway.
19344 * New "-showcerts" option for s_client.
19348 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19349 application. Various cleanups and fixes.
19353 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19354 modify error routines to work internally. Add error codes and PBE init
19355 to library startup routines.
19359 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19360 packing functions to asn1 and evp. Changed function names and error
19361 codes along the way.
19365 * PKCS12 integration: and so it begins... First of several patches to
19366 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19367 objects to objects.h
19371 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19372 and display support for Thawte strong extranet extension.
19376 * Add LinuxPPC support.
19378 *Jeff Dubrule <igor@pobox.org>*
19380 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19381 bn_div_words in alpha.s.
19383 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19385 * Make sure the RSA OAEP test is skipped under -DRSAref because
19386 OAEP isn't supported when OpenSSL is built with RSAref.
19388 *Ulf Moeller <ulf@fitug.de>*
19390 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19391 so they no longer are missing under -DNOPROTO.
19393 *Soren S. Jorvang <soren@t.dk>*
19395 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19397 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19398 doesn't work when the session is reused. Coming soon!
19402 * Fix a security hole, that allows sessions to be reused in the wrong
19403 context thus bypassing client cert protection! All software that uses
19404 client certs and session caches in multiple contexts NEEDS PATCHING to
19405 allow session reuse! A fuller solution is in the works.
19407 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19409 * Some more source tree cleanups (removed obsolete files
19410 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19411 permission on "config" script to be executable) and a fix for the INSTALL
19414 *Ulf Moeller <ulf@fitug.de>*
19416 * Remove some legacy and erroneous uses of malloc, free instead of
19419 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19421 * Make rsa_oaep_test return non-zero on error.
19423 *Ulf Moeller <ulf@fitug.de>*
19425 * Add support for native Solaris shared libraries. Configure
19426 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19427 if someone would make that last step automatic.
19429 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19431 * ctx_size was not built with the right compiler during "make links". Fixed.
19435 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19436 except NULL ciphers". This means the default cipher list will no longer
19437 enable NULL ciphers. They need to be specifically enabled e.g. with
19438 the string "DEFAULT:eNULL".
19442 * Fix to RSA private encryption routines: if p < q then it would
19443 occasionally produce an invalid result. This will only happen with
19444 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19448 * Be less restrictive and allow also `perl util/perlpath.pl
19449 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19450 because this way one can also use an interpreter named `perl5` (which is
19451 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19452 installed as `perl`).
19454 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19456 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19458 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19460 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19461 advapi32.lib to Win32 build and change the pem test comparison
19462 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19463 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19464 and crypto/des/ede_cbcm_enc.c.
19468 * DES quad checksum was broken on big-endian architectures. Fixed.
19472 * Comment out two functions in bio.h that aren't implemented. Fix up the
19473 Win32 test batch file so it (might) work again. The Win32 test batch file
19474 is horrible: I feel ill....
19478 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19479 in e_os.h. Audit of header files to check ANSI and non ANSI
19480 sections: 10 functions were absent from non ANSI section and not exported
19481 from Windows DLLs. Fixed up libeay.num for new functions.
19485 * Make `openssl version` output lines consistent.
19487 *Ralf S. Engelschall*
19489 * Fix Win32 symbol export lists for BIO functions: Added
19490 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19491 to ms/libeay{16,32}.def.
19493 *Ralf S. Engelschall*
19495 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19496 fine under Unix and passes some trivial tests I've now added. But the
19497 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19498 added to make sure no one expects that this stuff really works in the
19499 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19500 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19503 *Ralf S. Engelschall*
19505 * Fix the generation of two part addresses in perl.
19507 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19509 * Add config entry for Linux on MIPS.
19511 *John Tobey <jtobey@channel1.com>*
19513 * Make links whenever Configure is run, unless we are on Windoze.
19517 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19518 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19523 * Add a useful kludge to allow package maintainers to specify compiler and
19524 other platforms details on the command line without having to patch the
19525 Configure script every time: One now can use
19526 `perl Configure <id>:<details>`,
19527 i.e. platform ids are allowed to have details appended
19528 to them (separated by colons). This is treated as there would be a static
19529 pre-configured entry in Configure's %table under key `<id>` with value
19530 `<details>` and `perl Configure <id>` is called. So, when you want to
19531 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19532 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19533 now, which overrides the FreeBSD-elf entry on-the-fly.
19535 *Ralf S. Engelschall*
19537 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19541 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19542 on the `perl Configure ...` command line. This way one can compile
19543 OpenSSL libraries with Position Independent Code (PIC) which is needed
19544 for linking it into DSOs.
19546 *Ralf S. Engelschall*
19548 * Remarkably, export ciphers were totally broken and no-one had noticed!
19553 * Cleaned up the LICENSE document: The official contact for any license
19554 questions now is the OpenSSL core team under openssl-core@openssl.org.
19555 And add a paragraph about the dual-license situation to make sure people
19556 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19557 to the OpenSSL toolkit.
19559 *Ralf S. Engelschall*
19561 * General source tree makefile cleanups: Made `making xxx in yyy...`
19562 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19563 Additionally cleaned up the `make links` target: Remove unnecessary
19564 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19565 to speed processing and no longer clutter the display with confusing
19566 stuff. Instead only the actually done links are displayed.
19568 *Ralf S. Engelschall*
19570 * Permit null encryption ciphersuites, used for authentication only. It used
19571 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19572 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19577 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19578 signed attributes when verifying signatures (this would break them),
19579 the detached data encoding was wrong and public keys obtained using
19580 X509_get_pubkey() weren't freed.
19584 * Add text documentation for the BUFFER functions. Also added a work around
19585 to a Win95 console bug. This was triggered by the password read stuff: the
19586 last character typed gets carried over to the next fread(). If you were
19587 generating a new cert request using 'req' for example then the last
19588 character of the passphrase would be CR which would then enter the first
19593 * Added the new 'Includes OpenSSL Cryptography Software' button as
19594 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19595 button and can be used by applications based on OpenSSL to show the
19596 relationship to the OpenSSL project.
19598 *Ralf S. Engelschall*
19600 * Remove confusing variables in function signatures in files
19601 ssl/ssl_lib.c and ssl/ssl.h.
19603 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19605 * Don't install bss_file.c under PREFIX/include/
19607 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19609 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19610 functions that return function pointers and has support for NT specific
19611 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19612 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19613 unsigned to signed types: this was killing the Win32 compile.
19617 * Add new certificate file to stack functions,
19618 SSL_add_dir_cert_subjects_to_stack() and
19619 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19620 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19621 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19622 This means that Apache-SSL and similar packages don't have to mess around
19623 to add as many CAs as they want to the preferred list.
19627 * Experiment with doxygen documentation. Currently only partially applied to
19629 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19630 openssl.doxy as the configuration file.
19634 * Get rid of remaining C++-style comments which strict C compilers hate.
19636 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19638 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19639 compiled in by default: it has problems with large keys.
19643 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19644 DH private keys and/or callback functions which directly correspond to
19645 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19646 is needed for applications which have to configure certificates on a
19647 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19649 For the RSA certificate situation is makes no difference, but
19650 for the DSA certificate situation this fixes the "no shared cipher"
19651 problem where the OpenSSL cipher selection procedure failed because the
19652 temporary keys were not overtaken from the context and the API provided
19653 no way to reconfigure them.
19654 The new functions now let applications reconfigure the stuff and they
19655 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19656 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19657 non-public-API function ssl_cert_instantiate() is used as a helper
19658 function and also to reduce code redundancy inside ssl_rsa.c.
19660 *Ralf S. Engelschall*
19662 * Move s_server -dcert and -dkey options out of the undocumented feature
19663 area because they are useful for the DSA situation and should be
19664 recognized by the users.
19666 *Ralf S. Engelschall*
19668 * Fix the cipher decision scheme for export ciphers: the export bits are
19669 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19670 SSL_EXP_MASK. So, the original variable has to be used instead of the
19671 already masked variable.
19673 *Richard Levitte <levitte@stacken.kth.se>*
19675 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19677 *Richard Levitte <levitte@stacken.kth.se>*
19679 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19680 from `int` to `unsigned int` because it is a length and initialized by
19681 EVP_DigestFinal() which expects an `unsigned int *`.
19683 *Richard Levitte <levitte@stacken.kth.se>*
19685 * Don't hard-code path to Perl interpreter on shebang line of Configure
19686 script. Instead use the usual Shell->Perl transition trick.
19688 *Ralf S. Engelschall*
19690 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19691 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19692 -noout -modulus` as it's already the case for `openssl rsa -noout
19693 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19694 currently the public key is printed (a decision which was already done by
19695 `openssl dsa -modulus` in the past) which serves a similar purpose.
19696 Additionally the NO_RSA no longer completely removes the whole -modulus
19697 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19700 *Ralf S. Engelschall*
19702 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19703 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19705 *Arne Ansper <arne@ats.cyber.ee>*
19707 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19708 to be added. Now both 'req' and 'ca' can use new objects defined in the
19713 * Add cool BIO that does syslog (or event log on NT).
19715 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19717 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19718 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19719 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19720 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19724 * Add preliminary config info for new extension code.
19728 * Make RSA_NO_PADDING really use no padding.
19730 *Ulf Moeller <ulf@fitug.de>*
19732 * Generate errors when private/public key check is done.
19736 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19737 for some CRL extensions and new objects added.
19741 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19742 key usage extension and fuller support for authority key id.
19746 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19747 padding method for RSA, which is recommended for new applications in PKCS
19748 #1 v2.0 (RFC 2437, October 1998).
19749 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19750 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19751 against Bleichbacher's attack on RSA.
19752 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19755 * Updates to the new SSL compression code
19757 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19759 * Fix so that the version number in the master secret, when passed
19760 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19761 (because the server will not accept higher), that the version number
19762 is 0x03,0x01, not 0x03,0x00
19764 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19766 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19767 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19768 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19772 * Support for RAW extensions where an arbitrary extension can be
19773 created by including its DER encoding. See `apps/openssl.cnf` for
19778 * Make sure latest Perl versions don't interpret some generated C array
19779 code as Perl array code in the crypto/err/err_genc.pl script.
19781 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19783 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19784 not many people have the assembler. Various Win32 compilation fixes and
19785 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19786 build instructions.
19790 * Modify configure script 'Configure' to automatically create crypto/date.h
19791 file under Win32 and also build pem.h from pem.org. New script
19792 util/mkfiles.pl to create the MINFO file on environments that can't do a
19793 'make files': perl util/mkfiles.pl >MINFO should work.
19797 * Major rework of DES function declarations, in the pursuit of correctness
19798 and purity. As a result, many evil casts evaporated, and some weirdness,
19799 too. You may find this causes warnings in your code. Zapping your evil
19800 casts will probably fix them. Mostly.
19804 * Fix for a typo in asn1.h. Bug fix to object creation script
19805 obj_dat.pl. It considered a zero in an object definition to mean
19806 "end of object": none of the objects in objects.h have any zeros
19807 so it wasn't spotted.
19809 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19811 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19812 Masking (CBCM). In the absence of test vectors, the best I have been able
19813 to do is check that the decrypt undoes the encrypt, so far. Send me test
19814 vectors if you have them.
19818 * Correct calculation of key length for export ciphers (too much space was
19819 allocated for null ciphers). This has not been tested!
19823 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19824 message is now correct (it understands "crypto" and "ssl" on its
19825 command line). There is also now an "update" option. This will update
19826 the util/ssleay.num and util/libeay.num files with any new functions.
19828 perl util/mkdef.pl crypto ssl update
19829 it will update them.
19833 * Overhauled the Perl interface:
19834 - ported BN stuff to OpenSSL's different BN library
19835 - made the perl/ source tree CVS-aware
19836 - renamed the package from SSLeay to OpenSSL (the files still contain
19837 their history because I've copied them in the repository)
19838 - removed obsolete files (the test scripts will be replaced
19839 by better Test::Harness variants in the future)
19841 *Ralf S. Engelschall*
19843 * First cut for a very conservative source tree cleanup:
19844 1. merge various obsolete readme texts into doc/ssleay.txt
19845 where we collect the old documents and readme texts.
19846 2. remove the first part of files where I'm already sure that we no
19847 longer need them because of three reasons: either they are just temporary
19848 files which were left by Eric or they are preserved original files where
19849 I've verified that the diff is also available in the CVS via "cvs diff
19850 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19851 the crypto/md/ stuff).
19853 *Ralf S. Engelschall*
19855 * More extension code. Incomplete support for subject and issuer alt
19856 name, issuer and authority key id. Change the i2v function parameters
19857 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19858 what that's for :-) Fix to ASN1 macro which messed up
19859 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19863 * Preliminary support for ENUMERATED type. This is largely copied from the
19868 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19870 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19872 * Make sure `make rehash` target really finds the `openssl` program.
19874 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19876 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19877 like to hear about it if this slows down other processors.
19881 * Add CygWin32 platform information to Configure script.
19883 *Alan Batie <batie@aahz.jf.intel.com>*
19885 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19887 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19889 * New program nseq to manipulate netscape certificate sequences
19893 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19898 * Fixes to BN code. Previously the default was to define BN_RECURSION
19899 but the BN code had some problems that would cause failures when
19900 doing certificate verification and some other functions.
19902 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19904 * Add ASN1 and PEM code to support netscape certificate sequences.
19908 * Add ASN1 and PEM code to support netscape certificate sequences.
19912 * Add several PKIX and private extended key usage OIDs.
19916 * Modify the 'ca' program to handle the new extension code. Modify
19917 openssl.cnf for new extension format, add comments.
19921 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19922 and add a sample to openssl.cnf so req -x509 now adds appropriate
19927 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19928 error code, add initial support to X509_print() and x509 application.
19932 * Takes a deep breath and start adding X509 V3 extension support code. Add
19933 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19934 stuff is currently isolated and isn't even compiled yet.
19938 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19939 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19940 Removed the versions check from X509 routines when loading extensions:
19941 this allows certain broken certificates that don't set the version
19942 properly to be processed.
19946 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19947 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19948 can still be regenerated with "make depend".
19952 * Spelling mistake in C version of CAST-128.
19954 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19956 * Changes to the error generation code. The perl script err-code.pl
19957 now reads in the old error codes and retains the old numbers, only
19958 adding new ones if necessary. It also only changes the .err files if new
19959 codes are added. The makefiles have been modified to only insert errors
19960 when needed (to avoid needlessly modifying header files). This is done
19961 by only inserting errors if the .err file is newer than the auto generated
19962 C file. To rebuild all the error codes from scratch (the old behaviour)
19963 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19964 or delete all the .err files.
19968 * CAST-128 was incorrectly implemented for short keys. The C version has
19969 been fixed, but is untested. The assembler versions are also fixed, but
19970 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19971 to regenerate it if needed.
19972 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19973 Hagino <itojun@kame.net>*
19975 * File was opened incorrectly in randfile.c.
19977 *Ulf Möller <ulf@fitug.de>*
19979 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19980 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19981 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19982 al: it's just almost always a UTCTime. Note this patch adds new error
19983 codes so do a "make errors" if there are problems.
19987 * Correct Linux 1 recognition in config.
19989 *Ulf Möller <ulf@fitug.de>*
19991 * Remove pointless MD5 hash when using DSA keys in ca.
19993 *Anonymous <nobody@replay.com>*
19995 * Generate an error if given an empty string as a cert directory. Also
19996 generate an error if handed NULL (previously returned 0 to indicate an
19997 error, but didn't set one).
19999 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20001 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20005 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20006 parameters. This was causing a warning which killed off the Win32 compile.
20010 * Remove C++ style comments from crypto/bn/bn_local.h.
20012 *Neil Costigan <neil.costigan@celocom.com>*
20014 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20015 based on a text string, looking up short and long names and finally
20016 "dot" format. The "dot" format stuff didn't work. Added new function
20017 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20018 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20019 OID is not part of the table.
20023 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20024 X509_LOOKUP_by_alias().
20028 * Sort openssl functions by name.
20032 * Get the `gendsa` command working and add it to the `list` command. Remove
20033 encryption from sample DSA keys (in case anyone is interested the password
20038 * Make *all* `*_free` functions accept a NULL pointer.
20040 *Frans Heymans <fheymans@isaserver.be>*
20042 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20045 *Anonymous <nobody@replay.com>*
20047 * s_server should send the CAfile as acceptable CAs, not its own cert.
20049 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20051 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20053 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20055 * Temp key "for export" tests were wrong in s3_srvr.c.
20057 *Anonymous <nobody@replay.com>*
20059 * Add prototype for temp key callback functions
20060 SSL_CTX_set_tmp_{rsa,dh}_callback().
20064 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20065 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20069 * X509_name_add_entry() freed the wrong thing after an error.
20071 *Arne Ansper <arne@ats.cyber.ee>*
20073 * rsa_eay.c would attempt to free a NULL context.
20075 *Arne Ansper <arne@ats.cyber.ee>*
20077 * BIO_s_socket() had a broken should_retry() on Windoze.
20079 *Arne Ansper <arne@ats.cyber.ee>*
20081 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20083 *Arne Ansper <arne@ats.cyber.ee>*
20085 * Make sure the already existing X509_STORE->depth variable is initialized
20086 in X509_STORE_new(), but document the fact that this variable is still
20087 unused in the certificate verification process.
20089 *Ralf S. Engelschall*
20091 * Fix the various library and `apps/` files to free up pkeys obtained from
20092 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20096 * Fix reference counting in X509_PUBKEY_get(). This makes
20097 demos/maurice/example2.c work, amongst others, probably.
20099 *Steve Henson and Ben Laurie*
20101 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20102 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20103 are no longer created. This way we have a single and consistent command
20104 line interface `openssl <command>`, similar to `cvs <command>`.
20106 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20108 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20109 BIT STRING wrapper always have zero unused bits.
20113 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20117 * Make the top-level INSTALL documentation easier to understand.
20121 * Makefiles updated to exit if an error occurs in a sub-directory
20122 make (including if user presses ^C) [Paul Sutton]
20124 * Make Montgomery context stuff explicit in RSA data structure.
20128 * Fix build order of pem and err to allow for generated pem.h.
20132 * Fix renumbering bug in X509_NAME_delete_entry().
20136 * Enhanced the err-ins.pl script so it makes the error library number
20137 global and can add a library name. This is needed for external ASN1 and
20138 other error libraries.
20142 * Fixed sk_insert which never worked properly.
20146 * Fix ASN1 macros so they can handle indefinite length constructed
20147 EXPLICIT tags. Some non standard certificates use these: they can now
20152 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20153 into a single doc/ssleay.txt bundle. This way the information is still
20154 preserved but no longer messes up this directory. Now it's new room for
20155 the new set of documentation files.
20157 *Ralf S. Engelschall*
20159 * SETs were incorrectly DER encoded. This was a major pain, because they
20160 shared code with SEQUENCEs, which aren't coded the same. This means that
20161 almost everything to do with SETs or SEQUENCEs has either changed name or
20162 number of arguments.
20164 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20166 * Fix test data to work with the above.
20170 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20171 was already fixed by Eric for 0.9.1 it seems.
20173 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20175 * Autodetect FreeBSD3.
20179 * Fix various bugs in Configure. This affects the following platforms:
20183 unixware-2.0-pentium
20188 * Eliminate generated files from CVS. Reorder tests to regenerate files
20189 before they are needed.
20193 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20197 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20199 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20200 changed SSLeay to OpenSSL in version strings.
20202 *Ralf S. Engelschall*
20204 * Some fixups to the top-level documents.
20208 * Fixed the nasty bug where rsaref.h was not found under compile-time
20209 because the symlink to include/ was missing.
20211 *Ralf S. Engelschall*
20213 * Incorporated the popular no-RSA/DSA-only patches
20214 which allow to compile an RSA-free SSLeay.
20216 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20218 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20219 when "ssleay" is still not found.
20221 *Ralf S. Engelschall*
20223 * Added more platforms to Configure: Cray T3E, HPUX 11,
20225 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20227 * Updated the README file.
20229 *Ralf S. Engelschall*
20231 * Added various .cvsignore files in the CVS repository subdirs
20232 to make a "cvs update" really silent.
20234 *Ralf S. Engelschall*
20236 * Recompiled the error-definition header files and added
20237 missing symbols to the Win32 linker tables.
20239 *Ralf S. Engelschall*
20241 * Cleaned up the top-level documents;
20242 o new files: CHANGES and LICENSE
20243 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20244 o merged COPYRIGHT into LICENSE
20245 o removed obsolete TODO file
20246 o renamed MICROSOFT to INSTALL.W32
20248 *Ralf S. Engelschall*
20250 * Removed dummy files from the 0.9.1b source tree:
20251 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20252 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20253 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20254 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20255 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20257 *Ralf S. Engelschall*
20259 * Added various platform portability fixes.
20263 * The Genesis of the OpenSSL rpject:
20264 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20265 Young and Tim J. Hudson created while they were working for C2Net until
20268 *The OpenSSL Project*
20270 ### Changes between 0.9.0b and 0.9.1b [not released]
20272 * Updated a few CA certificates under certs/
20276 * Changed some BIGNUM api stuff.
20280 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20281 DGUX x86, Linux Alpha, etc.
20285 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20286 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20291 * Add -strparse option to asn1pars program which parses nested
20294 *Dr Stephen Henson <shenson@bigfoot.com>*
20296 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20300 * DSA fix for "ca" program.
20304 * Added "-genkey" option to "dsaparam" program.
20308 * Added RIPE MD160 (rmd160) message digest.
20312 * Added -a (all) option to "ssleay version" command.
20316 * Added PLATFORM define which is the id given to Configure.
20320 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20324 * Extended the ASN.1 parser routines.
20328 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20332 * Added a BN_CTX to the BN library.
20336 * Fixed the weak key values in DES library
20340 * Changed API in EVP library for cipher aliases.
20344 * Added support for RC2/64bit cipher.
20348 * Converted the lhash library to the crypto/mem.c functions.
20352 * Added more recognized ASN.1 object ids.
20356 * Added more RSA padding checks for SSL/TLS.
20360 * Added BIO proxy/filter functionality.
20364 * Added extra_certs to SSL_CTX which can be used
20365 send extra CA certificates to the client in the CA cert chain sending
20366 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20370 * Now Fortezza is denied in the authentication phase because
20371 this is key exchange mechanism is not supported by SSLeay at all.
20375 * Additional PKCS1 checks.
20379 * Support the string "TLSv1" for all TLS v1 ciphers.
20383 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20384 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20388 * Fixed a few memory leaks.
20392 * Fixed various code and comment typos.
20396 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20397 bytes sent in the client random.
20399 *Edward Bishop <ebishop@spyglass.com>*
20403 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20404 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20405 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20406 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20407 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20408 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20409 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20410 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20411 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20412 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20413 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20414 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20415 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20416 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20417 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20418 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20419 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20420 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20421 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20422 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20423 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20424 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20425 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20426 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20427 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20428 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20429 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20430 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20431 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20432 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20433 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20434 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20435 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20436 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20437 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20438 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20439 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20440 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20441 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20442 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20443 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20444 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20445 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20446 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20447 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20448 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20449 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20450 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20451 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20452 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20453 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20454 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20455 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20456 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20457 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20458 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20459 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20460 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20461 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20462 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20463 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20464 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20465 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20466 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20467 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20468 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20469 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20470 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20471 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20472 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20473 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20474 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20475 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20476 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20477 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20478 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20479 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20480 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20481 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20482 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20483 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20484 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20485 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20486 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20487 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20488 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20489 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20490 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20491 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20492 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20493 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20494 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20495 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20496 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20497 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20498 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20499 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20500 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20501 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20502 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20503 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20504 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20505 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20506 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20507 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20508 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20509 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20510 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20511 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20512 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20513 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20514 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20515 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20516 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20517 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20518 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20519 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20520 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20521 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20522 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20523 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20524 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20525 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20526 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20527 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20528 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20529 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20530 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20531 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20532 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20533 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20534 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20535 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20536 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20537 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20538 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20539 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20540 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20541 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20542 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20543 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20544 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20545 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20546 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20547 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20548 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20549 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20550 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20551 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20552 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20553 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20554 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20555 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20556 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20557 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20558 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20559 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20560 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20561 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20562 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20563 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20564 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20565 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20566 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20567 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20568 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20569 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20570 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20571 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20572 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20573 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20574 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20575 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20576 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20577 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20578 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20579 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20580 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20581 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20582 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20583 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20584 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20585 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20586 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20587 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20588 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655