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 configuration setting for providers in openssl.cnf has been
32 updated to require a value of [1|yes|true|on] (in lower or UPPER case) to
33 activate the provider. Conversely a setting [0|no|false|off] will prevent
34 provider activation. All other values, or the omission of a value for this
35 setting will result in an error.
39 * In `openssl speed`, changed the default hash function used with `hmac` from
44 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
45 - `certProfile` request message header and respective `-profile` CLI option
49 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
50 be less hard coded in the build file templates, and to allow easier
51 addition of more exporters. With that, an exporter for CMake is also
56 * The BLAKE2s hash algorithm matches BLAKE2b's support
57 for configurable output length.
61 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
62 server to prefer session resumption using PSK-only key exchange over PSK
63 with DHE, if both are available.
65 *Markus Minichmayr, Tapkey GmbH*
70 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
72 * The BLAKE2b hash algorithm supports a configurable output length
73 by setting the "size" parameter.
75 *Čestmír Kalina and Tomáš Mráz*
77 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
81 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
82 and the corresponding provider-storemgmt API function
83 OSSL_FUNC_store_delete().
87 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
88 a passphrase callback when opening a store.
92 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
93 from 8 bytes to 16 bytes.
94 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
95 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
96 requires a salt length of 128 bits. This affects OpenSSL command line
97 applications such as "genrsa" and "pkcs8" and API's such as
98 PEM_write_bio_PrivateKey() that are reliant on the default value.
99 The additional commandline option 'saltlen' has been added to the
100 OpenSSL command line applications for "pkcs8" and "enc" to allow the
101 salt length to be set to a non default value.
105 * Changed the default value of the `ess_cert_id_alg` configuration
106 option which is used to calculate the TSA's public key certificate
107 identifier. The default algorithm is updated to be sha256 instead
110 *Małgorzata Olszówka*
112 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
113 table for point multiplication of the base point, which increases the size of
114 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
115 been added to disable the precomputed table.
119 * Added client side support for QUIC
121 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
123 * Added multiple tutorials on the OpenSSL library and in particular
124 on writing various clients (using TLS and QUIC protocols) with libssl.
128 * Added secp384r1 implementation using Solinas' reduction to improve
129 speed of the NIST P-384 elliptic curve. To enable the implementation
130 the build option `enable-ec_nistp_64_gcc_128` must be used.
134 * Improved RFC7468 compliance of the asn1parse command.
136 *Matthias St. Pierre*
138 * Added SHA256/192 algorithm support.
142 * Added support for securely getting root CA certificate update in
147 * Improved contention on global write locks by using more read locks where
152 * Improved performance of OSSL_PARAM lookups in performance critical
157 * Added the SSL_get0_group_name() function to provide access to the
158 name of the group used for the TLS key exchange.
162 * Provide a new configure option `no-http` that can be used to disable the
163 HTTP support. Provide new configure options `no-apps` and `no-docs` to
164 disable building the openssl command line application and the documentation.
168 * Provide a new configure option `no-ecx` that can be used to disable the
169 X25519, X448, and EdDSA support.
173 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
174 the EVP_KDF_CTX_set_params() function they are now concatenated not just
175 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
179 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
180 the provider context as a parameter.
184 * TLS round-trip time calculation was added by a Brigham Young University
185 Capstone team partnering with Sandia National Laboratories. A new function
186 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
191 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
192 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
193 option. Use of the "advanced" s_client command command via the "-adv" option
198 * Added an "advanced" command mode to s_client. Use this with the "-adv"
199 option. The old "basic" command mode recognises certain letters that must
200 always appear at the start of a line and cannot be escaped. The advanced
201 command mode enables commands to be entered anywhere and there is an
202 escaping mechanism. After starting s_client with "-adv" type "{help}"
203 to show a list of available commands.
207 * Add Raw Public Key (RFC7250) support. Authentication is supported
208 by matching keys against either local policy (TLSA records synthesised
209 from the expected keys) or DANE (TLSA records obtained by the
210 application from DNS). TLSA records will also match the same key in
211 the server certificate, should RPK use not happen to be negotiated.
215 * Added support for modular exponentiation and CRT offloading for the
220 * Added further assembler code for the RISC-V architecture.
224 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
225 from a given EC_GROUP.
229 * Improved support for non-default library contexts and property queries
230 when parsing PKCS#12 files.
234 * Implemented support for all five instances of EdDSA from RFC8032:
235 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
236 The streaming is not yet supported for the HashEdDSA variants
237 (Ed25519ph and Ed448ph).
241 * Added SM4 optimization for ARM processors using ASIMD and AES HW
246 * Implemented SM4-XTS support.
250 * Added platform-agnostic OSSL_sleep() function.
254 * Implemented deterministic ECDSA signatures (RFC6979) support.
258 * Implemented AES-GCM-SIV (RFC8452) support.
262 * Added support for pluggable (provider-based) TLS signature algorithms.
263 This enables TLS 1.3 authentication operations with algorithms embedded
264 in providers not included by default in OpenSSL. In combination with
265 the already available pluggable KEM and X.509 support, this enables
266 for example suitable providers to deliver post-quantum or quantum-safe
267 cryptography to OpenSSL users.
271 * Added support for pluggable (provider-based) CMS signature algorithms.
272 This enables CMS sign and verify operations with algorithms embedded
273 in providers not included by default in OpenSSL.
277 * Added support for Hybrid Public Key Encryption (HPKE) as defined
278 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
279 Message Layer Security (MLS) and other IETF specifications.
280 HPKE can also be used by other applications that require
281 encrypting "to" an ECDH public key. External APIs are defined in
282 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
286 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
291 * Add support for certificate compression (RFC8879), including
292 library support for Brotli and Zstandard compression.
296 * Add the ability to add custom attributes to PKCS12 files. Add a new API
297 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
298 for a user specified callback and optional argument.
299 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
300 added to the existing STACK_OF attrs.
304 * Major refactor of the libssl record layer.
308 * Add a mac salt length option for the pkcs12 command.
312 * Add more SRTP protection profiles from RFC8723 and RFC8269.
316 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
318 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
320 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
321 supported and enabled.
325 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
326 to the list of ciphersuites providing Perfect Forward Secrecy as
327 required by SECLEVEL >= 3.
329 *Dmitry Belyavskiy, Nicola Tuveri*
331 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
332 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
333 SSL_get0_iana_groups() function-like macro, retrieves the list of
334 supported groups sent by the peer.
335 The function SSL_client_hello_get_extension_order() populates
336 a caller-supplied array with the list of extension types present in the
337 ClientHello, in order of appearance.
341 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
342 to make it possible to use empty passphrase strings.
346 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
350 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
351 to change functions used for allocating the memory of asynchronous call stack.
355 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
359 * A failure exit code is returned when using the openssl x509 command to check
360 certificate attributes and the checks fail.
364 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
365 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
366 of 160 bits and above and less than 224 bits were previously accepted by
367 default but are now no longer allowed. By default TLS compression was
368 already disabled in previous OpenSSL versions. At security level 2 it cannot
373 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
378 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
379 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
380 will need to load the legacy crypto provider.
384 * CCM8 cipher suites in TLS have been downgraded to security level zero
385 because they use a short authentication tag which lowers their strength.
389 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
394 * Add X.509 certificate codeSigning purpose and related checks on key usage and
395 extended key usage of the leaf certificate according to the CA/Browser Forum.
399 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
400 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
401 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
402 X.509 version 3 if the certificate information includes X.509 extensions.
406 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
407 such as adding a trace facility for debugging certificate chain building.
411 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
412 in particular supporting requests for central key generation, generalized
413 polling, and various types of genm/genp exchanges defined in CMP Updates.
417 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
418 like correcting the TLS and proxy support and adding tracing for debugging.
422 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
426 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
427 a certificate to be added is already present. `CMS_sign_ex()` and
428 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
429 and no longer throw an error for them.
433 * Fixed and extended `util/check-format.pl` for checking adherence to the
434 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
435 The checks are meanwhile more complete and yield fewer false positives.
439 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
440 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
441 calls. They can be used as the transport BIOs for QUIC.
443 *Hugo Landau, Matt Caswell and Tomáš Mráz*
445 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
446 sending and receiving multiple messages in a single call. An implementation
447 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
451 * Support for loading root certificates from the Windows certificate store
452 has been added. The support is in the form of a store which recognises the
453 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
454 arguments. This store is built by default and can be disabled using the new
455 compile-time option `no-winstore`. This store is not currently used by
456 default and must be loaded explicitly using the above store URI. It is
457 expected to be loaded by default in the future.
461 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
462 kernel versions that support KTLS have a known bug in CCM processing. That
463 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
464 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
469 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
474 * Zerocopy KTLS sendfile() support on Linux.
478 * The OBJ_ calls are now thread safe using a global lock.
482 * New parameter `-digest` for openssl cms command allowing signing
483 pre-computed digests and new CMS API functions supporting that
488 * OPENSSL_malloc() and other allocation functions now raise errors on
489 allocation failures. The callers do not need to explicitly raise errors
490 unless they want to for tracing purposes.
494 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
495 decryption as a protection against Bleichenbacher-like attacks.
496 The RSA decryption API will now return a randomly generated deterministic
497 message instead of an error in case it detects an error when checking
498 padding during PKCS#1 v1.5 decryption. This is a general protection against
499 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
501 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
502 on the RSA decryption context.
506 * Added support for Brainpool curves in TLS-1.3.
508 *Bernd Edlinger and Matt Caswell*
510 * Added OpenBSD specific build targets.
514 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
515 a basic thread pool implementation for select platforms.
522 ### Changes between 3.1.4 and 3.1.5 [xx XXX xxxx]
524 * Fix excessive time spent in DH check / generation with large Q parameter
527 Applications that use the functions DH_generate_key() to generate an
528 X9.42 DH key may experience long delays. Likewise, applications that use
529 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
530 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
531 Where the key or parameters that are being checked have been obtained from
532 an untrusted source this may lead to a Denial of Service.
538 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
540 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
541 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
542 that alter the key or IV length ([CVE-2023-5363]).
546 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
548 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
550 The POLY1305 MAC (message authentication code) implementation in OpenSSL
551 does not save the contents of non-volatile XMM registers on Windows 64
552 platform when calculating the MAC of data larger than 64 bytes. Before
553 returning to the caller all the XMM registers are set to zero rather than
554 restoring their previous content. The vulnerable code is used only on newer
555 x86_64 processors supporting the AVX512-IFMA instructions.
557 The consequences of this kind of internal application state corruption can
558 be various - from no consequences, if the calling application does not
559 depend on the contents of non-volatile XMM registers at all, to the worst
560 consequences, where the attacker could get complete control of the
561 application process. However given the contents of the registers are just
562 zeroized so the attacker cannot put arbitrary values inside, the most likely
563 consequence, if any, would be an incorrect result of some application
564 dependent calculations or a crash leading to a denial of service.
570 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
572 * Fix excessive time spent checking DH q parameter value.
574 The function DH_check() performs various checks on DH parameters. After
575 fixing CVE-2023-3446 it was discovered that a large q parameter value can
576 also trigger an overly long computation during some of these checks.
577 A correct q value, if present, cannot be larger than the modulus p
578 parameter, thus it is unnecessary to perform these checks if q is larger
581 If DH_check() is called with such q parameter value,
582 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
583 intensive checks are skipped.
589 * Fix DH_check() excessive time with over sized modulus.
591 The function DH_check() performs various checks on DH parameters. One of
592 those checks confirms that the modulus ("p" parameter) is not too large.
593 Trying to use a very large modulus is slow and OpenSSL will not normally use
594 a modulus which is over 10,000 bits in length.
596 However the DH_check() function checks numerous aspects of the key or
597 parameters that have been supplied. Some of those checks use the supplied
598 modulus value even if it has already been found to be too large.
600 A new limit has been added to DH_check of 32,768 bits. Supplying a
601 key/parameters with a modulus over this size will simply cause DH_check() to
608 * Do not ignore empty associated data entries with AES-SIV.
610 The AES-SIV algorithm allows for authentication of multiple associated
611 data entries along with the encryption. To authenticate empty data the
612 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
613 with NULL pointer as the output buffer and 0 as the input buffer length.
614 The AES-SIV implementation in OpenSSL just returns success for such call
615 instead of performing the associated data authentication operation.
616 The empty data thus will not be authenticated. ([CVE-2023-2975])
618 Thanks to Juerg Wullschleger (Google) for discovering the issue.
620 The fix changes the authentication tag value and the ciphertext for
621 applications that use empty associated data entries with AES-SIV.
622 To decrypt data encrypted with previous versions of OpenSSL the application
623 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
628 * When building with the `enable-fips` option and using the resulting
629 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
630 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
631 not operate with truncated digests (FIPS 140-3 IG G.R).
635 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
637 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
638 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
640 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
641 numeric text form. For gigantic sub-identifiers, this would take a very
642 long time, the time complexity being O(n^2) where n is the size of that
643 sub-identifier. ([CVE-2023-2650])
645 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
646 IDENTIFIER to canonical numeric text form if the size of that OBJECT
647 IDENTIFIER is 586 bytes or less, and fail otherwise.
649 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
650 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
651 most 128 sub-identifiers, and that the maximum value that each sub-
652 identifier may have is 2^32-1 (4294967295 decimal).
654 For each byte of every sub-identifier, only the 7 lower bits are part of
655 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
656 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
661 * Multiple algorithm implementation fixes for ARM BE platforms.
665 * Added a -pedantic option to fipsinstall that adjusts the various
666 settings to ensure strict FIPS compliance rather than backwards
671 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
672 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
673 trigger a crash of an application using AES-XTS decryption if the memory
674 just after the buffer being decrypted is not mapped.
675 Thanks to Anton Romanov (Amazon) for discovering the issue.
680 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
681 The previous fix for this timing side channel turned out to cause
682 a severe 2-3x performance regression in the typical use case
683 compared to 3.0.7. The new fix uses existing constant time
684 code paths, and restores the previous performance level while
685 fully eliminating all existing timing side channels.
686 The fix was developed by Bernd Edlinger with testing support
691 * Add FIPS provider configuration option to disallow the use of
692 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
693 The option '-no_drbg_truncated_digests' can optionally be
694 supplied to 'openssl fipsinstall'.
698 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
699 that it does not enable policy checking. Thanks to David Benjamin for
700 discovering this issue.
705 * Fixed an issue where invalid certificate policies in leaf certificates are
706 silently ignored by OpenSSL and other certificate policy checks are skipped
707 for that certificate. A malicious CA could use this to deliberately assert
708 invalid certificate policies in order to circumvent policy checking on the
709 certificate altogether.
714 * Limited the number of nodes created in a policy tree to mitigate
715 against CVE-2023-0464. The default limit is set to 1000 nodes, which
716 should be sufficient for most installations. If required, the limit
717 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
718 time define to a desired maximum number of nodes or zero to allow
724 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
726 * Add FIPS provider configuration option to enforce the
727 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
728 The option '-ems-check' can optionally be supplied to
729 'openssl fipsinstall'.
733 * The FIPS provider includes a few non-approved algorithms for
734 backward compatibility purposes and the "fips=yes" property query
735 must be used for all algorithm fetches to ensure FIPS compliance.
737 The algorithms that are included but not approved are Triple DES ECB,
738 Triple DES CBC and EdDSA.
742 * Added support for KMAC in KBKDF.
746 * RNDR and RNDRRS support in provider functions to provide
747 random number generation for Arm CPUs (aarch64).
751 * s_client and s_server apps now explicitly say when the TLS version
752 does not include the renegotiation mechanism. This avoids confusion
753 between that scenario versus when the TLS version includes secure
754 renegotiation but the peer lacks support for it.
758 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
760 *Tomasz Kantecki, Andrey Matyukov*
762 * The various OBJ_* functions have been made thread safe.
766 * Parallel dual-prime 1536/2048-bit modular exponentiation for
767 AVX512_IFMA capable processors.
769 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
771 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
772 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
773 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
774 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
775 `OPENSSL_NO_DEPRECATED_3_1`.
777 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
778 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
779 definitions for these functions regardless of whether
780 `OPENSSL_NO_DEPRECATED_3_1` is defined.
782 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
783 functions regardless of whether they are using them. It is recommended that
784 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
788 * When generating safe-prime DH parameters set the recommended private key
789 length equivalent to minimum key lengths as in RFC 7919.
793 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
794 maximum size that is smaller or equal to the digest length to comply with
795 FIPS 186-4 section 5. This is implemented by a new option
796 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
797 `rsa_pss_saltlen` parameter, which is now the default. Signature
798 verification is not affected by this change and continues to work as before.
805 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
806 listed here are only a brief description.
807 The migration guide contains more detailed information related to new features,
808 breaking changes, and mappings for the large list of deprecated functions.
810 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
812 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
814 * Fixed NULL dereference during PKCS7 data verification.
816 A NULL pointer can be dereferenced when signatures are being
817 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
818 algorithm used for the signature is known to the OpenSSL library but
819 the implementation of the hash algorithm is not available the digest
820 initialization will fail. There is a missing check for the return
821 value from the initialization function which later leads to invalid
822 usage of the digest API most likely leading to a crash.
825 PKCS7 data is processed by the SMIME library calls and also by the
826 time stamp (TS) library calls. The TLS implementation in OpenSSL does
827 not call these functions however third party applications would be
828 affected if they call these functions to verify signatures on untrusted
833 * Fixed X.400 address type confusion in X.509 GeneralName.
835 There is a type confusion vulnerability relating to X.400 address processing
836 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
837 but the public structure definition for GENERAL_NAME incorrectly specified
838 the type of the x400Address field as ASN1_TYPE. This field is subsequently
839 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
842 When CRL checking is enabled (i.e. the application sets the
843 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
844 pass arbitrary pointers to a memcmp call, enabling them to read memory
845 contents or enact a denial of service.
850 * Fixed NULL dereference validating DSA public key.
852 An invalid pointer dereference on read can be triggered when an
853 application tries to check a malformed DSA public key by the
854 EVP_PKEY_public_check() function. This will most likely lead
855 to an application crash. This function can be called on public
856 keys supplied from untrusted sources which could allow an attacker
857 to cause a denial of service attack.
859 The TLS implementation in OpenSSL does not call this function
860 but applications might call the function if there are additional
861 security requirements imposed by standards such as FIPS 140-3.
864 *Shane Lontis, Tomáš Mráz*
866 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
868 An invalid pointer dereference on read can be triggered when an
869 application tries to load malformed PKCS7 data with the
870 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
872 The result of the dereference is an application crash which could
873 lead to a denial of service attack. The TLS implementation in OpenSSL
874 does not call this function however third party applications might
875 call these functions on untrusted data.
880 * Fixed Use-after-free following BIO_new_NDEF.
882 The public API function BIO_new_NDEF is a helper function used for
883 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
884 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
885 be called directly by end user applications.
887 The function receives a BIO from the caller, prepends a new BIO_f_asn1
888 filter BIO onto the front of it to form a BIO chain, and then returns
889 the new head of the BIO chain to the caller. Under certain conditions,
890 for example if a CMS recipient public key is invalid, the new filter BIO
891 is freed and the function returns a NULL result indicating a failure.
892 However, in this case, the BIO chain is not properly cleaned up and the
893 BIO passed by the caller still retains internal pointers to the previously
894 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
895 then a use-after-free will occur. This will most likely result in a crash.
898 *Viktor Dukhovni, Matt Caswell*
900 * Fixed Double free after calling PEM_read_bio_ex.
902 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
903 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
904 data. If the function succeeds then the "name_out", "header" and "data"
905 arguments are populated with pointers to buffers containing the relevant
906 decoded data. The caller is responsible for freeing those buffers. It is
907 possible to construct a PEM file that results in 0 bytes of payload data.
908 In this case PEM_read_bio_ex() will return a failure code but will populate
909 the header argument with a pointer to a buffer that has already been freed.
910 If the caller also frees this buffer then a double free will occur. This
911 will most likely lead to a crash.
913 The functions PEM_read_bio() and PEM_read() are simple wrappers around
914 PEM_read_bio_ex() and therefore these functions are also directly affected.
916 These functions are also called indirectly by a number of other OpenSSL
917 functions including PEM_X509_INFO_read_bio_ex() and
918 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
919 internal uses of these functions are not vulnerable because the caller does
920 not free the header argument if PEM_read_bio_ex() returns a failure code.
923 *Kurt Roeckx, Matt Caswell*
925 * Fixed Timing Oracle in RSA Decryption.
927 A timing based side channel exists in the OpenSSL RSA Decryption
928 implementation which could be sufficient to recover a plaintext across
929 a network in a Bleichenbacher style attack. To achieve a successful
930 decryption an attacker would have to be able to send a very large number
931 of trial messages for decryption. The vulnerability affects all RSA padding
932 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
935 *Dmitry Belyavsky, Hubert Kario*
937 * Fixed X.509 Name Constraints Read Buffer Overflow.
939 A read buffer overrun can be triggered in X.509 certificate verification,
940 specifically in name constraint checking. The read buffer overrun might
941 result in a crash which could lead to a denial of service attack.
942 In a TLS client, this can be triggered by connecting to a malicious
943 server. In a TLS server, this can be triggered if the server requests
944 client authentication and a malicious client connects.
949 * Fixed X.509 Policy Constraints Double Locking security issue.
951 If an X.509 certificate contains a malformed policy constraint and
952 policy processing is enabled, then a write lock will be taken twice
953 recursively. On some operating systems (most widely: Windows) this
954 results in a denial of service when the affected process hangs. Policy
955 processing being enabled on a publicly facing server is not considered
956 to be a common setup.
961 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
962 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
963 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
964 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
965 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
966 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
967 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
968 for legacy EC and SM2 keys is also changed similarly to honor the
969 equivalent conversion format flag as specified in the underlying
970 `EC_KEY` object being exported to a provider, when this function is
971 called through `EVP_PKEY_export()`.
975 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
977 * Fixed two buffer overflows in punycode decoding functions.
979 A buffer overrun can be triggered in X.509 certificate verification,
980 specifically in name constraint checking. Note that this occurs after
981 certificate chain signature verification and requires either a CA to
982 have signed the malicious certificate or for the application to continue
983 certificate verification despite failure to construct a path to a trusted
986 In a TLS client, this can be triggered by connecting to a malicious
987 server. In a TLS server, this can be triggered if the server requests
988 client authentication and a malicious client connects.
990 An attacker can craft a malicious email address to overflow
991 an arbitrary number of bytes containing the `.` character (decimal 46)
992 on the stack. This buffer overflow could result in a crash (causing a
996 An attacker can craft a malicious email address to overflow four
997 attacker-controlled bytes on the stack. This buffer overflow could
998 result in a crash (causing a denial of service) or potentially remote code
999 execution depending on stack layout for any given platform/compiler.
1004 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1005 parameters in OpenSSL code.
1006 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1007 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1008 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1009 Using these invalid names may cause algorithms to use slower methods
1010 that ignore the CRT parameters.
1014 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1019 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1020 data to be signed before signing the certificate.
1024 * Added RIPEMD160 to the default provider.
1028 * Ensured that the key share group sent or accepted for the key exchange
1029 is allowed for the protocol version.
1033 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1035 * OpenSSL supports creating a custom cipher via the legacy
1036 EVP_CIPHER_meth_new() function and associated function calls. This function
1037 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1038 to use the new provider mechanism in order to implement custom ciphers.
1040 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1041 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1042 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1043 and decryption initialisation functions). Instead of using the custom cipher
1044 directly it incorrectly tries to fetch an equivalent cipher from the
1045 available providers. An equivalent cipher is found based on the NID passed to
1046 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1047 given cipher. However it is possible for an application to incorrectly pass
1048 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1049 is used in this way the OpenSSL encryption/decryption initialisation function
1050 will match the NULL cipher as being equivalent and will fetch this from the
1051 available providers. This will succeed if the default provider has been
1052 loaded (or if a third party provider has been loaded that offers this
1053 cipher). Using the NULL cipher means that the plaintext is emitted as the
1056 Applications are only affected by this issue if they call
1057 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1058 encryption/decryption initialisation function. Applications that only use
1059 SSL/TLS are not impacted by this issue.
1064 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1069 * Fixed the linux-mips64 Configure target which was missing the
1070 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1075 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1080 * Correctly handle a retransmitted ClientHello in DTLS
1084 * Fixed detection of ktls support in cross-compile environment on Linux
1088 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1093 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1094 report correct results in some cases
1098 * Fix UWP builds by defining VirtualLock
1102 * For known safe primes use the minimum key length according to RFC 7919.
1103 Longer private key sizes unnecessarily raise the cycles needed to compute the
1104 shared secret without any increase of the real security. This fixes a
1105 regression from 1.1.1 where these shorter keys were generated for the known
1110 * Added the loongarch64 target
1114 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1115 only passed to the FIPS provider and not to the default or legacy provider.
1119 * Fixed reported performance degradation on aarch64. Restored the
1120 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1121 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1122 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1123 The new algorithm is still used for 32 bit targets.
1127 * Added a missing header for memcmp that caused compilation failure on some
1132 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1134 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1135 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1136 This issue makes the RSA implementation with 2048 bit private keys
1137 incorrect on such machines and memory corruption will happen during
1138 the computation. As a consequence of the memory corruption an attacker
1139 may be able to trigger a remote code execution on the machine performing
1142 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1143 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1144 are affected by this issue.
1149 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1150 implementation would not encrypt the entirety of the data under some
1151 circumstances. This could reveal sixteen bytes of data that was
1152 preexisting in the memory that wasn't written. In the special case of
1153 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1155 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1156 they are both unaffected.
1159 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1161 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1163 * In addition to the c_rehash shell command injection identified in
1164 CVE-2022-1292, further bugs where the c_rehash script does not
1165 properly sanitise shell metacharacters to prevent command injection have been
1168 When the CVE-2022-1292 was fixed it was not discovered that there
1169 are other places in the script where the file names of certificates
1170 being hashed were possibly passed to a command executed through the shell.
1172 This script is distributed by some operating systems in a manner where
1173 it is automatically executed. On such operating systems, an attacker
1174 could execute arbitrary commands with the privileges of the script.
1176 Use of the c_rehash script is considered obsolete and should be replaced
1177 by the OpenSSL rehash command line tool.
1180 *Daniel Fiala, Tomáš Mráz*
1182 * Case insensitive string comparison no longer uses locales. It has instead
1183 been directly implemented.
1187 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1189 * Case insensitive string comparison is reimplemented via new locale-agnostic
1190 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1191 comparison. The previous implementation had problems when the Turkish locale
1196 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1197 metacharacters to prevent command injection. This script is distributed by
1198 some operating systems in a manner where it is automatically executed. On
1199 such operating systems, an attacker could execute arbitrary commands with the
1200 privileges of the script.
1202 Use of the c_rehash script is considered obsolete and should be replaced
1203 by the OpenSSL rehash command line tool.
1208 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1209 certificate on an OCSP response. The bug caused the function in the case
1210 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1211 response (meaning a successful verification) even in the case where the
1212 response signing certificate fails to verify.
1214 It is anticipated that most users of `OCSP_basic_verify` will not use the
1215 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1216 a negative value (indicating a fatal error) in the case of a certificate
1217 verification failure. The normal expected return value in this case would be
1220 This issue also impacts the command line OpenSSL "ocsp" application. When
1221 verifying an ocsp response with the "-no_cert_checks" option the command line
1222 application will report that the verification is successful even though it
1223 has in fact failed. In this case the incorrect successful response will also
1224 be accompanied by error messages showing the failure and contradicting the
1225 apparently successful result.
1230 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1231 AAD data as the MAC key. This made the MAC key trivially predictable.
1233 An attacker could exploit this issue by performing a man-in-the-middle attack
1234 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1235 that the modified data would still pass the MAC integrity check.
1237 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1238 endpoint will always be rejected by the recipient and the connection will
1239 fail at that point. Many application protocols require data to be sent from
1240 the client to the server first. Therefore, in such a case, only an OpenSSL
1241 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1243 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1244 sent in both directions. In this case both clients and servers could be
1245 affected, regardless of the application protocol.
1247 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1248 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1249 the handshake when using this ciphersuite.
1251 The confidentiality of data is not impacted by this issue, i.e. an attacker
1252 cannot decrypt data that has been encrypted using this ciphersuite - they can
1255 In order for this attack to work both endpoints must legitimately negotiate
1256 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1257 OpenSSL 3.0, and is not available within the default provider or the default
1258 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1259 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1260 following must have occurred:
1262 1) OpenSSL must have been compiled with the (non-default) compile time option
1263 enable-weak-ssl-ciphers
1265 2) OpenSSL must have had the legacy provider explicitly loaded (either
1266 through application code or via configuration)
1268 3) The ciphersuite must have been explicitly added to the ciphersuite list
1270 4) The libssl security level must have been set to 0 (default is 1)
1272 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1274 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1275 others that both endpoints have in common
1280 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1281 occupied by the removed hash table entries.
1283 This function is used when decoding certificates or keys. If a long lived
1284 process periodically decodes certificates or keys its memory usage will
1285 expand without bounds and the process might be terminated by the operating
1286 system causing a denial of service. Also traversing the empty hash table
1287 entries will take increasingly more time.
1289 Typically such long lived processes might be TLS clients or TLS servers
1290 configured to accept client certificate authentication.
1293 *Hugo Landau, Aliaksei Levin*
1295 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1296 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1297 statistics are no longer supported. For compatibility, these statistics are
1298 still listed in the output but are now always reported as zero.
1302 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1304 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1305 for non-prime moduli.
1307 Internally this function is used when parsing certificates that contain
1308 elliptic curve public keys in compressed form or explicit elliptic curve
1309 parameters with a base point encoded in compressed form.
1311 It is possible to trigger the infinite loop by crafting a certificate that
1312 has invalid explicit curve parameters.
1314 Since certificate parsing happens prior to verification of the certificate
1315 signature, any process that parses an externally supplied certificate may thus
1316 be subject to a denial of service attack. The infinite loop can also be
1317 reached when parsing crafted private keys as they can contain explicit
1318 elliptic curve parameters.
1320 Thus vulnerable situations include:
1322 - TLS clients consuming server certificates
1323 - TLS servers consuming client certificates
1324 - Hosting providers taking certificates or private keys from customers
1325 - Certificate authorities parsing certification requests from subscribers
1326 - Anything else which parses ASN.1 elliptic curve parameters
1328 Also any other applications that use the BN_mod_sqrt() where the attacker
1329 can control the parameter values are vulnerable to this DoS issue.
1334 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1335 to the list of ciphersuites providing Perfect Forward Secrecy as
1336 required by SECLEVEL >= 3.
1338 *Dmitry Belyavskiy, Nicola Tuveri*
1340 * Made the AES constant time code for no-asm configurations
1341 optional due to the resulting 95% performance degradation.
1342 The AES constant time code can be enabled, for no assembly
1343 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1347 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1352 * The negative return value handling of the certificate verification callback
1353 was reverted. The replacement is to set the verification retry state with
1354 the SSL_set_retry_verify() function.
1358 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1360 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1361 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1362 verify a certificate supplied by a server. That function may return a
1363 negative return value to indicate an internal error (for example out of
1364 memory). Such a negative return value is mishandled by OpenSSL and will cause
1365 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1366 success and a subsequent call to SSL_get_error() to return the value
1367 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1368 returned by OpenSSL if the application has previously called
1369 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1370 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1371 totally unexpected and applications may not behave correctly as a result. The
1372 exact behaviour will depend on the application but it could result in
1373 crashes, infinite loops or other similar incorrect responses.
1375 This issue is made more serious in combination with a separate bug in OpenSSL
1376 3.0 that will cause X509_verify_cert() to indicate an internal error when
1377 processing a certificate chain. This will occur where a certificate does not
1378 include the Subject Alternative Name extension but where a Certificate
1379 Authority has enforced name constraints. This issue can occur even with valid
1385 * Corrected a few file name and file reference bugs in the build,
1386 installation and setup scripts, which lead to installation verification
1387 failures. Slightly enhanced the installation verification script.
1391 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1396 * Fixed PVK encoder to properly query for the passphrase.
1400 * Multiple fixes in the OSSL_HTTP API functions.
1404 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1405 OSSL_PARAM_INTEGER data type and return error on negative numbers
1406 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1407 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1411 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1415 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1419 * Multiple threading fixes.
1423 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1427 * Allow fetching an operation from the provider that owns an unexportable key
1428 as a fallback if that is still allowed by the property query.
1432 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1434 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1439 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1440 S390X capability vector to zero. This simplifies testing of different code
1441 paths on S390X architecture.
1445 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1446 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1447 SP 800-38D". The communication will fail at this point.
1451 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1452 confidential in EC_GROUP data.
1456 * The byte order mark (BOM) character is ignored if encountered at the
1457 beginning of a PEM-formatted file.
1461 * Added CMS support for the Russian GOST algorithms.
1465 * Due to move of the implementation of cryptographic operations
1466 to the providers, validation of various operation parameters can
1467 be postponed until the actual operation is executed where previously
1468 it happened immediately when an operation parameter was set.
1470 For example when setting an unsupported curve with
1471 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1472 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1474 *OpenSSL team members and many third party contributors*
1476 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1477 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1478 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1479 instead to retrieve these algorithms from a provider.
1483 * On build targets where the multilib postfix is set in the build
1484 configuration the libdir directory was changing based on whether
1485 the lib directory with the multilib postfix exists on the system
1486 or not. This unpredictable behavior was removed and eventual
1487 multilib postfix is now always added to the default libdir. Use
1488 `--libdir=lib` to override the libdir if adding the postfix is
1493 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1494 no longer interoperable with OpenSSL 1.1.1.
1498 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1499 function codes, this function can only cause problems for calling
1504 * Add a configurable flag to output date formats as ISO 8601. Does not
1505 change the default date format.
1509 * Version of MSVC earlier than 1300 could get link warnings, which could
1510 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1511 Support for this flag has been removed.
1515 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1516 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1517 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1518 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1519 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1523 * The signatures of the functions to get and set options on SSL and
1524 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1525 Some source code changes may be required.
1529 * The public definitions of conf_method_st and conf_st have been
1530 deprecated. They will be made opaque in a future release.
1532 *Rich Salz and Tomáš Mráz*
1534 * Client-initiated renegotiation is disabled by default. To allow it, use
1535 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1536 flag, or the "ClientRenegotiation" config parameter as appropriate.
1540 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1541 or modify relative pathname inclusion.
1545 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1546 validated. Please consult the README-FIPS and
1547 README-PROVIDERS files, as well as the migration guide.
1549 *OpenSSL team members and many third party contributors*
1551 * For the key types DH and DHX the allowed settable parameters are now different.
1555 * The openssl commands that read keys, certificates, and CRLs now
1556 automatically detect the PEM or DER format of the input files.
1558 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1560 * Added enhanced PKCS#12 APIs which accept a library context.
1564 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1568 * Added support for Kernel TLS (KTLS).
1570 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1572 * Support for RFC 5746 secure renegotiation is now required by default for
1573 SSL or TLS connections to succeed.
1577 * The signature of the `copy` functional parameter of the
1578 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1579 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1580 the signature of the `pub_decode` functional parameter of the
1581 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1582 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1586 * The error return values from some control calls (ctrl) have changed.
1590 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1594 * Many functions in the EVP_ namespace that are getters of values from
1595 implementations or contexts were renamed to include get or get0 in their
1596 names. Old names are provided as macro aliases for compatibility and
1601 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1602 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1603 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1608 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1611 * The output from the command line applications may have minor
1616 * The output from numerous "printing" may have minor changes.
1620 * Windows thread synchronization uses read/write primitives (SRWLock) when
1621 supported by the OS, otherwise CriticalSection continues to be used.
1625 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1626 work on read only BIO source/sinks that do not support these functions.
1627 This allows piping or redirection of a file BIO using stdin to be buffered
1628 into memory. This is used internally in OSSL_DECODER_from_bio().
1632 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1633 this function would return one of the values OSSL_STORE_INFO_NAME,
1634 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1635 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1636 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1637 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1638 using this function should be amended to handle the changed return value.
1642 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1643 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1644 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1645 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1646 contains more than one certificate identifier: This means that all
1647 certificates referenced there MUST be part of the validation chain.
1651 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1652 RC5, DESX and DES have been moved to the legacy provider.
1656 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1657 RIPEMD-160 have been moved to the legacy provider.
1661 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1666 * The deprecated functions EVP_PKEY_get0_RSA(),
1667 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1668 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1669 well as the similarly named "get1" functions behave differently in
1674 * A number of functions handling low-level keys or engines were deprecated
1675 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1676 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1677 EVP_PKEY_get0_siphash().
1681 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1682 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1683 will need to load the legacy crypto provider. This includes these PBE
1684 algorithms which use this KDF:
1685 - NID_pbeWithMD2AndDES_CBC
1686 - NID_pbeWithMD5AndDES_CBC
1687 - NID_pbeWithSHA1AndRC2_CBC
1688 - NID_pbeWithMD2AndRC2_CBC
1689 - NID_pbeWithMD5AndRC2_CBC
1690 - NID_pbeWithSHA1AndDES_CBC
1694 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1695 BIO_debug_callback() functions.
1699 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1700 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1704 * The RAND_METHOD APIs have been deprecated.
1708 * The SRP APIs have been deprecated.
1712 * Add a compile time option to prevent the caching of provider fetched
1713 algorithms. This is enabled by including the no-cached-fetch option
1714 at configuration time.
1718 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1719 count of PKCS12_DEFAULT_ITER.
1721 *Tomáš Mráz and Sahana Prasad*
1723 * The openssl speed command does not use low-level API calls anymore.
1727 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1730 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1732 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1736 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1737 providers may supply their own group implementations (using either the "key
1738 exchange" or the "key encapsulation" methods) which will automatically be
1739 detected and used by libssl.
1741 *Matt Caswell, Nicola Tuveri*
1743 * The undocumented function X509_certificate_type() has been deprecated;
1747 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1751 * Removed RSA padding mode for SSLv23 (which was only used for
1752 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1753 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1758 * Deprecated the obsolete X9.31 RSA key generation related functions.
1760 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1761 is not allowed to return a value > 1, this is no more taken as failure.
1763 *Viktor Dukhovni and David von Oheimb*
1765 * Deprecated the obsolete X9.31 RSA key generation related functions
1766 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1767 BN_X931_generate_prime_ex().
1771 * The default key generation method for the regular 2-prime RSA keys was
1772 changed to the FIPS 186-4 B.3.6 method.
1776 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1780 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1784 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1785 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1787 *Rich Salz, Richard Levitte, and David von Oheimb*
1789 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1793 * Deprecated `OCSP_parse_url()`.
1797 * Validation of SM2 keys has been separated from the validation of regular EC
1802 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1803 switches: a validation failure triggers an early exit, returning a failure
1804 exit status to the parent process.
1808 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1809 to ignore unknown ciphers.
1813 * The `-cipher-commands` and `-digest-commands` options
1814 of the command line utility `list` have been deprecated.
1815 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1819 * Added convenience functions for generating asymmetric key pairs:
1820 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1821 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1825 * All of the low-level EC_KEY functions have been deprecated.
1827 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1829 * Deprecated all the libcrypto and libssl error string loading
1834 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1835 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1840 * The `-crypt` option to the `passwd` command line tool has been removed.
1844 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1849 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1853 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1854 EVP_PKEY_get1_tls_encodedpoint().
1858 * The security callback, which can be customised by application code, supports
1859 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1860 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1864 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1865 interface. Their functionality remains unchanged.
1869 * Added new option for 'openssl list', '-providers', which will display the
1870 list of loaded providers, their names, version and status. It optionally
1871 displays their gettable parameters.
1875 * Removed EVP_PKEY_set_alias_type().
1879 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1880 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1884 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1885 parameters are still checked for type safety at compile time via helper
1890 * Remove the RAND_DRBG API
1892 *Paul Dale and Matthias St. Pierre*
1894 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1895 as well as actual hostnames.
1899 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1900 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1901 conversely, silently ignore DTLS protocol version bounds when configuring
1902 TLS-based contexts. The commands can be repeated to set bounds of both
1903 types. The same applies with the corresponding "min_protocol" and
1904 "max_protocol" command-line switches, in case some application uses both TLS
1907 SSL_CTX instances that are created for a fixed protocol version (e.g.
1908 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1909 attempts to apply bounds to these protocol versions would result in an
1910 error. Now only the "version-flexible" SSL_CTX instances are subject to
1911 limits in configuration files in command-line options.
1915 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1920 * Reworked the recorded ERR codes to make better space for system errors.
1921 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1922 given code is a system error (true) or an OpenSSL error (false).
1926 * Reworked the test perl framework to better allow parallel testing.
1928 *Nicola Tuveri and David von Oheimb*
1930 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1931 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1935 * 'Configure' has been changed to figure out the configuration target if
1936 none is given on the command line. Consequently, the 'config' script is
1937 now only a mere wrapper. All documentation is changed to only mention
1940 *Rich Salz and Richard Levitte*
1942 * Added a library context `OSSL_LIB_CTX` that applications as well as
1943 other libraries can use to form a separate context within which
1944 libcrypto operations are performed.
1948 * Added various `_ex` functions to the OpenSSL API that support using
1949 a non-default `OSSL_LIB_CTX`.
1953 * Handshake now fails if Extended Master Secret extension is dropped
1958 * Dropped interactive mode from the `openssl` program.
1962 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1964 *David von Oheimb and Shane Lontis*
1966 * Deprecated `EC_METHOD_get_field_type()`.
1970 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1971 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1972 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1976 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1980 * Add CAdES-BES signature verification support, mostly derived
1981 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1983 *Filipe Raimundo da Silva*
1985 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1989 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1990 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1994 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1998 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1999 EC_KEY_precompute_mult().
2003 * Deprecated EC_POINTs_mul().
2007 * Removed FIPS_mode() and FIPS_mode_set().
2011 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2015 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2016 EC_POINT_get_Jprojective_coordinates_GFp().
2020 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2021 arrays to be more easily constructed via a series of utility functions.
2022 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2023 the various push functions and finally convert to a passable OSSL_PARAM
2024 array using OSSL_PARAM_BLD_to_param().
2028 * The security strength of SHA1 and MD5 based signatures in TLS has been
2033 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2034 contain a provider side internal key.
2038 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2042 * Project text documents not yet having a proper file name extension
2043 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2044 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2048 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2049 have been converted to Markdown with the goal to produce documents
2050 which not only look pretty when viewed online in the browser, but
2051 remain well readable inside a plain text editor.
2053 To achieve this goal, a 'minimalistic' Markdown style has been applied
2054 which avoids formatting elements that interfere too much with the
2055 reading flow in the text file. For example, it
2057 * avoids [ATX headings][] and uses [setext headings][] instead
2058 (which works for `<h1>` and `<h2>` headings only).
2059 * avoids [inline links][] and uses [reference links][] instead.
2060 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2062 [ATX headings]: https://github.github.com/gfm/#atx-headings
2063 [setext headings]: https://github.github.com/gfm/#setext-headings
2064 [inline links]: https://github.github.com/gfm/#inline-link
2065 [reference links]: https://github.github.com/gfm/#reference-link
2066 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2067 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2069 *Matthias St. Pierre*
2071 * The test suite is changed to preserve results of each test recipe.
2072 A new directory test-runs/ with subdirectories named like the
2073 test recipes are created in the build tree for this purpose.
2077 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2078 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2079 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2081 *David von Oheimb, Martin Peylo*
2083 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2084 It supports arbitrary request and response content types, GET redirection,
2085 TLS, connections via HTTP(S) proxies, connections and exchange via
2086 user-defined BIOs (allowing implicit connections), persistent connections,
2087 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2088 The legacy OCSP-focused (and only partly documented) API
2089 is retained for backward compatibility, while most of it is deprecated.
2093 * Added `util/check-format.pl`, a tool for checking adherence to the
2094 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2095 The checks performed are incomplete and yield some false positives.
2096 Still the tool should be useful for detecting most typical glitches.
2100 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2101 If domain name resolution yields multiple IP addresses all of them are tried
2102 after `connect()` failures.
2106 * All of the low-level RSA functions have been deprecated.
2110 * X509 certificates signed using SHA1 are no longer allowed at security
2115 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2116 modified to use PKEY APIs. These commands are now in maintenance mode
2117 and no new features will be added to them.
2121 * The command line utility rsautl has been deprecated.
2125 * The command line utilities genrsa and rsa have been modified to use PKEY
2126 APIs. They now write PKCS#8 keys by default. These commands are now in
2127 maintenance mode and no new features will be added to them.
2131 * All of the low-level DH functions have been deprecated.
2133 *Paul Dale and Matt Caswell*
2135 * All of the low-level DSA functions have been deprecated.
2139 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2140 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2144 * Deprecated low-level ECDH and ECDSA functions.
2148 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2152 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2153 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2154 a new formulation to include all the things it can be used for,
2155 as well as words of caution.
2159 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2163 * All of the low-level HMAC functions have been deprecated.
2165 *Paul Dale and David von Oheimb*
2167 * Over two thousand fixes were made to the documentation, including:
2168 - Common options (such as -rand/-writerand, TLS version control, etc)
2169 were refactored and point to newly-enhanced descriptions in openssl.pod.
2170 - Added style conformance for all options (with help from Richard Levitte),
2171 documented all reported missing options, added a CI build to check
2172 that all options are documented and that no unimplemented options
2174 - Documented some internals, such as all use of environment variables.
2175 - Addressed all internal broken L<> references.
2179 * All of the low-level CMAC functions have been deprecated.
2183 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2184 functions have been deprecated.
2186 *Paul Dale and David von Oheimb*
2188 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2189 set of functions. The documentation mentioned negative values for some
2190 errors, but this was never the case, so the mention of negative values
2193 Code that followed the documentation and thereby check with something
2194 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2198 * All of the low-level cipher functions have been deprecated.
2200 *Matt Caswell and Paul Dale*
2202 * Removed include/openssl/opensslconf.h.in and replaced it with
2203 include/openssl/configuration.h.in, which differs in not including
2204 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2205 was added to include both.
2207 This allows internal hacks where one might need to modify the set
2208 of configured macros, for example this if deprecated symbols are
2209 still supposed to be available internally:
2211 #include <openssl/configuration.h>
2213 #undef OPENSSL_NO_DEPRECATED
2214 #define OPENSSL_SUPPRESS_DEPRECATED
2216 #include <openssl/macros.h>
2218 This should not be used by applications that use the exported
2219 symbols, as that will lead to linking errors.
2223 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2224 used in exponentiation with 512-bit moduli. No EC algorithms are
2225 affected. Analysis suggests that attacks against 2-prime RSA1024,
2226 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2227 difficult to perform and are not believed likely. Attacks against DH512
2228 are considered just feasible. However, for an attack the target would
2229 have to reuse the DH512 private key, which is not recommended anyway.
2230 Also applications directly using the low-level API BN_mod_exp may be
2231 affected if they use BN_FLG_CONSTTIME.
2236 * Most memory-debug features have been deprecated, and the functionality
2237 replaced with no-ops.
2241 * Added documentation for the STACK API.
2245 * Introduced a new method type and API, OSSL_ENCODER, to represent
2246 generic encoders. These do the same sort of job that PEM writers
2247 and d2i functions do, but with support for methods supplied by
2248 providers, and the possibility for providers to support other
2253 * Introduced a new method type and API, OSSL_DECODER, to represent
2254 generic decoders. These do the same sort of job that PEM readers
2255 and i2d functions do, but with support for methods supplied by
2256 providers, and the possibility for providers to support other
2261 * Added a .pragma directive to the syntax of configuration files, to
2262 allow varying behavior in a supported and predictable manner.
2263 Currently added pragma:
2267 This allows dollar signs to be a keyword character unless it's
2268 followed by a opening brace or parenthesis. This is useful for
2269 platforms where dollar signs are commonly used in names, such as
2270 volume names and system directory names on VMS.
2274 * Added functionality to create an EVP_PKEY from user data.
2278 * Change the interpretation of the '--api' configuration option to
2279 mean that this is a desired API compatibility level with no
2280 further meaning. The previous interpretation, that this would
2281 also mean to remove all deprecated symbols up to and including
2282 the given version, no requires that 'no-deprecated' is also used
2283 in the configuration.
2285 When building applications, the desired API compatibility level
2286 can be set with the OPENSSL_API_COMPAT macro like before. For
2287 API compatibility version below 3.0, the old style numerical
2288 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2289 For version 3.0 and on, the value is expected to be the decimal
2290 value calculated from the major and minor version like this:
2292 MAJOR * 10000 + MINOR * 100
2296 -DOPENSSL_API_COMPAT=30000 For 3.0
2297 -DOPENSSL_API_COMPAT=30200 For 3.2
2299 To hide declarations that are deprecated up to and including the
2300 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2301 given when building the application as well.
2305 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2306 access to certificate and CRL stores via URIs and OSSL_STORE
2309 This adds the following functions:
2311 - X509_LOOKUP_store()
2312 - X509_STORE_load_file()
2313 - X509_STORE_load_path()
2314 - X509_STORE_load_store()
2315 - SSL_add_store_cert_subjects_to_stack()
2316 - SSL_CTX_set_default_verify_store()
2317 - SSL_CTX_load_verify_file()
2318 - SSL_CTX_load_verify_dir()
2319 - SSL_CTX_load_verify_store()
2323 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2324 The presence of this system service is determined at run-time.
2328 * Added functionality to create an EVP_PKEY context based on data
2329 for methods from providers. This takes an algorithm name and a
2330 property query string and simply stores them, with the intent
2331 that any operation that uses this context will use those strings
2332 to fetch the needed methods implicitly, thereby making the port
2333 of application written for pre-3.0 OpenSSL easier.
2337 * The undocumented function NCONF_WIN32() has been deprecated; for
2338 conversion details see the HISTORY section of doc/man5/config.pod
2342 * Introduced the new functions EVP_DigestSignInit_ex() and
2343 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2344 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2345 pages for further details.
2349 * Over two thousand fixes were made to the documentation, including:
2350 adding missing command flags, better style conformance, documentation
2353 *Rich Salz, Richard Levitte*
2355 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2356 X25519, X448, Ed25519 and Ed448.
2360 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2365 * Deprecated the public definition of `ERR_STATE` as well as the function
2366 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2371 * Added ERR functionality to give callers access to the stored function
2372 names that have replaced the older function code based functions.
2374 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2375 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2376 ERR_peek_error_all() and ERR_peek_last_error_all().
2378 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2379 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2380 ERR_func_error_string().
2384 * Extended testing to be verbose for failing tests only. The make variables
2385 VERBOSE_FAILURE or VF can be used to enable this:
2387 $ make VF=1 test # Unix
2388 $ mms /macro=(VF=1) test ! OpenVMS
2389 $ nmake VF=1 test # Windows
2393 * Added the `-copy_extensions` option to the `x509` command for use with
2394 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2395 all extensions in the request are copied to the certificate or vice versa.
2397 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2399 * Added the `-copy_extensions` option to the `req` command for use with
2400 `-x509`. When given with the `copy` or `copyall` argument,
2401 all extensions in the certification request are copied to the certificate.
2405 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2406 they generate are by default RFC 5280 compliant in the following sense:
2407 There is a subjectKeyIdentifier extension with a hash value of the public key
2408 and for not self-signed certs there is an authorityKeyIdentifier extension
2409 with a keyIdentifier field or issuer information identifying the signing key.
2410 This is done unless some configuration overrides the new default behavior,
2411 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2415 * Added several checks to `X509_verify_cert()` according to requirements in
2416 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2417 (which may be done by using the CLI option `-x509_strict`):
2418 * The basicConstraints of CA certificates must be marked critical.
2419 * CA certificates must explicitly include the keyUsage extension.
2420 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2421 * The issuer name of any certificate must not be empty.
2422 * The subject name of CA certs, certs with keyUsage crlSign,
2423 and certs without subjectAlternativeName must not be empty.
2424 * If a subjectAlternativeName extension is given it must not be empty.
2425 * The signatureAlgorithm field and the cert signature must be consistent.
2426 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2427 must not be marked critical.
2428 * The authorityKeyIdentifier must be given for X.509v3 certs
2429 unless they are self-signed.
2430 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2434 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2435 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2439 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2440 used even when parsing explicit parameters, when loading a encoded key
2441 or calling `EC_GROUP_new_from_ecpkparameters()`/
2442 `EC_GROUP_new_from_ecparameters()`.
2443 This prevents bypass of security hardening and performance gains,
2444 especially for curves with specialized EC_METHODs.
2445 By default, if a key encoded with explicit parameters is loaded and later
2446 encoded, the output is still encoded with explicit parameters, even if
2447 internally a "named" EC_GROUP is used for computation.
2451 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2452 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2453 NULL. After this change, only the cofactor parameter can be NULL. It also
2454 does some minimal sanity checks on the passed order.
2459 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2460 An attack is simple, if the first CMS_recipientInfo is valid but the
2461 second CMS_recipientInfo is chosen ciphertext. If the second
2462 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2463 encryption key will be replaced by garbage, and the message cannot be
2464 decoded, but if the RSA decryption fails, the correct encryption key is
2465 used and the recipient will not notice the attack.
2466 As a work around for this potential attack the length of the decrypted
2467 key must be equal to the cipher default key length, in case the
2468 certificate is not given and all recipientInfo are tried out.
2469 The old behaviour can be re-enabled in the CMS code by setting the
2470 CMS_DEBUG_DECRYPT flag.
2474 * Early start up entropy quality from the DEVRANDOM seed source has been
2475 improved for older Linux systems. The RAND subsystem will wait for
2476 /dev/random to be producing output before seeding from /dev/urandom.
2477 The seeded state is stored for future library initialisations using
2478 a system global shared memory segment. The shared memory identifier
2479 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2480 the desired value. The default identifier is 114.
2484 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2485 when primes for RSA keys are computed.
2486 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2487 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2488 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2489 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2490 This avoids possible fingerprinting of newly generated RSA modules.
2494 * Correct the extended master secret constant on EBCDIC systems. Without this
2495 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2496 negotiate EMS will fail. Unfortunately this also means that TLS connections
2497 between EBCDIC systems with this fix, and EBCDIC systems without this
2498 fix will fail if they negotiate EMS.
2502 * Changed the library initialisation so that the config file is now loaded
2503 by default. This was already the case for libssl. It now occurs for both
2504 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2505 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2509 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2510 where the former acts as a replacement for `ERR_put_error()`, and the
2511 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2512 `ERR_raise_data()` adds more flexibility by taking a format string and
2513 an arbitrary number of arguments following it, to be processed with
2518 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2519 to check if a named provider is loaded and available. When called, it
2520 will also activate all fallback providers if such are still present.
2524 * Enforce a minimum DH modulus size of 512 bits.
2528 * Changed DH parameters to generate the order q subgroup instead of 2q.
2529 Previously generated DH parameters are still accepted by DH_check
2530 but DH_generate_key works around that by clearing bit 0 of the
2531 private key for those. This avoids leaking bit 0 of the private key.
2535 * Significantly reduce secure memory usage by the randomness pools.
2539 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2544 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2545 algorithms. An implementation of a key exchange algorithm can be obtained
2546 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2547 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2548 the older EVP_PKEY_derive_init() function. See the man pages for the new
2549 functions for further details.
2553 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2557 * Removed the function names from error messages and deprecated the
2562 * Removed NextStep support and the macro OPENSSL_UNISTD
2566 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2567 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2568 Also removed "export var as function" capability; we do not export
2569 variables, only functions.
2573 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2574 an error and 1 indicating success. In previous versions of OpenSSL this
2575 was a void type. If a key was set longer than the maximum possible this
2580 * Support SM2 signing and verification schemes with X509 certificate.
2584 * Use SHA256 as the default digest for TS query in the `ts` app.
2588 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2592 * Default cipher lists/suites are now available via a function, the
2593 #defines are deprecated.
2597 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2598 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2599 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2603 * Join the directories crypto/x509 and crypto/x509v3
2607 * Added command 'openssl kdf' that uses the EVP_KDF API.
2611 * Added command 'openssl mac' that uses the EVP_MAC API.
2615 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2616 as default directories. Also added the command 'openssl info'
2617 for scripting purposes.
2621 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2626 * Add prediction resistance to the DRBG reseeding process.
2630 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2631 mandated by IEEE Std 1619-2018.
2635 * Added newline escaping functionality to a filename when using openssl dgst.
2636 This output format is to replicate the output format found in the `*sum`
2637 checksum programs. This aims to preserve backward compatibility.
2639 *Matt Eaton, Richard Levitte, and Paul Dale*
2641 * Removed the heartbeat message in DTLS feature, as it has very
2642 little usage and doesn't seem to fulfill a valuable purpose.
2643 The configuration option is now deprecated.
2647 * Changed the output of 'openssl {digestname} < file' to display the
2648 digest name in its output.
2652 * Added a new generic trace API which provides support for enabling
2653 instrumentation through trace output.
2655 *Richard Levitte & Matthias St. Pierre*
2657 * Added build tests for C++. These are generated files that only do one
2658 thing, to include one public OpenSSL head file each. This tests that
2659 the public header files can be usefully included in a C++ application.
2661 This test isn't enabled by default. It can be enabled with the option
2662 'enable-buildtest-c++'.
2666 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2670 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2674 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2678 * Added KMAC to EVP_MAC.
2682 * Added property based algorithm implementation selection framework to
2687 * Added SCA hardening for modular field inversion in EC_GROUP through
2688 a new dedicated field_inv() pointer in EC_METHOD.
2689 This also addresses a leakage affecting conversions from projective
2690 to affine coordinates.
2692 *Billy Bob Brumley, Nicola Tuveri*
2694 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2695 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2696 those algorithms that were already supported through the EVP_PKEY API
2697 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2698 and scrypt are now wrappers that call EVP_KDF.
2702 * Build devcrypto engine as a dynamic engine.
2704 *Eneas U de Queiroz*
2706 * Add keyed BLAKE2 to EVP_MAC.
2710 * Fix a bug in the computation of the endpoint-pair shared secret used
2711 by DTLS over SCTP. This breaks interoperability with older versions
2712 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2713 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2714 interoperability with such broken implementations. However, enabling
2715 this switch breaks interoperability with correct implementations.
2717 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2718 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2722 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2726 * Changed the license to the Apache License v2.0.
2730 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2732 - Major releases (indicated by incrementing the MAJOR release number)
2733 may introduce incompatible API/ABI changes.
2734 - Minor releases (indicated by incrementing the MINOR release number)
2735 may introduce new features but retain API/ABI compatibility.
2736 - Patch releases (indicated by incrementing the PATCH number)
2737 are intended for bug fixes and other improvements of existing
2738 features only (like improving performance or adding documentation)
2739 and retain API/ABI compatibility.
2743 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2747 * Remove the 'dist' target and add a tarball building script. The
2748 'dist' target has fallen out of use, and it shouldn't be
2749 necessary to configure just to create a source distribution.
2753 * Recreate the OS390-Unix config target. It no longer relies on a
2754 special script like it did for OpenSSL pre-1.1.0.
2758 * Instead of having the source directories listed in Configure, add
2759 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2764 * Add GMAC to EVP_MAC.
2768 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2772 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2773 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2774 to facilitate the continued use of MACs through raw private keys in
2775 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2779 * Deprecate ECDH_KDF_X9_62().
2783 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2784 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2785 are retained for backwards compatibility.
2789 * AES-XTS mode now enforces that its two keys are different to mitigate
2790 the attacked described in "Efficient Instantiations of Tweakable
2791 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2792 Details of this attack can be obtained from:
2793 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2797 * Rename the object files, i.e. give them other names than in previous
2798 versions. Their names now include the name of the final product, as
2799 well as its type mnemonic (bin, lib, shlib).
2803 * Added new option for 'openssl list', '-objects', which will display the
2804 list of built in objects, i.e. OIDs with names.
2808 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2809 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2814 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2815 improves application performance by removing data copies and providing
2816 applications with zero-copy system calls such as sendfile and splice.
2820 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2824 * `PKCS12_parse` now maintains the order of the parsed certificates
2825 when outputting them via `*ca` (rather than reversing it).
2829 * Deprecated pthread fork support methods.
2833 * Added support for FFDHE key exchange in TLS 1.3.
2837 * Added a new concept for OpenSSL plugability: providers. This
2838 functionality is designed to replace the ENGINE API and ENGINE
2839 implementations, and to be much more dynamic, allowing provider
2840 authors to introduce new algorithms among other things, as long as
2841 there's an API that supports the algorithm type.
2843 With this concept comes a new core API for interaction between
2844 libcrypto and provider implementations. Public libcrypto functions
2845 that want to use providers do so through this core API.
2847 The main documentation for this core API is found in
2848 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2849 refer to other manuals describing the API specific for supported
2850 algorithm types (also called operations).
2857 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2859 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2861 * Avoid loading of a dynamic engine twice.
2865 * Prioritise DANE TLSA issuer certs over peer certs
2869 * Fixed random API for MacOS prior to 10.12
2871 These MacOS versions don't support the CommonCrypto APIs
2875 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2877 * Fixed an SM2 Decryption Buffer Overflow.
2879 In order to decrypt SM2 encrypted data an application is expected to
2880 call the API function EVP_PKEY_decrypt(). Typically an application will
2881 call this function twice. The first time, on entry, the "out" parameter
2882 can be NULL and, on exit, the "outlen" parameter is populated with the
2883 buffer size required to hold the decrypted plaintext. The application
2884 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2885 again, but this time passing a non-NULL value for the "out" parameter.
2887 A bug in the implementation of the SM2 decryption code means that the
2888 calculation of the buffer size required to hold the plaintext returned
2889 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2890 size required by the second call. This can lead to a buffer overflow
2891 when EVP_PKEY_decrypt() is called by the application a second time with
2892 a buffer that is too small.
2894 A malicious attacker who is able present SM2 content for decryption to
2895 an application could cause attacker chosen data to overflow the buffer
2896 by up to a maximum of 62 bytes altering the contents of other data held
2897 after the buffer, possibly changing application behaviour or causing
2898 the application to crash. The location of the buffer is application
2899 dependent but is typically heap allocated.
2904 * Fixed various read buffer overruns processing ASN.1 strings
2906 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2907 structure which contains a buffer holding the string data and a field
2908 holding the buffer length. This contrasts with normal C strings which
2909 are represented as a buffer for the string data which is terminated
2910 with a NUL (0) byte.
2912 Although not a strict requirement, ASN.1 strings that are parsed using
2913 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2914 well as any string whose value has been set with the ASN1_STRING_set()
2915 function will additionally NUL terminate the byte array in the
2916 ASN1_STRING structure.
2918 However, it is possible for applications to directly construct valid
2919 ASN1_STRING structures which do not NUL terminate the byte array by
2920 directly setting the "data" and "length" fields in the ASN1_STRING
2921 array. This can also happen by using the ASN1_STRING_set0() function.
2923 Numerous OpenSSL functions that print ASN.1 data have been found to
2924 assume that the ASN1_STRING byte array will be NUL terminated, even
2925 though this is not guaranteed for strings that have been directly
2926 constructed. Where an application requests an ASN.1 structure to be
2927 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2928 been directly constructed by the application without NUL terminating
2929 the "data" field, then a read buffer overrun can occur.
2931 The same thing can also occur during name constraints processing
2932 of certificates (for example if a certificate has been directly
2933 constructed by the application instead of loading it via the OpenSSL
2934 parsing functions, and the certificate contains non NUL terminated
2935 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2936 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2938 If a malicious actor can cause an application to directly construct an
2939 ASN1_STRING and then process it through one of the affected OpenSSL
2940 functions then this issue could be hit. This might result in a crash
2941 (causing a Denial of Service attack). It could also result in the
2942 disclosure of private memory contents (such as private keys, or
2943 sensitive plaintext).
2948 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2950 * Fixed a problem with verifying a certificate chain when using the
2951 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2952 the certificates present in a certificate chain. It is not set by default.
2954 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2955 the chain that have explicitly encoded elliptic curve parameters was added
2956 as an additional strict check.
2958 An error in the implementation of this check meant that the result of a
2959 previous check to confirm that certificates in the chain are valid CA
2960 certificates was overwritten. This effectively bypasses the check
2961 that non-CA certificates must not be able to issue other certificates.
2963 If a "purpose" has been configured then there is a subsequent opportunity
2964 for checks that the certificate is a valid CA. All of the named "purpose"
2965 values implemented in libcrypto perform this check. Therefore, where
2966 a purpose is set the certificate chain will still be rejected even when the
2967 strict flag has been used. A purpose is set by default in libssl client and
2968 server certificate verification routines, but it can be overridden or
2969 removed by an application.
2971 In order to be affected, an application must explicitly set the
2972 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2973 for the certificate verification or, in the case of TLS client or server
2974 applications, override the default purpose.
2979 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2980 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2981 renegotiation ClientHello omits the signature_algorithms extension (where it
2982 was present in the initial ClientHello), but includes a
2983 signature_algorithms_cert extension then a NULL pointer dereference will
2984 result, leading to a crash and a denial of service attack.
2986 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2987 (which is the default configuration). OpenSSL TLS clients are not impacted by
2991 *Peter Kästle and Samuel Sapalski*
2993 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2995 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2996 create a unique hash value based on the issuer and serial number data
2997 contained within an X509 certificate. However, it was failing to correctly
2998 handle any errors that may occur while parsing the issuer field (which might
2999 occur if the issuer field is maliciously constructed). This may subsequently
3000 result in a NULL pointer deref and a crash leading to a potential denial of
3006 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3007 padding mode to correctly check for rollback attacks. This is considered a
3008 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3013 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3014 functions. Previously they could overflow the output length argument in some
3015 cases where the input length is close to the maximum permissible length for
3016 an integer on the platform. In such cases the return value from the function
3017 call would be 1 (indicating success), but the output length value would be
3018 negative. This could cause applications to behave incorrectly or crash.
3023 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3024 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3025 could be exploited in a side channel attack to recover the password. Since
3026 the attack is local host only this is outside of the current OpenSSL
3027 threat model and therefore no CVE is assigned.
3029 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3034 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3036 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3037 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3038 If an attacker can control both items being compared then this could lead
3039 to a possible denial of service attack. OpenSSL itself uses the
3040 GENERAL_NAME_cmp function for two purposes:
3041 1) Comparing CRL distribution point names between an available CRL and a
3042 CRL distribution point embedded in an X509 certificate
3043 2) When verifying that a timestamp response token signer matches the
3044 timestamp authority name (exposed via the API functions
3045 TS_RESP_verify_response and TS_RESP_verify_token)
3050 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3052 * Certificates with explicit curve parameters are now disallowed in
3053 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3057 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3058 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3059 conversely, silently ignore DTLS protocol version bounds when configuring
3060 TLS-based contexts. The commands can be repeated to set bounds of both
3061 types. The same applies with the corresponding "min_protocol" and
3062 "max_protocol" command-line switches, in case some application uses both TLS
3065 SSL_CTX instances that are created for a fixed protocol version (e.g.
3066 TLSv1_server_method()) also silently ignore version bounds. Previously
3067 attempts to apply bounds to these protocol versions would result in an
3068 error. Now only the "version-flexible" SSL_CTX instances are subject to
3069 limits in configuration files in command-line options.
3073 * Handshake now fails if Extended Master Secret extension is dropped
3078 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3080 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3082 * Fixed segmentation fault in SSL_check_chain()
3083 Server or client applications that call the SSL_check_chain() function
3084 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3085 dereference as a result of incorrect handling of the
3086 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3087 or unrecognised signature algorithm is received from the peer. This could
3088 be exploited by a malicious peer in a Denial of Service attack.
3093 * Added AES consttime code for no-asm configurations
3094 an optional constant time support for AES was added
3095 when building openssl for no-asm.
3096 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3097 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3098 At this time this feature is by default disabled.
3099 It will be enabled by default in 3.0.
3103 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3105 * Revert the change of EOF detection while reading in libssl to avoid
3106 regressions in applications depending on the current way of reporting
3107 the EOF. As the existing method is not fully accurate the change to
3108 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3109 branch and will be present in the 3.0 release.
3113 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3114 when primes for RSA keys are computed.
3115 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3116 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3117 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3118 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3119 This avoids possible fingerprinting of newly generated RSA modules.
3123 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3125 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3126 while reading in libssl then we would report an error back to the
3127 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3128 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3129 therefore give a hint as to what went wrong.
3133 * Check that ed25519 and ed448 are allowed by the security level. Previously
3134 signature algorithms not using an MD were not being checked that they were
3135 allowed by the security level.
3139 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3140 was not quite right. The behaviour was not consistent between resumption
3141 and normal handshakes, and also not quite consistent with historical
3142 behaviour. The behaviour in various scenarios has been clarified and
3143 it has been updated to make it match historical behaviour as closely as
3148 * *[VMS only]* The header files that the VMS compilers include automatically,
3149 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3150 that the C++ compiler doesn't understand. This is a shortcoming in the
3151 compiler, but can be worked around with `__cplusplus` guards.
3153 C++ applications that use OpenSSL libraries must be compiled using the
3154 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3155 functions. Otherwise, only functions with symbols of less than 31
3156 characters can be used, as the linker will not be able to successfully
3157 resolve symbols with longer names.
3161 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3162 The presence of this system service is determined at run-time.
3166 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3171 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3173 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3174 number generator (RNG). This was intended to include protection in the
3175 event of a fork() system call in order to ensure that the parent and child
3176 processes did not share the same RNG state. However, this protection was not
3177 being used in the default case.
3179 A partial mitigation for this issue is that the output from a high
3180 precision timer is mixed into the RNG state so the likelihood of a parent
3181 and child process sharing state is significantly reduced.
3183 If an application already calls OPENSSL_init_crypto() explicitly using
3184 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3187 *Matthias St. Pierre*
3189 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3190 used even when parsing explicit parameters, when loading a encoded key
3191 or calling `EC_GROUP_new_from_ecpkparameters()`/
3192 `EC_GROUP_new_from_ecparameters()`.
3193 This prevents bypass of security hardening and performance gains,
3194 especially for curves with specialized EC_METHODs.
3195 By default, if a key encoded with explicit parameters is loaded and later
3196 encoded, the output is still encoded with explicit parameters, even if
3197 internally a "named" EC_GROUP is used for computation.
3201 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3202 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3203 NULL. After this change, only the cofactor parameter can be NULL. It also
3204 does some minimal sanity checks on the passed order.
3209 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3210 An attack is simple, if the first CMS_recipientInfo is valid but the
3211 second CMS_recipientInfo is chosen ciphertext. If the second
3212 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3213 encryption key will be replaced by garbage, and the message cannot be
3214 decoded, but if the RSA decryption fails, the correct encryption key is
3215 used and the recipient will not notice the attack.
3216 As a work around for this potential attack the length of the decrypted
3217 key must be equal to the cipher default key length, in case the
3218 certificate is not given and all recipientInfo are tried out.
3219 The old behaviour can be re-enabled in the CMS code by setting the
3220 CMS_DEBUG_DECRYPT flag.
3225 * Early start up entropy quality from the DEVRANDOM seed source has been
3226 improved for older Linux systems. The RAND subsystem will wait for
3227 /dev/random to be producing output before seeding from /dev/urandom.
3228 The seeded state is stored for future library initialisations using
3229 a system global shared memory segment. The shared memory identifier
3230 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3231 the desired value. The default identifier is 114.
3235 * Correct the extended master secret constant on EBCDIC systems. Without this
3236 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3237 negotiate EMS will fail. Unfortunately this also means that TLS connections
3238 between EBCDIC systems with this fix, and EBCDIC systems without this
3239 fix will fail if they negotiate EMS.
3243 * Use Windows installation paths in the mingw builds
3245 Mingw isn't a POSIX environment per se, which means that Windows
3246 paths should be used for installation.
3251 * Changed DH_check to accept parameters with order q and 2q subgroups.
3252 With order 2q subgroups the bit 0 of the private key is not secret
3253 but DH_generate_key works around that by clearing bit 0 of the
3254 private key for those. This avoids leaking bit 0 of the private key.
3258 * Significantly reduce secure memory usage by the randomness pools.
3262 * Revert the DEVRANDOM_WAIT feature for Linux systems
3264 The DEVRANDOM_WAIT feature added a select() call to wait for the
3265 /dev/random device to become readable before reading from the
3266 /dev/urandom device.
3268 It turned out that this change had negative side effects on
3269 performance which were not acceptable. After some discussion it
3270 was decided to revert this feature and leave it up to the OS
3271 resp. the platform maintainer to ensure a proper initialization
3272 during early boot time.
3274 *Matthias St. Pierre*
3276 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3278 * Add build tests for C++. These are generated files that only do one
3279 thing, to include one public OpenSSL head file each. This tests that
3280 the public header files can be usefully included in a C++ application.
3282 This test isn't enabled by default. It can be enabled with the option
3283 'enable-buildtest-c++'.
3287 * Enable SHA3 pre-hashing for ECDSA and DSA.
3291 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3292 This changes the size when using the `genpkey` command when no size is given.
3293 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3294 generation commands to use 2048 bits by default.
3298 * Reorganize the manual pages to consistently have RETURN VALUES,
3299 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3300 util/fix-doc-nits accordingly.
3302 *Paul Yang, Joshua Lock*
3304 * Add the missing accessor EVP_PKEY_get0_engine()
3308 * Have commands like `s_client` and `s_server` output the signature scheme
3309 along with other cipher suite parameters when debugging.
3313 * Make OPENSSL_config() error agnostic again.
3317 * Do the error handling in RSA decryption constant time.
3321 * Prevent over long nonces in ChaCha20-Poly1305.
3323 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3324 for every encryption operation. RFC 7539 specifies that the nonce value
3325 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3326 and front pads the nonce with 0 bytes if it is less than 12
3327 bytes. However it also incorrectly allows a nonce to be set of up to 16
3328 bytes. In this case only the last 12 bytes are significant and any
3329 additional leading bytes are ignored.
3331 It is a requirement of using this cipher that nonce values are
3332 unique. Messages encrypted using a reused nonce value are susceptible to
3333 serious confidentiality and integrity attacks. If an application changes
3334 the default nonce length to be longer than 12 bytes and then makes a
3335 change to the leading bytes of the nonce expecting the new value to be a
3336 new unique nonce then such an application could inadvertently encrypt
3337 messages with a reused nonce.
3339 Additionally the ignored bytes in a long nonce are not covered by the
3340 integrity guarantee of this cipher. Any application that relies on the
3341 integrity of these ignored leading bytes of a long nonce may be further
3342 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3343 is safe because no such use sets such a long nonce value. However user
3344 applications that use this cipher directly and set a non-default nonce
3345 length to be longer than 12 bytes may be vulnerable.
3347 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3353 * Add DEVRANDOM_WAIT feature for Linux systems
3355 On older Linux systems where the getrandom() system call is not available,
3356 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3357 Contrary to getrandom(), the /dev/urandom device will not block during
3358 early boot when the kernel CSPRNG has not been seeded yet.
3360 To mitigate this known weakness, use select() to wait for /dev/random to
3361 become readable before reading from /dev/urandom.
3363 * Ensure that SM2 only uses SM3 as digest algorithm
3367 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3369 * Change the info callback signals for the start and end of a post-handshake
3370 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3371 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3372 confused by this and assume that a TLSv1.2 renegotiation has started. This
3373 can break KeyUpdate handling. Instead we no longer signal the start and end
3374 of a post handshake message exchange (although the messages themselves are
3375 still signalled). This could break some applications that were expecting
3376 the old signals. However without this KeyUpdate is not usable for many
3381 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3383 * Timing vulnerability in DSA signature generation
3385 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3386 timing side channel attack. An attacker could use variations in the signing
3387 algorithm to recover the private key.
3389 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3394 * Timing vulnerability in ECDSA signature generation
3396 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3397 timing side channel attack. An attacker could use variations in the signing
3398 algorithm to recover the private key.
3400 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3405 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3406 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3407 of two gigabytes and the error handling improved.
3409 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3410 categorized as a normal bug, not a security issue, because the DRBG reseeds
3411 automatically and is fully functional even without additional randomness
3412 provided by the application.
3414 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3416 * Add a new ClientHello callback. Provides a callback interface that gives
3417 the application the ability to adjust the nascent SSL object at the
3418 earliest stage of ClientHello processing, immediately after extensions have
3419 been collected but before they have been processed. In particular, this
3420 callback can adjust the supported TLS versions in response to the contents
3425 * Add SM2 base algorithm support.
3429 * s390x assembly pack: add (improved) hardware-support for the following
3430 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3431 aes-cfb/cfb8, aes-ecb.
3435 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3436 parameter is no longer accepted, as it leads to a corrupt table. NULL
3437 pem_str is reserved for alias entries only.
3441 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3442 step for prime curves. The new implementation is based on formulae from
3443 differential addition-and-doubling in homogeneous projective coordinates
3444 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3445 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3446 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3447 to work in projective coordinates.
3449 *Billy Bob Brumley, Nicola Tuveri*
3451 * Change generating and checking of primes so that the error rate of not
3452 being prime depends on the intended use based on the size of the input.
3453 For larger primes this will result in more rounds of Miller-Rabin.
3454 The maximal error rate for primes with more than 1080 bits is lowered
3457 *Kurt Roeckx, Annie Yousar*
3459 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3463 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3464 moving between systems, and to avoid confusion when a Windows build is
3465 done with mingw vs with MSVC. For POSIX installs, there's still a
3466 symlink or copy named 'tsget' to avoid that confusion as well.
3470 * Revert blinding in ECDSA sign and instead make problematic addition
3471 length-invariant. Switch even to fixed-length Montgomery multiplication.
3475 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3476 step for binary curves. The new implementation is based on formulae from
3477 differential addition-and-doubling in mixed Lopez-Dahab projective
3478 coordinates, modified to independently blind the operands.
3480 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3482 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3483 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3484 EC_METHODs to implement their own specialized "ladder step", to take
3485 advantage of more favorable coordinate systems or more efficient
3486 differential addition-and-doubling algorithms.
3488 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3490 * Modified the random device based seed sources to keep the relevant
3491 file descriptors open rather than reopening them on each access.
3492 This allows such sources to operate in a chroot() jail without
3493 the associated device nodes being available. This behaviour can be
3494 controlled using RAND_keep_random_devices_open().
3498 * Numerous side-channel attack mitigations have been applied. This may have
3499 performance impacts for some algorithms for the benefit of improved
3500 security. Specific changes are noted in this change log by their respective
3505 * AIX shared library support overhaul. Switch to AIX "natural" way of
3506 handling shared libraries, which means collecting shared objects of
3507 different versions and bitnesses in one common archive. This allows to
3508 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3509 doesn't affect the way 3rd party applications are linked, only how
3510 multi-version installation is managed.
3514 * Make ec_group_do_inverse_ord() more robust and available to other
3515 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3516 mitigations are applied to the fallback BN_mod_inverse().
3517 When using this function rather than BN_mod_inverse() directly, new
3518 EC cryptosystem implementations are then safer-by-default.
3522 * Add coordinate blinding for EC_POINT and implement projective
3523 coordinate blinding for generic prime curves as a countermeasure to
3524 chosen point SCA attacks.
3526 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3528 * Add blinding to ECDSA and DSA signatures to protect against side channel
3529 attacks discovered by Keegan Ryan (NCC Group).
3533 * Enforce checking in the `pkeyutl` command to ensure that the input
3534 length does not exceed the maximum supported digest length when performing
3535 a sign, verify or verifyrecover operation.
3539 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3540 I/O in combination with something like select() or poll() will hang. This
3541 can be turned off again using SSL_CTX_clear_mode().
3542 Many applications do not properly handle non-application data records, and
3543 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3544 around the problems in those applications, but can also break some.
3545 It's recommended to read the manpages about SSL_read(), SSL_write(),
3546 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3547 SSL_CTX_set_read_ahead() again.
3551 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3552 now allow empty (zero character) pass phrases.
3556 * Apply blinding to binary field modular inversion and remove patent
3557 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3561 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3562 binary and prime elliptic curves.
3566 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3567 constant time fixed point multiplication.
3571 * Revise elliptic curve scalar multiplication with timing attack
3572 defenses: ec_wNAF_mul redirects to a constant time implementation
3573 when computing fixed point and variable point multiplication (which
3574 in OpenSSL are mostly used with secret scalars in keygen, sign,
3575 ECDH derive operations).
3576 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3579 * Updated CONTRIBUTING
3583 * Updated DRBG / RAND to request nonce and additional low entropy
3584 randomness from the system.
3586 *Matthias St. Pierre*
3588 * Updated 'openssl rehash' to use OpenSSL consistent default.
3592 * Moved the load of the ssl_conf module to libcrypto, which helps
3593 loading engines that libssl uses before libssl is initialised.
3597 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3601 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3603 *Ingo Schwarze, Rich Salz*
3605 * Added output of accepting IP address and port for 'openssl s_server'
3609 * Added a new API for TLSv1.3 ciphersuites:
3610 SSL_CTX_set_ciphersuites()
3611 SSL_set_ciphersuites()
3615 * Memory allocation failures consistently add an error to the error
3620 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3621 in libcrypto when run as setuid/setgid.
3625 * Load any config file by default when libssl is used.
3629 * Added new public header file <openssl/rand_drbg.h> and documentation
3630 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3632 *Matthias St. Pierre*
3634 * QNX support removed (cannot find contributors to get their approval
3635 for the license change).
3639 * TLSv1.3 replay protection for early data has been implemented. See the
3640 SSL_read_early_data() man page for further details.
3644 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3645 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3646 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3647 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3648 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3649 configuration has been separated out. See the ciphers man page or the
3650 SSL_CTX_set_ciphersuites() man page for more information.
3654 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3655 in responder mode now supports the new "-multi" option, which
3656 spawns the specified number of child processes to handle OCSP
3657 requests. The "-timeout" option now also limits the OCSP
3658 responder's patience to wait to receive the full client request
3659 on a newly accepted connection. Child processes are respawned
3660 as needed, and the CA index file is automatically reloaded
3661 when changed. This makes it possible to run the "ocsp" responder
3662 as a long-running service, making the OpenSSL CA somewhat more
3663 feature-complete. In this mode, most diagnostic messages logged
3664 after entering the event loop are logged via syslog(3) rather than
3669 * Added support for X448 and Ed448. Heavily based on original work by
3674 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3675 objects loaded. This adds the functions OSSL_STORE_expect() and
3676 OSSL_STORE_find() as well as needed tools to construct searches and
3677 get the search data out of them.
3681 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3682 version of OpenSSL should review their configuration settings to ensure
3683 that they are still appropriate for TLSv1.3. For further information see:
3684 <https://wiki.openssl.org/index.php/TLS1.3>
3688 * Grand redesign of the OpenSSL random generator
3690 The default RAND method now utilizes an AES-CTR DRBG according to
3691 NIST standard SP 800-90Ar1. The new random generator is essentially
3692 a port of the default random generator from the OpenSSL FIPS 2.0
3693 object module. It is a hybrid deterministic random bit generator
3694 using an AES-CTR bit stream and which seeds and reseeds itself
3695 automatically using trusted system entropy sources.
3697 Some of its new features are:
3698 - Support for multiple DRBG instances with seed chaining.
3699 - The default RAND method makes use of a DRBG.
3700 - There is a public and private DRBG instance.
3701 - The DRBG instances are fork-safe.
3702 - Keep all global DRBG instances on the secure heap if it is enabled.
3703 - The public and private DRBG instance are per thread for lock free
3706 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3708 * Changed Configure so it only says what it does and doesn't dump
3709 so much data. Instead, ./configdata.pm should be used as a script
3710 to display all sorts of configuration data.
3714 * Added processing of "make variables" to Configure.
3718 * Added SHA512/224 and SHA512/256 algorithm support.
3722 * The last traces of Netware support, first removed in 1.1.0, have
3727 * Get rid of Makefile.shared, and in the process, make the processing
3728 of certain files (rc.obj, or the .def/.map/.opt files produced from
3729 the ordinal files) more visible and hopefully easier to trace and
3730 debug (or make silent).
3734 * Make it possible to have environment variable assignments as
3735 arguments to config / Configure.
3739 * Add multi-prime RSA (RFC 8017) support.
3743 * Add SM3 implemented according to GB/T 32905-2016
3744 *Jack Lloyd <jack.lloyd@ribose.com>,*
3745 *Ronald Tse <ronald.tse@ribose.com>,*
3746 *Erick Borsboom <erick.borsboom@ribose.com>*
3748 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3749 as documented in RFC6066.
3750 Based on a patch from Tomasz Moń
3752 *Filipe Raimundo da Silva*
3754 * Add SM4 implemented according to GB/T 32907-2016.
3755 *Jack Lloyd <jack.lloyd@ribose.com>,*
3756 *Ronald Tse <ronald.tse@ribose.com>,*
3757 *Erick Borsboom <erick.borsboom@ribose.com>*
3759 * Reimplement -newreq-nodes and ERR_error_string_n; the
3760 original author does not agree with the license change.
3764 * Add ARIA AEAD TLS support.
3768 * Some macro definitions to support VS6 have been removed. Visual
3769 Studio 6 has not worked since 1.1.0
3773 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3774 without clearing the errors.
3778 * Add "atfork" functions. If building on a system that without
3779 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3780 requirements. The RAND facility now uses/requires this.
3788 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3789 not possible to disable entirely. However, it's still possible to
3790 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3793 To disable, configure with 'no-ui-console'. 'no-ui' is still
3794 possible to use as an alias. Check at compile time with the
3795 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3796 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3800 * Add a STORE module, which implements a uniform and URI based reader of
3801 stores that can contain keys, certificates, CRLs and numerous other
3802 objects. The main API is loosely based on a few stdio functions,
3803 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3804 OSSL_STORE_error and OSSL_STORE_close.
3805 The implementation uses backends called "loaders" to implement arbitrary
3806 URI schemes. There is one built in "loader" for the 'file' scheme.
3810 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3811 then adjusted to work on FreeBSD 8.4 as well.
3812 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3813 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3817 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3818 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3819 error code calls like this:
3821 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3823 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3824 that can be encoded in C. For the foreseeable future, this will only
3827 *Richard Levitte and Tim Hudson*
3829 * Removed BSD cryptodev engine.
3833 * Add a build target 'build_all_generated', to build all generated files
3834 and only that. This can be used to prepare everything that requires
3835 things like perl for a system that lacks perl and then move everything
3836 to that system and do the rest of the build there.
3840 * In the UI interface, make it possible to duplicate the user data. This
3841 can be used by engines that need to retain the data for a longer time
3842 than just the call where this user data is passed.
3846 * Ignore the '-named_curve auto' value for compatibility of applications
3849 *Tomáš Mráz <tmraz@fedoraproject.org>*
3851 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3852 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3853 alerts across multiple records (some of which could be empty). In practice
3854 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3855 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3856 support this at all. Supporting it adds significant complexity to the
3857 record layer, and its removal is unlikely to cause interoperability
3862 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3863 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3864 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3869 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3870 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3872 *Richard Levitte, Andy Polyakov*
3874 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3879 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3880 platform rather than 'mingw'.
3884 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3885 success if they are asked to add an object which already exists
3886 in the store. This change cascades to other functions which load
3887 certificates and CRLs.
3891 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3892 facilitate stack unwinding even from assembly subroutines.
3896 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3897 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3901 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3902 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3903 which is the minimum version we support.
3907 * Certificate time validation (X509_cmp_time) enforces stricter
3908 compliance with RFC 5280. Fractional seconds and timezone offsets
3909 are no longer allowed.
3913 * Add support for ARIA
3917 * s_client will now send the Server Name Indication (SNI) extension by
3918 default unless the new "-noservername" option is used. The server name is
3919 based on the host provided to the "-connect" option unless overridden by
3920 using "-servername".
3924 * Add support for SipHash
3928 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3929 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3930 prevent issues where no progress is being made and the peer continually
3931 sends unrecognised record types, using up resources processing them.
3935 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3936 using the algorithm defined in
3937 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3941 * Heartbeat support has been removed; the ABI is changed for now.
3943 *Richard Levitte, Rich Salz*
3945 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3949 * The RSA "null" method, which was partially supported to avoid patent
3950 issues, has been replaced to always returns NULL.
3957 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3959 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3960 used even when parsing explicit parameters, when loading a encoded key
3961 or calling `EC_GROUP_new_from_ecpkparameters()`/
3962 `EC_GROUP_new_from_ecparameters()`.
3963 This prevents bypass of security hardening and performance gains,
3964 especially for curves with specialized EC_METHODs.
3965 By default, if a key encoded with explicit parameters is loaded and later
3966 encoded, the output is still encoded with explicit parameters, even if
3967 internally a "named" EC_GROUP is used for computation.
3971 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3972 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3973 NULL. After this change, only the cofactor parameter can be NULL. It also
3974 does some minimal sanity checks on the passed order.
3979 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3980 An attack is simple, if the first CMS_recipientInfo is valid but the
3981 second CMS_recipientInfo is chosen ciphertext. If the second
3982 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3983 encryption key will be replaced by garbage, and the message cannot be
3984 decoded, but if the RSA decryption fails, the correct encryption key is
3985 used and the recipient will not notice the attack.
3986 As a work around for this potential attack the length of the decrypted
3987 key must be equal to the cipher default key length, in case the
3988 certificate is not given and all recipientInfo are tried out.
3989 The old behaviour can be re-enabled in the CMS code by setting the
3990 CMS_DEBUG_DECRYPT flag.
3995 * Use Windows installation paths in the mingw builds
3997 Mingw isn't a POSIX environment per se, which means that Windows
3998 paths should be used for installation.
4003 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4005 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4006 This changes the size when using the `genpkey` command when no size is given.
4007 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4008 generation commands to use 2048 bits by default.
4012 * Prevent over long nonces in ChaCha20-Poly1305.
4014 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4015 for every encryption operation. RFC 7539 specifies that the nonce value
4016 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4017 and front pads the nonce with 0 bytes if it is less than 12
4018 bytes. However it also incorrectly allows a nonce to be set of up to 16
4019 bytes. In this case only the last 12 bytes are significant and any
4020 additional leading bytes are ignored.
4022 It is a requirement of using this cipher that nonce values are
4023 unique. Messages encrypted using a reused nonce value are susceptible to
4024 serious confidentiality and integrity attacks. If an application changes
4025 the default nonce length to be longer than 12 bytes and then makes a
4026 change to the leading bytes of the nonce expecting the new value to be a
4027 new unique nonce then such an application could inadvertently encrypt
4028 messages with a reused nonce.
4030 Additionally the ignored bytes in a long nonce are not covered by the
4031 integrity guarantee of this cipher. Any application that relies on the
4032 integrity of these ignored leading bytes of a long nonce may be further
4033 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4034 is safe because no such use sets such a long nonce value. However user
4035 applications that use this cipher directly and set a non-default nonce
4036 length to be longer than 12 bytes may be vulnerable.
4038 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4044 * Added SCA hardening for modular field inversion in EC_GROUP through
4045 a new dedicated field_inv() pointer in EC_METHOD.
4046 This also addresses a leakage affecting conversions from projective
4047 to affine coordinates.
4049 *Billy Bob Brumley, Nicola Tuveri*
4051 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4052 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4056 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4060 * Remove the 'dist' target and add a tarball building script. The
4061 'dist' target has fallen out of use, and it shouldn't be
4062 necessary to configure just to create a source distribution.
4066 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4068 * Timing vulnerability in DSA signature generation
4070 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4071 timing side channel attack. An attacker could use variations in the signing
4072 algorithm to recover the private key.
4074 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4079 * Timing vulnerability in ECDSA signature generation
4081 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4082 timing side channel attack. An attacker could use variations in the signing
4083 algorithm to recover the private key.
4085 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4090 * Add coordinate blinding for EC_POINT and implement projective
4091 coordinate blinding for generic prime curves as a countermeasure to
4092 chosen point SCA attacks.
4094 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4096 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4098 * Client DoS due to large DH parameter
4100 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4101 malicious server can send a very large prime value to the client. This will
4102 cause the client to spend an unreasonably long period of time generating a
4103 key for this prime resulting in a hang until the client has finished. This
4104 could be exploited in a Denial Of Service attack.
4106 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4111 * Cache timing vulnerability in RSA Key Generation
4113 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4114 a cache timing side channel attack. An attacker with sufficient access to
4115 mount cache timing attacks during the RSA key generation process could
4116 recover the private key.
4118 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4119 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4124 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4125 parameter is no longer accepted, as it leads to a corrupt table. NULL
4126 pem_str is reserved for alias entries only.
4130 * Revert blinding in ECDSA sign and instead make problematic addition
4131 length-invariant. Switch even to fixed-length Montgomery multiplication.
4135 * Change generating and checking of primes so that the error rate of not
4136 being prime depends on the intended use based on the size of the input.
4137 For larger primes this will result in more rounds of Miller-Rabin.
4138 The maximal error rate for primes with more than 1080 bits is lowered
4141 *Kurt Roeckx, Annie Yousar*
4143 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4147 * Add blinding to ECDSA and DSA signatures to protect against side channel
4148 attacks discovered by Keegan Ryan (NCC Group).
4152 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4153 now allow empty (zero character) pass phrases.
4157 * Certificate time validation (X509_cmp_time) enforces stricter
4158 compliance with RFC 5280. Fractional seconds and timezone offsets
4159 are no longer allowed.
4163 * Fixed a text canonicalisation bug in CMS
4165 Where a CMS detached signature is used with text content the text goes
4166 through a canonicalisation process first prior to signing or verifying a
4167 signature. This process strips trailing space at the end of lines, converts
4168 line terminators to CRLF and removes additional trailing line terminators
4169 at the end of a file. A bug in the canonicalisation process meant that
4170 some characters, such as form-feed, were incorrectly treated as whitespace
4171 and removed. This is contrary to the specification (RFC5485). This fix
4172 could mean that detached text data signed with an earlier version of
4173 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4174 signed with a fixed OpenSSL may fail to verify with an earlier version of
4175 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4176 and use the "-binary" flag (for the "cms" command line application) or set
4177 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4181 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4183 * Constructed ASN.1 types with a recursive definition could exceed the stack
4185 Constructed ASN.1 types with a recursive definition (such as can be found
4186 in PKCS7) could eventually exceed the stack given malicious input with
4187 excessive recursion. This could result in a Denial Of Service attack. There
4188 are no such structures used within SSL/TLS that come from untrusted sources
4189 so this is considered safe.
4191 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4197 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4199 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4200 effectively reduced to only comparing the least significant bit of each
4201 byte. This allows an attacker to forge messages that would be considered as
4202 authenticated in an amount of tries lower than that guaranteed by the
4203 security claims of the scheme. The module can only be compiled by the
4204 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4206 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4212 * Add a build target 'build_all_generated', to build all generated files
4213 and only that. This can be used to prepare everything that requires
4214 things like perl for a system that lacks perl and then move everything
4215 to that system and do the rest of the build there.
4219 * Backport SSL_OP_NO_RENGOTIATION
4221 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4222 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4223 changes this is no longer possible in 1.1.0. Therefore, the new
4224 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4225 1.1.0 to provide equivalent functionality.
4227 Note that if an application built against 1.1.0h headers (or above) is run
4228 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4229 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4233 * Removed the OS390-Unix config target. It relied on a script that doesn't
4238 * rsaz_1024_mul_avx2 overflow bug on x86_64
4240 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4241 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4242 Analysis suggests that attacks against RSA and DSA as a result of this
4243 defect would be very difficult to perform and are not believed likely.
4244 Attacks against DH1024 are considered just feasible, because most of the
4245 work necessary to deduce information about a private key may be performed
4246 offline. The amount of resources required for such an attack would be
4247 significant. However, for an attack on TLS to be meaningful, the server
4248 would have to share the DH1024 private key among multiple clients, which is
4249 no longer an option since CVE-2016-0701.
4251 This only affects processors that support the AVX2 but not ADX extensions
4252 like Intel Haswell (4th generation).
4254 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4255 was originally found via the OSS-Fuzz project.
4260 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4262 * bn_sqrx8x_internal carry bug on x86_64
4264 There is a carry propagating bug in the x86_64 Montgomery squaring
4265 procedure. No EC algorithms are affected. Analysis suggests that attacks
4266 against RSA and DSA as a result of this defect would be very difficult to
4267 perform and are not believed likely. Attacks against DH are considered just
4268 feasible (although very difficult) because most of the work necessary to
4269 deduce information about a private key may be performed offline. The amount
4270 of resources required for such an attack would be very significant and
4271 likely only accessible to a limited number of attackers. An attacker would
4272 additionally need online access to an unpatched system using the target
4273 private key in a scenario with persistent DH parameters and a private
4274 key that is shared between multiple clients.
4276 This only affects processors that support the BMI1, BMI2 and ADX extensions
4277 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4279 This issue was reported to OpenSSL by the OSS-Fuzz project.
4284 * Malformed X.509 IPAddressFamily could cause OOB read
4286 If an X.509 certificate has a malformed IPAddressFamily extension,
4287 OpenSSL could do a one-byte buffer overread. The most likely result
4288 would be an erroneous display of the certificate in text format.
4290 This issue was reported to OpenSSL by the OSS-Fuzz project.
4295 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4297 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4298 platform rather than 'mingw'.
4302 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4303 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4304 which is the minimum version we support.
4308 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4310 * Encrypt-Then-Mac renegotiation crash
4312 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4313 negotiated where it was not in the original handshake (or vice-versa) then
4314 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4315 and servers are affected.
4317 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4322 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4324 * Truncated packet could crash via OOB read
4326 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4327 cipher is being used, then a truncated packet can cause that host to
4328 perform an out-of-bounds read, usually resulting in a crash.
4330 This issue was reported to OpenSSL by Robert Święcki of Google.
4335 * Bad (EC)DHE parameters cause a client crash
4337 If a malicious server supplies bad parameters for a DHE or ECDHE key
4338 exchange then this can result in the client attempting to dereference a
4339 NULL pointer leading to a client crash. This could be exploited in a Denial
4342 This issue was reported to OpenSSL by Guido Vranken.
4347 * BN_mod_exp may produce incorrect results on x86_64
4349 There is a carry propagating bug in the x86_64 Montgomery squaring
4350 procedure. No EC algorithms are affected. Analysis suggests that attacks
4351 against RSA and DSA as a result of this defect would be very difficult to
4352 perform and are not believed likely. Attacks against DH are considered just
4353 feasible (although very difficult) because most of the work necessary to
4354 deduce information about a private key may be performed offline. The amount
4355 of resources required for such an attack would be very significant and
4356 likely only accessible to a limited number of attackers. An attacker would
4357 additionally need online access to an unpatched system using the target
4358 private key in a scenario with persistent DH parameters and a private
4359 key that is shared between multiple clients. For example this can occur by
4360 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4361 similar to CVE-2015-3193 but must be treated as a separate problem.
4363 This issue was reported to OpenSSL by the OSS-Fuzz project.
4368 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4370 * ChaCha20/Poly1305 heap-buffer-overflow
4372 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4373 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4374 crash. This issue is not considered to be exploitable beyond a DoS.
4376 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4381 * CMS Null dereference
4383 Applications parsing invalid CMS structures can crash with a NULL pointer
4384 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4385 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4386 structure callback if an attempt is made to free certain invalid encodings.
4387 Only CHOICE structures using a callback which do not handle NULL value are
4390 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4395 * Montgomery multiplication may produce incorrect results
4397 There is a carry propagating bug in the Broadwell-specific Montgomery
4398 multiplication procedure that handles input lengths divisible by, but
4399 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4400 and DH private keys are impossible. This is because the subroutine in
4401 question is not used in operations with the private key itself and an input
4402 of the attacker's direct choice. Otherwise the bug can manifest itself as
4403 transient authentication and key negotiation failures or reproducible
4404 erroneous outcome of public-key operations with specially crafted input.
4405 Among EC algorithms only Brainpool P-512 curves are affected and one
4406 presumably can attack ECDH key negotiation. Impact was not analyzed in
4407 detail, because pre-requisites for attack are considered unlikely. Namely
4408 multiple clients have to choose the curve in question and the server has to
4409 share the private key among them, neither of which is default behaviour.
4410 Even then only clients that chose the curve will be affected.
4412 This issue was publicly reported as transient failures and was not
4413 initially recognized as a security issue. Thanks to Richard Morgan for
4414 providing reproducible case.
4419 * Removed automatic addition of RPATH in shared libraries and executables,
4420 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4424 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4426 * Fix Use After Free for large message sizes
4428 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4429 message larger than approx 16k is received then the underlying buffer to
4430 store the incoming message is reallocated and moved. Unfortunately a
4431 dangling pointer to the old location is left which results in an attempt to
4432 write to the previously freed location. This is likely to result in a
4433 crash, however it could potentially lead to execution of arbitrary code.
4435 This issue only affects OpenSSL 1.1.0a.
4437 This issue was reported to OpenSSL by Robert Święcki.
4442 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4444 * OCSP Status Request extension unbounded memory growth
4446 A malicious client can send an excessively large OCSP Status Request
4447 extension. If that client continually requests renegotiation, sending a
4448 large OCSP Status Request extension each time, then there will be unbounded
4449 memory growth on the server. This will eventually lead to a Denial Of
4450 Service attack through memory exhaustion. Servers with a default
4451 configuration are vulnerable even if they do not support OCSP. Builds using
4452 the "no-ocsp" build time option are not affected.
4454 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4459 * SSL_peek() hang on empty record
4461 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4462 sends an empty record. This could be exploited by a malicious peer in a
4463 Denial Of Service attack.
4465 This issue was reported to OpenSSL by Alex Gaynor.
4470 * Excessive allocation of memory in tls_get_message_header() and
4471 dtls1_preprocess_fragment()
4473 A (D)TLS message includes 3 bytes for its length in the header for the
4474 message. This would allow for messages up to 16Mb in length. Messages of
4475 this length are excessive and OpenSSL includes a check to ensure that a
4476 peer is sending reasonably sized messages in order to avoid too much memory
4477 being consumed to service a connection. A flaw in the logic of version
4478 1.1.0 means that memory for the message is allocated too early, prior to
4479 the excessive message length check. Due to way memory is allocated in
4480 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4481 to service a connection. This could lead to a Denial of Service through
4482 memory exhaustion. However, the excessive message length check still takes
4483 place, and this would cause the connection to immediately fail. Assuming
4484 that the application calls SSL_free() on the failed connection in a timely
4485 manner then the 21Mb of allocated memory will then be immediately freed
4486 again. Therefore, the excessive memory allocation will be transitory in
4487 nature. This then means that there is only a security impact if:
4489 1) The application does not call SSL_free() in a timely manner in the event
4490 that the connection fails
4492 2) The application is working in a constrained environment where there is
4493 very little free memory
4495 3) The attacker initiates multiple connection attempts such that there are
4496 multiple connections in a state where memory has been allocated for the
4497 connection; SSL_free() has not yet been called; and there is insufficient
4498 memory to service the multiple requests.
4500 Except in the instance of (1) above any Denial Of Service is likely to be
4501 transitory because as soon as the connection fails the memory is
4502 subsequently freed again in the SSL_free() call. However there is an
4503 increased risk during this period of application crashes due to the lack of
4504 memory - which would then mean a more serious Denial of Service.
4506 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4507 (CVE-2016-6307 and CVE-2016-6308)
4511 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4512 had to be removed. Primary reason is that vendor assembler can't
4513 assemble our modules with -KPIC flag. As result it, assembly
4514 support, was not even available as option. But its lack means
4515 lack of side-channel resistant code, which is incompatible with
4516 security by todays standards. Fortunately gcc is readily available
4517 prepackaged option, which we firmly point at...
4521 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4523 * Windows command-line tool supports UTF-8 opt-in option for arguments
4524 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4525 (to any value) allows Windows user to access PKCS#12 file generated
4526 with Windows CryptoAPI and protected with non-ASCII password, as well
4527 as files generated under UTF-8 locale on Linux also protected with
4532 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4533 have been disabled by default and removed from DEFAULT, just like RC4.
4534 See the RC4 item below to re-enable both.
4538 * The method for finding the storage location for the Windows RAND seed file
4539 has changed. First we check %RANDFILE%. If that is not set then we check
4540 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4541 all else fails we fall back to C:\.
4545 * The EVP_EncryptUpdate() function has had its return type changed from void
4546 to int. A return of 0 indicates and error while a return of 1 indicates
4551 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4552 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4553 off the constant time implementation for RSA, DSA and DH have been made
4554 no-ops and deprecated.
4558 * Windows RAND implementation was simplified to only get entropy by
4559 calling CryptGenRandom(). Various other RAND-related tickets
4562 *Joseph Wylie Yandle, Rich Salz*
4564 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4565 and `OPENSSL_LH_`, respectively. The old names are available
4566 with API compatibility. They new names are now completely documented.
4570 * Unify TYPE_up_ref(obj) methods signature.
4571 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4572 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4573 int (instead of void) like all others TYPE_up_ref() methods.
4574 So now these methods also check the return value of CRYPTO_atomic_add(),
4575 and the validity of object reference counter.
4577 *fdasilvayy@gmail.com*
4579 * With Windows Visual Studio builds, the .pdb files are installed
4580 alongside the installed libraries and executables. For a static
4581 library installation, ossl_static.pdb is the associate compiler
4582 generated .pdb file to be used when linking programs.
4586 * Remove openssl.spec. Packaging files belong with the packagers.
4590 * Automatic Darwin/OSX configuration has had a refresh, it will now
4591 recognise x86_64 architectures automatically. You can still decide
4592 to build for a different bitness with the environment variable
4593 KERNEL_BITS (can be 32 or 64), for example:
4595 KERNEL_BITS=32 ./config
4599 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4600 256 bit AES and HMAC with SHA256.
4604 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4608 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4612 * To enable users to have their own config files and build file templates,
4613 Configure looks in the directory indicated by the environment variable
4614 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4615 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4616 name and is used as is.
4620 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4621 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4622 X509_CERT_FILE_CTX was removed.
4626 * "shared" builds are now the default. To create only static libraries use
4627 the "no-shared" Configure option.
4631 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4632 All of these option have not worked for some while and are fundamental
4637 * Make various cleanup routines no-ops and mark them as deprecated. Most
4638 global cleanup functions are no longer required because they are handled
4639 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4640 Explicitly de-initing can cause problems (e.g. where a library that uses
4641 OpenSSL de-inits, but an application is still using it). The affected
4642 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4643 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4644 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4645 COMP_zlib_cleanup().
4649 * --strict-warnings no longer enables runtime debugging options
4650 such as REF_DEBUG. Instead, debug options are automatically
4651 enabled with '--debug' builds.
4653 *Andy Polyakov, Emilia Käsper*
4655 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4656 have been moved out of the public header files. New functions for managing
4657 these have been added.
4661 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4662 objects have been moved out of the public header files. New
4663 functions for managing these have been added.
4667 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4668 have been moved out of the public header files. New functions for managing
4669 these have been added.
4673 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4674 moved out of the public header files. New functions for managing these
4679 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4683 * Removed the mk1mf build scripts.
4687 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4688 it is always safe to #include a header now.
4692 * Removed the aged BC-32 config and all its supporting scripts
4696 * Removed support for Ultrix, Netware, and OS/2.
4700 * Add support for HKDF.
4702 *Alessandro Ghedini*
4704 * Add support for blake2b and blake2s
4708 * Added support for "pipelining". Ciphers that have the
4709 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4710 encryptions/decryptions simultaneously. There are currently no built-in
4711 ciphers with this property but the expectation is that engines will be able
4712 to offer it to significantly improve throughput. Support has been extended
4713 into libssl so that multiple records for a single connection can be
4714 processed in one go (for >=TLS 1.1).
4718 * Added the AFALG engine. This is an async capable engine which is able to
4719 offload work to the Linux kernel. In this initial version it only supports
4720 AES128-CBC. The kernel must be version 4.1.0 or greater.
4724 * OpenSSL now uses a new threading API. It is no longer necessary to
4725 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4726 are two supported threading models: pthreads and windows threads. It is
4727 also possible to configure OpenSSL at compile time for "no-threads". The
4728 old threading API should no longer be used. The functions have been
4729 replaced with "no-op" compatibility macros.
4731 *Alessandro Ghedini, Matt Caswell*
4733 * Modify behavior of ALPN to invoke callback after SNI/servername
4734 callback, such that updates to the SSL_CTX affect ALPN.
4738 * Add SSL_CIPHER queries for authentication and key-exchange.
4742 * Changes to the DEFAULT cipherlist:
4743 - Prefer (EC)DHE handshakes over plain RSA.
4744 - Prefer AEAD ciphers over legacy ciphers.
4745 - Prefer ECDSA over RSA when both certificates are available.
4746 - Prefer TLSv1.2 ciphers/PRF.
4747 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4752 * Change the ECC default curve list to be this, in order: x25519,
4753 secp256r1, secp521r1, secp384r1.
4757 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4758 disabled by default. They can be re-enabled using the
4759 enable-weak-ssl-ciphers option to Configure.
4763 * If the server has ALPN configured, but supports no protocols that the
4764 client advertises, send a fatal "no_application_protocol" alert.
4765 This behaviour is SHALL in RFC 7301, though it isn't universally
4766 implemented by other servers.
4770 * Add X25519 support.
4771 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4772 for public and private key encoding using the format documented in
4773 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4774 key generation and key derivation.
4776 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4781 * Deprecate SRP_VBASE_get_by_user.
4782 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4783 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4784 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4785 seed, even if the seed is configured.
4787 Users should use SRP_VBASE_get1_by_user instead. Note that in
4788 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4789 also that even though configuring the SRP seed attempts to hide
4790 invalid usernames by continuing the handshake with fake
4791 credentials, this behaviour is not constant time and no strong
4792 guarantees are made that the handshake is indistinguishable from
4793 that of a valid user.
4797 * Configuration change; it's now possible to build dynamic engines
4798 without having to build shared libraries and vice versa. This
4799 only applies to the engines in `engines/`, those in `crypto/engine/`
4800 will always be built into libcrypto (i.e. "static").
4802 Building dynamic engines is enabled by default; to disable, use
4803 the configuration option "disable-dynamic-engine".
4805 The only requirements for building dynamic engines are the
4806 presence of the DSO module and building with position independent
4807 code, so they will also automatically be disabled if configuring
4808 with "disable-dso" or "disable-pic".
4810 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4811 are also taken away from openssl/opensslconf.h, as they are
4816 * Configuration change; if there is a known flag to compile
4817 position independent code, it will always be applied on the
4818 libcrypto and libssl object files, and never on the application
4819 object files. This means other libraries that use routines from
4820 libcrypto / libssl can be made into shared libraries regardless
4821 of how OpenSSL was configured.
4823 If this isn't desirable, the configuration options "disable-pic"
4824 or "no-pic" can be used to disable the use of PIC. This will
4825 also disable building shared libraries and dynamic engines.
4829 * Removed JPAKE code. It was experimental and has no wide use.
4833 * The INSTALL_PREFIX Makefile variable has been renamed to
4834 DESTDIR. That makes for less confusion on what this variable
4835 is for. Also, the configuration option --install_prefix is
4840 * Heartbeat for TLS has been removed and is disabled by default
4841 for DTLS; configure with enable-heartbeats. Code that uses the
4842 old #define's might need to be updated.
4844 *Emilia Käsper, Rich Salz*
4846 * Rename REF_CHECK to REF_DEBUG.
4850 * New "unified" build system
4852 The "unified" build system is aimed to be a common system for all
4853 platforms we support. With it comes new support for VMS.
4855 This system builds supports building in a different directory tree
4856 than the source tree. It produces one Makefile (for unix family
4857 or lookalikes), or one descrip.mms (for VMS).
4859 The source of information to make the Makefile / descrip.mms is
4860 small files called 'build.info', holding the necessary
4861 information for each directory with source to compile, and a
4862 template in Configurations, like unix-Makefile.tmpl or
4865 With this change, the library names were also renamed on Windows
4866 and on VMS. They now have names that are closer to the standard
4867 on Unix, and include the major version number, and in certain
4868 cases, the architecture they are built for. See "Notes on shared
4869 libraries" in INSTALL.
4871 We rely heavily on the perl module Text::Template.
4875 * Added support for auto-initialisation and de-initialisation of the library.
4876 OpenSSL no longer requires explicit init or deinit routines to be called,
4877 except in certain circumstances. See the OPENSSL_init_crypto() and
4878 OPENSSL_init_ssl() man pages for further information.
4882 * The arguments to the DTLSv1_listen function have changed. Specifically the
4883 "peer" argument is now expected to be a BIO_ADDR object.
4885 * Rewrite of BIO networking library. The BIO library lacked consistent
4886 support of IPv6, and adding it required some more extensive
4887 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4888 which hold all types of addresses and chains of address information.
4889 It also introduces a new API, with functions like BIO_socket,
4890 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4891 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4892 have been adapted accordingly.
4896 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4901 * CRIME protection: disable compression by default, even if OpenSSL is
4902 compiled with zlib enabled. Applications can still enable compression
4903 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4904 using the SSL_CONF library to configure compression.
4908 * The signature of the session callback configured with
4909 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4910 was explicitly marked as `const unsigned char*` instead of
4915 * Always DPURIFY. Remove the use of uninitialized memory in the
4916 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4920 * Removed many obsolete configuration items, including
4921 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4922 MD2_CHAR, MD2_INT, MD2_LONG
4924 IDEA_SHORT, IDEA_LONG
4925 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4927 *Rich Salz, with advice from Andy Polyakov*
4929 * Many BN internals have been moved to an internal header file.
4931 *Rich Salz with help from Andy Polyakov*
4933 * Configuration and writing out the results from it has changed.
4934 Files such as Makefile include/openssl/opensslconf.h and are now
4935 produced through general templates, such as Makefile.in and
4936 crypto/opensslconf.h.in and some help from the perl module
4939 Also, the center of configuration information is no longer
4940 Makefile. Instead, Configure produces a perl module in
4941 configdata.pm which holds most of the config data (in the hash
4942 table %config), the target data that comes from the target
4943 configuration in one of the `Configurations/*.conf` files (in
4948 * To clarify their intended purposes, the Configure options
4949 --prefix and --openssldir change their semantics, and become more
4950 straightforward and less interdependent.
4952 --prefix shall be used exclusively to give the location INSTALLTOP
4953 where programs, scripts, libraries, include files and manuals are
4954 going to be installed. The default is now /usr/local.
4956 --openssldir shall be used exclusively to give the default
4957 location OPENSSLDIR where certificates, private keys, CRLs are
4958 managed. This is also where the default openssl.cnf gets
4960 If the directory given with this option is a relative path, the
4961 values of both the --prefix value and the --openssldir value will
4962 be combined to become OPENSSLDIR.
4963 The default for --openssldir is INSTALLTOP/ssl.
4965 Anyone who uses --openssldir to specify where OpenSSL is to be
4966 installed MUST change to use --prefix instead.
4970 * The GOST engine was out of date and therefore it has been removed. An up
4971 to date GOST engine is now being maintained in an external repository.
4972 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4973 support for GOST ciphersuites (these are only activated if a GOST engine
4978 * EGD is no longer supported by default; use enable-egd when
4981 *Ben Kaduk and Rich Salz*
4983 * The distribution now has Makefile.in files, which are used to
4984 create Makefile's when Configure is run. *Configure must be run
4985 before trying to build now.*
4989 * The return value for SSL_CIPHER_description() for error conditions
4994 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4996 Obtaining and performing DNSSEC validation of TLSA records is
4997 the application's responsibility. The application provides
4998 the TLSA records of its choice to OpenSSL, and these are then
4999 used to authenticate the peer.
5001 The TLSA records need not even come from DNS. They can, for
5002 example, be used to implement local end-entity certificate or
5003 trust-anchor "pinning", where the "pin" data takes the form
5004 of TLSA records, which can augment or replace verification
5005 based on the usual WebPKI public certification authorities.
5009 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5010 continues to support deprecated interfaces in default builds.
5011 However, applications are strongly advised to compile their
5012 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5013 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5014 or the 1.1.0 releases.
5016 In environments in which all applications have been ported to
5017 not use any deprecated interfaces OpenSSL's Configure script
5018 should be used with the --api=1.1.0 option to entirely remove
5019 support for the deprecated features from the library and
5020 unconditionally disable them in the installed headers.
5021 Essentially the same effect can be achieved with the "no-deprecated"
5022 argument to Configure, except that this will always restrict
5023 the build to just the latest API, rather than a fixed API
5026 As applications are ported to future revisions of the API,
5027 they should update their compile-time OPENSSL_API_COMPAT define
5028 accordingly, but in most cases should be able to continue to
5029 compile with later releases.
5031 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5032 0x10000000L and 0x00908000L, respectively. However those
5033 versions did not support the OPENSSL_API_COMPAT feature, and
5034 so applications are not typically tested for explicit support
5035 of just the undeprecated features of either release.
5039 * Add support for setting the minimum and maximum supported protocol.
5040 It can bet set via the SSL_set_min_proto_version() and
5041 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5042 MaxProtocol. It's recommended to use the new APIs to disable
5043 protocols instead of disabling individual protocols using
5044 SSL_set_options() or SSL_CONF's Protocol. This change also
5045 removes support for disabling TLS 1.2 in the OpenSSL TLS
5046 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5050 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5054 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5055 and integrates ECDSA and ECDH functionality into EC. Implementations can
5056 now redirect key generation and no longer need to convert to or from
5059 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5060 include the ec.h header file instead.
5064 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5065 ciphers who are no longer supported and drops support the ephemeral RSA key
5066 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5070 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5071 opaque. For HMAC_CTX, the following constructors and destructors
5074 HMAC_CTX *HMAC_CTX_new(void);
5075 void HMAC_CTX_free(HMAC_CTX *ctx);
5077 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5078 destroy such methods has been added. See EVP_MD_meth_new(3) and
5079 EVP_CIPHER_meth_new(3) for documentation.
5082 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5083 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5084 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5085 an already created structure.
5086 2) For consistency with the majority of our object creators and
5087 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5088 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5089 for deprecated builds.
5093 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5094 cryptographic operations to be performed asynchronously as long as an
5095 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5096 further details. Libssl has also had this capability integrated with the
5097 introduction of the new mode SSL_MODE_ASYNC and associated error
5098 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5099 pages. This work was developed in partnership with Intel Corp.
5103 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5104 always enabled now. If you want to disable the support you should
5105 exclude it using the list of supported ciphers. This also means that the
5106 "-no_ecdhe" option has been removed from s_server.
5110 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5111 SSL_{CTX_}set1_curves() which can set a list.
5115 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5116 curve you want to support using SSL_{CTX_}set1_curves().
5120 * State machine rewrite. The state machine code has been significantly
5121 refactored in order to remove much duplication of code and solve issues
5122 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5123 further details). This change does have some associated API changes.
5124 Notably the SSL_state() function has been removed and replaced by
5125 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5126 SSL_set_state() has been removed altogether. The previous handshake states
5127 defined in ssl.h and ssl3.h have also been removed.
5131 * All instances of the string "ssleay" in the public API were replaced
5132 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5133 Some error codes related to internal RSA_eay API's were renamed.
5137 * The demo files in crypto/threads were moved to demo/threads.
5141 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5144 *Matt Caswell, Rich Salz*
5146 * New ASN.1 embed macro.
5148 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5149 structure is not allocated: it is part of the parent. That is instead of
5157 This reduces memory fragmentation and make it impossible to accidentally
5158 set a mandatory field to NULL.
5160 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5161 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5162 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5167 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5171 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5172 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5173 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5174 DES and RC4 ciphersuites.
5178 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5179 This changes the decoding behaviour for some invalid messages,
5180 though the change is mostly in the more lenient direction, and
5181 legacy behaviour is preserved as much as possible.
5185 * Fix no-stdio build.
5186 *David Woodhouse <David.Woodhouse@intel.com> and also*
5187 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5189 * New testing framework
5190 The testing framework has been largely rewritten and is now using
5191 perl and the perl modules Test::Harness and an extended variant of
5192 Test::More called OpenSSL::Test to do its work. All test scripts in
5193 test/ have been rewritten into test recipes, and all direct calls to
5194 executables in test/Makefile have become individual recipes using the
5195 simplified testing OpenSSL::Test::Simple.
5197 For documentation on our testing modules, do:
5199 perldoc test/testlib/OpenSSL/Test/Simple.pm
5200 perldoc test/testlib/OpenSSL/Test.pm
5204 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5205 are used; the latter aborts on memory leaks (usually checked on exit).
5206 Some undocumented "set malloc, etc., hooks" functions were removed
5207 and others were changed. All are now documented.
5211 * In DSA_generate_parameters_ex, if the provided seed is too short,
5214 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5216 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5217 from RFC4279, RFC4785, RFC5487, RFC5489.
5219 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5220 original RSA_PSK patch.
5224 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5225 era flag was never set throughout the codebase (only read). Also removed
5226 SSL3_FLAGS_POP_BUFFER which was only used if
5227 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5231 * Changed the default name options in the "ca", "crl", "req" and "x509"
5232 to be "oneline" instead of "compat".
5236 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5237 not aware of clients that still exhibit this bug, and the workaround
5238 hasn't been working properly for a while.
5242 * The return type of BIO_number_read() and BIO_number_written() as well as
5243 the corresponding num_read and num_write members in the BIO structure has
5244 changed from unsigned long to uint64_t. On platforms where an unsigned
5245 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5250 * Given the pervasive nature of TLS extensions it is inadvisable to run
5251 OpenSSL without support for them. It also means that maintaining
5252 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5253 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5257 * Removed support for the two export grade static DH ciphersuites
5258 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5259 were newly added (along with a number of other static DH ciphersuites) to
5260 1.0.2. However the two export ones have *never* worked since they were
5261 introduced. It seems strange in any case to be adding new export
5262 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5266 * Version negotiation has been rewritten. In particular SSLv23_method(),
5267 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5268 and turned into macros which simply call the new preferred function names
5269 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5270 should use the new names instead. Also as part of this change the ssl23.h
5271 header file has been removed.
5275 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5276 code and the associated standard is no longer considered fit-for-purpose.
5280 * RT2547 was closed. When generating a private key, try to make the
5281 output file readable only by the owner. This behavior change might
5282 be noticeable when interacting with other software.
5284 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5289 * Added HTTP GET support to the ocsp command.
5293 * Changed default digest for the dgst and enc commands from MD5 to
5298 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5302 * Added support for TLS extended master secret from
5303 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5304 initial patch which was a great help during development.
5308 * All libssl internal structures have been removed from the public header
5309 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5310 now redundant). Users should not attempt to access internal structures
5311 directly. Instead they should use the provided API functions.
5315 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5316 Access to deprecated functions can be re-enabled by running config with
5317 "enable-deprecated". In addition applications wishing to use deprecated
5318 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5319 will, by default, disable some transitive includes that previously existed
5320 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5324 * Added support for OCB mode. OpenSSL has been granted a patent license
5325 compatible with the OpenSSL license for use of OCB. Details are available
5326 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5327 for OCB can be removed by calling config with no-ocb.
5331 * SSLv2 support has been removed. It still supports receiving an SSLv2
5332 compatible client hello.
5336 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5337 done while fixing the error code for the key-too-small case.
5339 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5341 * CA.sh has been removed; use CA.pl instead.
5345 * Removed old DES API.
5349 * Remove various unsupported platforms:
5355 Sinix/ReliantUNIX RM400
5360 16-bit platforms such as WIN16
5364 * Clean up OPENSSL_NO_xxx #define's
5365 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5366 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5367 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5368 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5369 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5370 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5371 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5372 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5373 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5374 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5378 * Cleaned up dead code
5379 Remove all but one '#ifdef undef' which is to be looked at.
5383 * Clean up calling of xxx_free routines.
5384 Just like free(), fix most of the xxx_free routines to accept
5385 NULL. Remove the non-null checks from callers. Save much code.
5389 * Add secure heap for storage of private keys (when possible).
5390 Add BIO_s_secmem(), CBIGNUM, etc.
5391 Contributed by Akamai Technologies under our Corporate CLA.
5395 * Experimental support for a new, fast, unbiased prime candidate generator,
5396 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5398 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5400 * New output format NSS in the sess_id command line tool. This allows
5401 exporting the session id and the master key in NSS keylog format.
5403 *Martin Kaiser <martin@kaiser.cx>*
5405 * Harmonize version and its documentation. -f flag is used to display
5408 *mancha <mancha1@zoho.com>*
5410 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5411 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5413 *mancha <mancha1@zoho.com>*
5415 * Fix some double frees. These are not thought to be exploitable.
5417 *mancha <mancha1@zoho.com>*
5419 * A missing bounds check in the handling of the TLS heartbeat extension
5420 can be used to reveal up to 64k of memory to a connected client or
5423 Thanks for Neel Mehta of Google Security for discovering this bug and to
5424 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5425 preparing the fix ([CVE-2014-0160])
5427 *Adam Langley, Bodo Moeller*
5429 * Fix for the attack described in the paper "Recovering OpenSSL
5430 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5431 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5432 <http://eprint.iacr.org/2014/140>
5434 Thanks to Yuval Yarom and Naomi Benger for discovering this
5435 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5437 *Yuval Yarom and Naomi Benger*
5439 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5440 this fixes a limitation in previous versions of OpenSSL.
5444 * Experimental encrypt-then-mac support.
5446 Experimental support for encrypt then mac from
5447 draft-gutmann-tls-encrypt-then-mac-02.txt
5449 To enable it set the appropriate extension number (0x42 for the test
5450 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5452 For non-compliant peers (i.e. just about everything) this should have no
5455 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5459 * Add EVP support for key wrapping algorithms, to avoid problems with
5460 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5461 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5462 algorithms and include tests cases.
5466 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5471 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5472 MGF1 digest and OAEP label.
5476 * Make openssl verify return errors.
5478 *Chris Palmer <palmer@google.com> and Ben Laurie*
5480 * New function ASN1_TIME_diff to calculate the difference between two
5481 ASN1_TIME structures or one structure and the current time.
5485 * Update fips_test_suite to support multiple command line options. New
5486 test to induce all self test errors in sequence and check expected
5491 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5492 sign or verify all in one operation.
5496 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5497 test programs and fips_test_suite. Includes functionality to parse
5498 the minimal script output of fipsalgest.pl directly.
5502 * Add authorisation parameter to FIPS_module_mode_set().
5506 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5510 * Use separate DRBG fields for internal and external flags. New function
5511 FIPS_drbg_health_check() to perform on demand health checking. Add
5512 generation tests to fips_test_suite with reduced health check interval to
5513 demonstrate periodic health checking. Add "nodh" option to
5514 fips_test_suite to skip very slow DH test.
5518 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5523 * More extensive health check for DRBG checking many more failure modes.
5524 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5525 combination: call this in fips_test_suite.
5529 * Add support for canonical generation of DSA parameter 'g'. See
5532 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5533 POST to handle HMAC cases.
5537 * Add functions FIPS_module_version() and FIPS_module_version_text()
5538 to return numerical and string versions of the FIPS module number.
5542 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5543 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5544 outside the validated module in the FIPS capable OpenSSL.
5548 * Minor change to DRBG entropy callback semantics. In some cases
5549 there is no multiple of the block length between min_len and
5550 max_len. Allow the callback to return more than max_len bytes
5551 of entropy but discard any extra: it is the callback's responsibility
5552 to ensure that the extra data discarded does not impact the
5553 requested amount of entropy.
5557 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5558 information in FIPS186-3, SP800-57 and SP800-131A.
5562 * CCM support via EVP. Interface is very similar to GCM case except we
5563 must supply all data in one chunk (i.e. no update, final) and the
5564 message length must be supplied if AAD is used. Add algorithm test
5569 * Initial version of POST overhaul. Add POST callback to allow the status
5570 of POST to be monitored and/or failures induced. Modify fips_test_suite
5571 to use callback. Always run all selftests even if one fails.
5575 * XTS support including algorithm test driver in the fips_gcmtest program.
5576 Note: this does increase the maximum key length from 32 to 64 bytes but
5577 there should be no binary compatibility issues as existing applications
5578 will never use XTS mode.
5582 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5583 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5584 performs algorithm blocking for unapproved PRNG types. Also do not
5585 set PRNG type in FIPS_mode_set(): leave this to the application.
5586 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5587 the standard OpenSSL PRNG: set additional data to a date time vector.
5591 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5592 This shouldn't present any incompatibility problems because applications
5593 shouldn't be using these directly and any that are will need to rethink
5594 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5598 * Extensive self tests and health checking required by SP800-90 DRBG.
5599 Remove strength parameter from FIPS_drbg_instantiate and always
5600 instantiate at maximum supported strength.
5604 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5608 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5612 * New function DH_compute_key_padded() to compute a DH key and pad with
5613 leading zeroes if needed: this complies with SP800-56A et al.
5617 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5618 anything, incomplete, subject to change and largely untested at present.
5622 * Modify fipscanisteronly build option to only build the necessary object
5623 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5627 * Add experimental option FIPSSYMS to give all symbols in
5628 fipscanister.o and FIPS or fips prefix. This will avoid
5629 conflicts with future versions of OpenSSL. Add perl script
5630 util/fipsas.pl to preprocess assembly language source files
5631 and rename any affected symbols.
5635 * Add selftest checks and algorithm block of non-fips algorithms in
5636 FIPS mode. Remove DES2 from selftests.
5640 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5641 return internal method without any ENGINE dependencies. Add new
5642 tiny fips sign and verify functions.
5646 * New build option no-ec2m to disable characteristic 2 code.
5650 * New build option "fipscanisteronly". This only builds fipscanister.o
5651 and (currently) associated fips utilities. Uses the file Makefile.fips
5652 instead of Makefile.org as the prototype.
5656 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5657 Update fips_gcmtest to use IV generator.
5661 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5662 setting output buffer to NULL. The `*Final` function must be
5663 called although it will not retrieve any additional data. The tag
5664 can be set or retrieved with a ctrl. The IV length is by default 12
5665 bytes (96 bits) but can be set to an alternative value. If the IV
5666 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5671 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5672 underlying do_cipher function handles all cipher semantics itself
5673 including padding and finalisation. This is useful if (for example)
5674 an ENGINE cipher handles block padding itself. The behaviour of
5675 do_cipher is subtly changed if this flag is set: the return value
5676 is the number of characters written to the output buffer (zero is
5677 no longer an error code) or a negative error code. Also if the
5678 input buffer is NULL and length 0 finalisation should be performed.
5682 * If a candidate issuer certificate is already part of the constructed
5683 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5687 * Improve forward-security support: add functions
5689 void SSL_CTX_set_not_resumable_session_callback(
5690 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5691 void SSL_set_not_resumable_session_callback(
5692 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5694 for use by SSL/TLS servers; the callback function will be called whenever a
5695 new session is created, and gets to decide whether the session may be
5696 cached to make it resumable (return 0) or not (return 1). (As by the
5697 SSL/TLS protocol specifications, the session_id sent by the server will be
5698 empty to indicate that the session is not resumable; also, the server will
5699 not generate RFC 4507 (RFC 5077) session tickets.)
5701 A simple reasonable callback implementation is to return is_forward_secure.
5702 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5703 by the SSL/TLS server library, indicating whether it can provide forward
5706 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5708 * New -verify_name option in command line utilities to set verification
5713 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5714 Add CMAC pkey methods.
5718 * Experimental renegotiation in s_server -www mode. If the client
5719 browses /reneg connection is renegotiated. If /renegcert it is
5720 renegotiated requesting a certificate.
5724 * Add an "external" session cache for debugging purposes to s_server. This
5725 should help trace issues which normally are only apparent in deployed
5726 multi-process servers.
5730 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5731 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5732 BIO_set_cipher() and some obscure PEM functions were changed so they
5733 can now return an error. The RAND changes required a change to the
5734 RAND_METHOD structure.
5738 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5739 a gcc attribute to warn if the result of a function is ignored. This
5740 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5741 whose return value is often ignored.
5745 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5746 These allow SCTs (signed certificate timestamps) to be requested and
5747 validated when establishing a connection.
5749 *Rob Percival <robpercival@google.com>*
5754 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5756 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5757 used even when parsing explicit parameters, when loading a encoded key
5758 or calling `EC_GROUP_new_from_ecpkparameters()`/
5759 `EC_GROUP_new_from_ecparameters()`.
5760 This prevents bypass of security hardening and performance gains,
5761 especially for curves with specialized EC_METHODs.
5762 By default, if a key encoded with explicit parameters is loaded and later
5763 encoded, the output is still encoded with explicit parameters, even if
5764 internally a "named" EC_GROUP is used for computation.
5768 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5769 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5770 NULL. After this change, only the cofactor parameter can be NULL. It also
5771 does some minimal sanity checks on the passed order.
5776 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5777 An attack is simple, if the first CMS_recipientInfo is valid but the
5778 second CMS_recipientInfo is chosen ciphertext. If the second
5779 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5780 encryption key will be replaced by garbage, and the message cannot be
5781 decoded, but if the RSA decryption fails, the correct encryption key is
5782 used and the recipient will not notice the attack.
5783 As a work around for this potential attack the length of the decrypted
5784 key must be equal to the cipher default key length, in case the
5785 certificate is not given and all recipientInfo are tried out.
5786 The old behaviour can be re-enabled in the CMS code by setting the
5787 CMS_DEBUG_DECRYPT flag.
5792 * Document issue with installation paths in diverse Windows builds
5794 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5795 binaries and run-time config file.
5800 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5802 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5803 This changes the size when using the `genpkey` command when no size is given.
5804 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5805 generation commands to use 2048 bits by default.
5809 * Add FIPS support for Android Arm 64-bit
5811 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5812 Module in Version 2.0.10. For some reason, the corresponding target
5813 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5814 built with FIPS support on Android Arm 64-bit. This omission has been
5817 *Matthias St. Pierre*
5819 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5821 * 0-byte record padding oracle
5823 If an application encounters a fatal protocol error and then calls
5824 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5825 then OpenSSL can respond differently to the calling application if a 0 byte
5826 record is received with invalid padding compared to if a 0 byte record is
5827 received with an invalid MAC. If the application then behaves differently
5828 based on that in a way that is detectable to the remote peer, then this
5829 amounts to a padding oracle that could be used to decrypt data.
5831 In order for this to be exploitable "non-stitched" ciphersuites must be in
5832 use. Stitched ciphersuites are optimised implementations of certain
5833 commonly used ciphersuites. Also the application must call SSL_shutdown()
5834 twice even if a protocol error has occurred (applications should not do
5835 this but some do anyway).
5837 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5838 Aviram, with additional investigation by Steven Collison and Andrew
5839 Hourselt. It was reported to OpenSSL on 10th December 2018.
5844 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5848 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5850 * Microarchitecture timing vulnerability in ECC scalar multiplication
5852 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5853 shown to be vulnerable to a microarchitecture timing side channel attack.
5854 An attacker with sufficient access to mount local timing attacks during
5855 ECDSA signature generation could recover the private key.
5857 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5858 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5864 * Timing vulnerability in DSA signature generation
5866 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5867 timing side channel attack. An attacker could use variations in the signing
5868 algorithm to recover the private key.
5870 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5875 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5876 Module, accidentally introduced while backporting security fixes from the
5877 development branch and hindering the use of ECC in FIPS mode.
5881 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5883 * Client DoS due to large DH parameter
5885 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5886 malicious server can send a very large prime value to the client. This will
5887 cause the client to spend an unreasonably long period of time generating a
5888 key for this prime resulting in a hang until the client has finished. This
5889 could be exploited in a Denial Of Service attack.
5891 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5896 * Cache timing vulnerability in RSA Key Generation
5898 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5899 a cache timing side channel attack. An attacker with sufficient access to
5900 mount cache timing attacks during the RSA key generation process could
5901 recover the private key.
5903 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5904 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5909 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5910 parameter is no longer accepted, as it leads to a corrupt table. NULL
5911 pem_str is reserved for alias entries only.
5915 * Revert blinding in ECDSA sign and instead make problematic addition
5916 length-invariant. Switch even to fixed-length Montgomery multiplication.
5920 * Change generating and checking of primes so that the error rate of not
5921 being prime depends on the intended use based on the size of the input.
5922 For larger primes this will result in more rounds of Miller-Rabin.
5923 The maximal error rate for primes with more than 1080 bits is lowered
5926 *Kurt Roeckx, Annie Yousar*
5928 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5932 * Add blinding to ECDSA and DSA signatures to protect against side channel
5933 attacks discovered by Keegan Ryan (NCC Group).
5937 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5938 now allow empty (zero character) pass phrases.
5942 * Certificate time validation (X509_cmp_time) enforces stricter
5943 compliance with RFC 5280. Fractional seconds and timezone offsets
5944 are no longer allowed.
5948 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5950 * Constructed ASN.1 types with a recursive definition could exceed the stack
5952 Constructed ASN.1 types with a recursive definition (such as can be found
5953 in PKCS7) could eventually exceed the stack given malicious input with
5954 excessive recursion. This could result in a Denial Of Service attack. There
5955 are no such structures used within SSL/TLS that come from untrusted sources
5956 so this is considered safe.
5958 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5964 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5966 * Read/write after SSL object in error state
5968 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5969 mechanism. The intent was that if a fatal error occurred during a handshake
5970 then OpenSSL would move into the error state and would immediately fail if
5971 you attempted to continue the handshake. This works as designed for the
5972 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5973 SSL_connect()), however due to a bug it does not work correctly if
5974 SSL_read() or SSL_write() is called directly. In that scenario, if the
5975 handshake fails then a fatal error will be returned in the initial function
5976 call. If SSL_read()/SSL_write() is subsequently called by the application
5977 for the same SSL object then it will succeed and the data is passed without
5978 being decrypted/encrypted directly from the SSL/TLS record layer.
5980 In order to exploit this issue an application bug would have to be present
5981 that resulted in a call to SSL_read()/SSL_write() being issued after having
5982 already received a fatal error.
5984 This issue was reported to OpenSSL by David Benjamin (Google).
5989 * rsaz_1024_mul_avx2 overflow bug on x86_64
5991 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5992 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5993 Analysis suggests that attacks against RSA and DSA as a result of this
5994 defect would be very difficult to perform and are not believed likely.
5995 Attacks against DH1024 are considered just feasible, because most of the
5996 work necessary to deduce information about a private key may be performed
5997 offline. The amount of resources required for such an attack would be
5998 significant. However, for an attack on TLS to be meaningful, the server
5999 would have to share the DH1024 private key among multiple clients, which is
6000 no longer an option since CVE-2016-0701.
6002 This only affects processors that support the AVX2 but not ADX extensions
6003 like Intel Haswell (4th generation).
6005 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6006 was originally found via the OSS-Fuzz project.
6011 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6013 * bn_sqrx8x_internal carry bug on x86_64
6015 There is a carry propagating bug in the x86_64 Montgomery squaring
6016 procedure. No EC algorithms are affected. Analysis suggests that attacks
6017 against RSA and DSA as a result of this defect would be very difficult to
6018 perform and are not believed likely. Attacks against DH are considered just
6019 feasible (although very difficult) because most of the work necessary to
6020 deduce information about a private key may be performed offline. The amount
6021 of resources required for such an attack would be very significant and
6022 likely only accessible to a limited number of attackers. An attacker would
6023 additionally need online access to an unpatched system using the target
6024 private key in a scenario with persistent DH parameters and a private
6025 key that is shared between multiple clients.
6027 This only affects processors that support the BMI1, BMI2 and ADX extensions
6028 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6030 This issue was reported to OpenSSL by the OSS-Fuzz project.
6035 * Malformed X.509 IPAddressFamily could cause OOB read
6037 If an X.509 certificate has a malformed IPAddressFamily extension,
6038 OpenSSL could do a one-byte buffer overread. The most likely result
6039 would be an erroneous display of the certificate in text format.
6041 This issue was reported to OpenSSL by the OSS-Fuzz project.
6045 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6047 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6048 platform rather than 'mingw'.
6052 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6054 * Truncated packet could crash via OOB read
6056 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6057 cipher is being used, then a truncated packet can cause that host to
6058 perform an out-of-bounds read, usually resulting in a crash.
6060 This issue was reported to OpenSSL by Robert Święcki of Google.
6065 * BN_mod_exp may produce incorrect results on x86_64
6067 There is a carry propagating bug in the x86_64 Montgomery squaring
6068 procedure. No EC algorithms are affected. Analysis suggests that attacks
6069 against RSA and DSA as a result of this defect would be very difficult to
6070 perform and are not believed likely. Attacks against DH are considered just
6071 feasible (although very difficult) because most of the work necessary to
6072 deduce information about a private key may be performed offline. The amount
6073 of resources required for such an attack would be very significant and
6074 likely only accessible to a limited number of attackers. An attacker would
6075 additionally need online access to an unpatched system using the target
6076 private key in a scenario with persistent DH parameters and a private
6077 key that is shared between multiple clients. For example this can occur by
6078 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6079 similar to CVE-2015-3193 but must be treated as a separate problem.
6081 This issue was reported to OpenSSL by the OSS-Fuzz project.
6086 * Montgomery multiplication may produce incorrect results
6088 There is a carry propagating bug in the Broadwell-specific Montgomery
6089 multiplication procedure that handles input lengths divisible by, but
6090 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6091 and DH private keys are impossible. This is because the subroutine in
6092 question is not used in operations with the private key itself and an input
6093 of the attacker's direct choice. Otherwise the bug can manifest itself as
6094 transient authentication and key negotiation failures or reproducible
6095 erroneous outcome of public-key operations with specially crafted input.
6096 Among EC algorithms only Brainpool P-512 curves are affected and one
6097 presumably can attack ECDH key negotiation. Impact was not analyzed in
6098 detail, because pre-requisites for attack are considered unlikely. Namely
6099 multiple clients have to choose the curve in question and the server has to
6100 share the private key among them, neither of which is default behaviour.
6101 Even then only clients that chose the curve will be affected.
6103 This issue was publicly reported as transient failures and was not
6104 initially recognized as a security issue. Thanks to Richard Morgan for
6105 providing reproducible case.
6110 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6111 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6112 prevent issues where no progress is being made and the peer continually
6113 sends unrecognised record types, using up resources processing them.
6117 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6119 * Missing CRL sanity check
6121 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6122 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6123 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6125 This issue only affects the OpenSSL 1.0.2i
6130 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6132 * OCSP Status Request extension unbounded memory growth
6134 A malicious client can send an excessively large OCSP Status Request
6135 extension. If that client continually requests renegotiation, sending a
6136 large OCSP Status Request extension each time, then there will be unbounded
6137 memory growth on the server. This will eventually lead to a Denial Of
6138 Service attack through memory exhaustion. Servers with a default
6139 configuration are vulnerable even if they do not support OCSP. Builds using
6140 the "no-ocsp" build time option are not affected.
6142 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6147 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6150 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6156 * OOB write in MDC2_Update()
6158 An overflow can occur in MDC2_Update() either if called directly or
6159 through the EVP_DigestUpdate() function using MDC2. If an attacker
6160 is able to supply very large amounts of input data after a previous
6161 call to EVP_EncryptUpdate() with a partial block then a length check
6162 can overflow resulting in a heap corruption.
6164 The amount of data needed is comparable to SIZE_MAX which is impractical
6167 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6172 * Malformed SHA512 ticket DoS
6174 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6175 DoS attack where a malformed ticket will result in an OOB read which will
6178 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6179 a custom server callback and ticket lookup mechanism.
6181 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6186 * OOB write in BN_bn2dec()
6188 The function BN_bn2dec() does not check the return value of BN_div_word().
6189 This can cause an OOB write if an application uses this function with an
6190 overly large BIGNUM. This could be a problem if an overly large certificate
6191 or CRL is printed out from an untrusted source. TLS is not affected because
6192 record limits will reject an oversized certificate before it is parsed.
6194 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6199 * OOB read in TS_OBJ_print_bio()
6201 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6202 the total length the OID text representation would use and not the amount
6203 of data written. This will result in OOB reads when large OIDs are
6206 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6211 * Pointer arithmetic undefined behaviour
6213 Avoid some undefined pointer arithmetic
6215 A common idiom in the codebase is to check limits in the following manner:
6218 Where "p" points to some malloc'd data of SIZE bytes and
6221 "len" here could be from some externally supplied data (e.g. from a TLS
6224 The rules of C pointer arithmetic are such that "p + len" is only well
6225 defined where len <= SIZE. Therefore the above idiom is actually
6226 undefined behaviour.
6228 For example this could cause problems if some malloc implementation
6229 provides an address for "p" such that "p + len" actually overflows for
6230 values of len that are too big and therefore p + len < limit.
6232 This issue was reported to OpenSSL by Guido Vranken
6237 * Constant time flag not preserved in DSA signing
6239 Operations in the DSA signing algorithm should run in constant time in
6240 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6241 implementation means that a non-constant time codepath is followed for
6242 certain operations. This has been demonstrated through a cache-timing
6243 attack to be sufficient for an attacker to recover the private DSA key.
6245 This issue was reported by César Pereida (Aalto University), Billy Brumley
6246 (Tampere University of Technology), and Yuval Yarom (The University of
6247 Adelaide and NICTA).
6252 * DTLS buffered message DoS
6254 In a DTLS connection where handshake messages are delivered out-of-order
6255 those messages that OpenSSL is not yet ready to process will be buffered
6256 for later use. Under certain circumstances, a flaw in the logic means that
6257 those messages do not get removed from the buffer even though the handshake
6258 has been completed. An attacker could force up to approx. 15 messages to
6259 remain in the buffer when they are no longer required. These messages will
6260 be cleared when the DTLS connection is closed. The default maximum size for
6261 a message is 100k. Therefore, the attacker could force an additional 1500k
6262 to be consumed per connection. By opening many simultaneous connections an
6263 attacker could cause a DoS attack through memory exhaustion.
6265 This issue was reported to OpenSSL by Quan Luo.
6270 * DTLS replay protection DoS
6272 A flaw in the DTLS replay attack protection mechanism means that records
6273 that arrive for future epochs update the replay protection "window" before
6274 the MAC for the record has been validated. This could be exploited by an
6275 attacker by sending a record for the next epoch (which does not have to
6276 decrypt or have a valid MAC), with a very large sequence number. This means
6277 that all subsequent legitimate packets are dropped causing a denial of
6278 service for a specific DTLS connection.
6280 This issue was reported to OpenSSL by the OCAP audit team.
6285 * Certificate message OOB reads
6287 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6288 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6289 theoretical DoS risk but this has not been observed in practice on common
6292 The messages affected are client certificate, client certificate request
6293 and server certificate. As a result the attack can only be performed
6294 against a client or a server which enables client authentication.
6296 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6301 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6303 * Prevent padding oracle in AES-NI CBC MAC check
6305 A MITM attacker can use a padding oracle attack to decrypt traffic
6306 when the connection uses an AES CBC cipher and the server support
6309 This issue was introduced as part of the fix for Lucky 13 padding
6310 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6311 constant time by making sure that always the same bytes are read and
6312 compared against either the MAC or padding bytes. But it no longer
6313 checked that there was enough data to have both the MAC and padding
6316 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6320 * Fix EVP_EncodeUpdate overflow
6322 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6323 Base64 encoding of binary data. If an attacker is able to supply very large
6324 amounts of input data then a length check can overflow resulting in a heap
6327 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6328 the `PEM_write_bio*` family of functions. These are mainly used within the
6329 OpenSSL command line applications, so any application which processes data
6330 from an untrusted source and outputs it as a PEM file should be considered
6331 vulnerable to this issue. User applications that call these APIs directly
6332 with large amounts of untrusted data may also be vulnerable.
6334 This issue was reported by Guido Vranken.
6339 * Fix EVP_EncryptUpdate overflow
6341 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6342 is able to supply very large amounts of input data after a previous call to
6343 EVP_EncryptUpdate() with a partial block then a length check can overflow
6344 resulting in a heap corruption. Following an analysis of all OpenSSL
6345 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6346 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6347 the first called function after an EVP_EncryptInit(), and therefore that
6348 specific call must be safe. The second form is where the length passed to
6349 EVP_EncryptUpdate() can be seen from the code to be some small value and
6350 therefore there is no possibility of an overflow. Since all instances are
6351 one of these two forms, it is believed that there can be no overflows in
6352 internal code due to this problem. It should be noted that
6353 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6354 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6355 of these calls have also been analysed too and it is believed there are no
6356 instances in internal usage where an overflow could occur.
6358 This issue was reported by Guido Vranken.
6363 * Prevent ASN.1 BIO excessive memory allocation
6365 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6366 a short invalid encoding can cause allocation of large amounts of memory
6367 potentially consuming excessive resources or exhausting memory.
6369 Any application parsing untrusted data through d2i BIO functions is
6370 affected. The memory based functions such as d2i_X509() are *not* affected.
6371 Since the memory based functions are used by the TLS library, TLS
6372 applications are not affected.
6374 This issue was reported by Brian Carpenter.
6381 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6382 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6383 in arbitrary stack data being returned in the buffer.
6385 This issue was reported by Guido Vranken.
6390 * Modify behavior of ALPN to invoke callback after SNI/servername
6391 callback, such that updates to the SSL_CTX affect ALPN.
6395 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6400 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6401 methods are enabled and ssl2 is disabled the methods return NULL.
6405 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6407 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6408 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6409 provide any "EXPORT" or "LOW" strength ciphers.
6413 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6414 is by default disabled at build-time. Builds that are not configured with
6415 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6416 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6417 will need to explicitly call either of:
6419 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6421 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6423 as appropriate. Even if either of those is used, or the application
6424 explicitly uses the version-specific SSLv2_method() or its client and
6425 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6426 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6427 ciphers, and SSLv2 56-bit DES are no longer available.
6432 * Fix a double-free in DSA code
6434 A double free bug was discovered when OpenSSL parses malformed DSA private
6435 keys and could lead to a DoS attack or memory corruption for applications
6436 that receive DSA private keys from untrusted sources. This scenario is
6439 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6445 * Disable SRP fake user seed to address a server memory leak.
6447 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6449 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6450 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6451 was changed to ignore the "fake user" SRP seed, even if the seed
6454 Users should use SRP_VBASE_get1_by_user instead. Note that in
6455 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6456 also that even though configuring the SRP seed attempts to hide
6457 invalid usernames by continuing the handshake with fake
6458 credentials, this behaviour is not constant time and no strong
6459 guarantees are made that the handshake is indistinguishable from
6460 that of a valid user.
6465 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6467 In the BN_hex2bn function the number of hex digits is calculated using an
6468 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6469 large values of `i` this can result in `bn_expand` not allocating any
6470 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6471 field as NULL leading to a subsequent NULL ptr deref. For very large values
6472 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6473 In this case memory is allocated to the internal BIGNUM data field, but it
6474 is insufficiently sized leading to heap corruption. A similar issue exists
6475 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6476 is ever called by user applications with very large untrusted hex/dec data.
6477 This is anticipated to be a rare occurrence.
6479 All OpenSSL internal usage of these functions use data that is not expected
6480 to be untrusted, e.g. config file data or application command line
6481 arguments. If user developed applications generate config file data based
6482 on untrusted data then it is possible that this could also lead to security
6483 consequences. This is also anticipated to be rare.
6485 This issue was reported to OpenSSL by Guido Vranken.
6490 * Fix memory issues in `BIO_*printf` functions
6492 The internal `fmtstr` function used in processing a "%s" format string in
6493 the `BIO_*printf` functions could overflow while calculating the length of a
6494 string and cause an OOB read when printing very long strings.
6496 Additionally the internal `doapr_outch` function can attempt to write to an
6497 OOB memory location (at an offset from the NULL pointer) in the event of a
6498 memory allocation failure. In 1.0.2 and below this could be caused where
6499 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6500 could be in processing a very long "%s" format string. Memory leaks can
6503 The first issue may mask the second issue dependent on compiler behaviour.
6504 These problems could enable attacks where large amounts of untrusted data
6505 is passed to the `BIO_*printf` functions. If applications use these functions
6506 in this way then they could be vulnerable. OpenSSL itself uses these
6507 functions when printing out human-readable dumps of ASN.1 data. Therefore
6508 applications that print this data could be vulnerable if the data is from
6509 untrusted sources. OpenSSL command line applications could also be
6510 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6511 as command line arguments.
6513 Libssl is not considered directly vulnerable. Additionally certificates etc
6514 received via remote connections via libssl are also unlikely to be able to
6515 trigger these issues because of message size limits enforced within libssl.
6517 This issue was reported to OpenSSL Guido Vranken.
6522 * Side channel attack on modular exponentiation
6524 A side-channel attack was found which makes use of cache-bank conflicts on
6525 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6526 of RSA keys. The ability to exploit this issue is limited as it relies on
6527 an attacker who has control of code in a thread running on the same
6528 hyper-threaded core as the victim thread which is performing decryptions.
6530 This issue was reported to OpenSSL by Yuval Yarom, The University of
6531 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6532 Nadia Heninger, University of Pennsylvania with more information at
6533 <http://cachebleed.info>.
6538 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6539 if no keysize is specified with default_bits. This fixes an
6540 omission in an earlier change that changed all RSA/DSA key generation
6541 commands to use 2048 bits by default.
6545 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6547 * DH small subgroups
6549 Historically OpenSSL only ever generated DH parameters based on "safe"
6550 primes. More recently (in version 1.0.2) support was provided for
6551 generating X9.42 style parameter files such as those required for RFC 5114
6552 support. The primes used in such files may not be "safe". Where an
6553 application is using DH configured with parameters based on primes that are
6554 not "safe" then an attacker could use this fact to find a peer's private
6555 DH exponent. This attack requires that the attacker complete multiple
6556 handshakes in which the peer uses the same private DH exponent. For example
6557 this could be used to discover a TLS server's private DH exponent if it's
6558 reusing the private DH exponent or it's using a static DH ciphersuite.
6560 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6561 TLS. It is not on by default. If the option is not set then the server
6562 reuses the same private DH exponent for the life of the server process and
6563 would be vulnerable to this attack. It is believed that many popular
6564 applications do set this option and would therefore not be at risk.
6566 The fix for this issue adds an additional check where a "q" parameter is
6567 available (as is the case in X9.42 based parameters). This detects the
6568 only known attack, and is the only possible defense for static DH
6569 ciphersuites. This could have some performance impact.
6571 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6572 default and cannot be disabled. This could have some performance impact.
6574 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6579 * SSLv2 doesn't block disabled ciphers
6581 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6582 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6583 been disabled, provided that the SSLv2 protocol was not also disabled via
6586 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6587 and Sebastian Schinzel.
6592 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6594 * BN_mod_exp may produce incorrect results on x86_64
6596 There is a carry propagating bug in the x86_64 Montgomery squaring
6597 procedure. No EC algorithms are affected. Analysis suggests that attacks
6598 against RSA and DSA as a result of this defect would be very difficult to
6599 perform and are not believed likely. Attacks against DH are considered just
6600 feasible (although very difficult) because most of the work necessary to
6601 deduce information about a private key may be performed offline. The amount
6602 of resources required for such an attack would be very significant and
6603 likely only accessible to a limited number of attackers. An attacker would
6604 additionally need online access to an unpatched system using the target
6605 private key in a scenario with persistent DH parameters and a private
6606 key that is shared between multiple clients. For example this can occur by
6607 default in OpenSSL DHE based SSL/TLS ciphersuites.
6609 This issue was reported to OpenSSL by Hanno Böck.
6614 * Certificate verify crash with missing PSS parameter
6616 The signature verification routines will crash with a NULL pointer
6617 dereference if presented with an ASN.1 signature using the RSA PSS
6618 algorithm and absent mask generation function parameter. Since these
6619 routines are used to verify certificate signature algorithms this can be
6620 used to crash any certificate verification operation and exploited in a
6621 DoS attack. Any application which performs certificate verification is
6622 vulnerable including OpenSSL clients and servers which enable client
6625 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6630 * X509_ATTRIBUTE memory leak
6632 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6633 memory. This structure is used by the PKCS#7 and CMS routines so any
6634 application which reads PKCS#7 or CMS data from untrusted sources is
6635 affected. SSL/TLS is not affected.
6637 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6643 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6644 This changes the decoding behaviour for some invalid messages,
6645 though the change is mostly in the more lenient direction, and
6646 legacy behaviour is preserved as much as possible.
6650 * In DSA_generate_parameters_ex, if the provided seed is too short,
6653 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6655 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6657 * Alternate chains certificate forgery
6659 During certificate verification, OpenSSL will attempt to find an
6660 alternative certificate chain if the first attempt to build such a chain
6661 fails. An error in the implementation of this logic can mean that an
6662 attacker could cause certain checks on untrusted certificates to be
6663 bypassed, such as the CA flag, enabling them to use a valid leaf
6664 certificate to act as a CA and "issue" an invalid certificate.
6666 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6671 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6673 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6674 incompatibility in the handling of HMAC. The previous ABI has now been
6679 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6681 * Malformed ECParameters causes infinite loop
6683 When processing an ECParameters structure OpenSSL enters an infinite loop
6684 if the curve specified is over a specially malformed binary polynomial
6687 This can be used to perform denial of service against any
6688 system which processes public keys, certificate requests or
6689 certificates. This includes TLS clients and TLS servers with
6690 client authentication enabled.
6692 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6697 * Exploitable out-of-bounds read in X509_cmp_time
6699 X509_cmp_time does not properly check the length of the ASN1_TIME
6700 string and can read a few bytes out of bounds. In addition,
6701 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6704 An attacker can use this to craft malformed certificates and CRLs of
6705 various sizes and potentially cause a segmentation fault, resulting in
6706 a DoS on applications that verify certificates or CRLs. TLS clients
6707 that verify CRLs are affected. TLS clients and servers with client
6708 authentication enabled may be affected if they use custom verification
6711 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6712 independently by Hanno Böck.
6717 * PKCS7 crash with missing EnvelopedContent
6719 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6720 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6721 with missing content and trigger a NULL pointer dereference on parsing.
6723 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6724 structures from untrusted sources are affected. OpenSSL clients and
6725 servers are not affected.
6727 This issue was reported to OpenSSL by Michal Zalewski (Google).
6732 * CMS verify infinite loop with unknown hash function
6734 When verifying a signedData message the CMS code can enter an infinite loop
6735 if presented with an unknown hash function OID. This can be used to perform
6736 denial of service against any system which verifies signedData messages using
6738 This issue was reported to OpenSSL by Johannes Bauer.
6743 * Race condition handling NewSessionTicket
6745 If a NewSessionTicket is received by a multi-threaded client when attempting to
6746 reuse a previous ticket then a race condition can occur potentially leading to
6747 a double free of the ticket data.
6752 * Only support 256-bit or stronger elliptic curves with the
6753 'ecdh_auto' setting (server) or by default (client). Of supported
6754 curves, prefer P-256 (both).
6758 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6760 * ClientHello sigalgs DoS fix
6762 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6763 invalid signature algorithms extension a NULL pointer dereference will
6764 occur. This can be exploited in a DoS attack against the server.
6766 This issue was was reported to OpenSSL by David Ramos of Stanford
6770 *Stephen Henson and Matt Caswell*
6772 * Multiblock corrupted pointer fix
6774 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6775 feature only applies on 64 bit x86 architecture platforms that support AES
6776 NI instructions. A defect in the implementation of "multiblock" can cause
6777 OpenSSL's internal write buffer to become incorrectly set to NULL when
6778 using non-blocking IO. Typically, when the user application is using a
6779 socket BIO for writing, this will only result in a failed connection.
6780 However if some other BIO is used then it is likely that a segmentation
6781 fault will be triggered, thus enabling a potential DoS attack.
6783 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6788 * Segmentation fault in DTLSv1_listen fix
6790 The DTLSv1_listen function is intended to be stateless and processes the
6791 initial ClientHello from many peers. It is common for user code to loop
6792 over the call to DTLSv1_listen until a valid ClientHello is received with
6793 an associated cookie. A defect in the implementation of DTLSv1_listen means
6794 that state is preserved in the SSL object from one invocation to the next
6795 that can lead to a segmentation fault. Errors processing the initial
6796 ClientHello can trigger this scenario. An example of such an error could be
6797 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6800 This issue was reported to OpenSSL by Per Allansson.
6805 * Segmentation fault in ASN1_TYPE_cmp fix
6807 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6808 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6809 certificate signature algorithm consistency this can be used to crash any
6810 certificate verification operation and exploited in a DoS attack. Any
6811 application which performs certificate verification is vulnerable including
6812 OpenSSL clients and servers which enable client authentication.
6817 * Segmentation fault for invalid PSS parameters fix
6819 The signature verification routines will crash with a NULL pointer
6820 dereference if presented with an ASN.1 signature using the RSA PSS
6821 algorithm and invalid parameters. Since these routines are used to verify
6822 certificate signature algorithms this can be used to crash any
6823 certificate verification operation and exploited in a DoS attack. Any
6824 application which performs certificate verification is vulnerable including
6825 OpenSSL clients and servers which enable client authentication.
6827 This issue was was reported to OpenSSL by Brian Carpenter.
6832 * ASN.1 structure reuse memory corruption fix
6834 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6835 memory corruption via an invalid write. Such reuse is and has been
6836 strongly discouraged and is believed to be rare.
6838 Applications that parse structures containing CHOICE or ANY DEFINED BY
6839 components may be affected. Certificate parsing (d2i_X509 and related
6840 functions) are however not affected. OpenSSL clients and servers are
6846 * PKCS7 NULL pointer dereferences fix
6848 The PKCS#7 parsing code does not handle missing outer ContentInfo
6849 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6850 missing content and trigger a NULL pointer dereference on parsing.
6852 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6853 otherwise parse PKCS#7 structures from untrusted sources are
6854 affected. OpenSSL clients and servers are not affected.
6856 This issue was reported to OpenSSL by Michal Zalewski (Google).
6861 * DoS via reachable assert in SSLv2 servers fix
6863 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6864 servers that both support SSLv2 and enable export cipher suites by sending
6865 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6867 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6868 (OpenSSL development team).
6873 * Empty CKE with client auth and DHE fix
6875 If client auth is used then a server can seg fault in the event of a DHE
6876 ciphersuite being selected and a zero length ClientKeyExchange message
6877 being sent by the client. This could be exploited in a DoS attack.
6882 * Handshake with unseeded PRNG fix
6884 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6885 with an unseeded PRNG. The conditions are:
6886 - The client is on a platform where the PRNG has not been seeded
6887 automatically, and the user has not seeded manually
6888 - A protocol specific client method version has been used (i.e. not
6889 SSL_client_methodv23)
6890 - A ciphersuite is used that does not require additional random data from
6891 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6893 If the handshake succeeds then the client random that has been used will
6894 have been generated from a PRNG with insufficient entropy and therefore the
6895 output may be predictable.
6897 For example using the following command with an unseeded openssl will
6898 succeed on an unpatched platform:
6900 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6905 * Use After Free following d2i_ECPrivatekey error fix
6907 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6908 could cause a use after free condition. This, in turn, could cause a double
6909 free in several private key parsing functions (such as d2i_PrivateKey
6910 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6911 for applications that receive EC private keys from untrusted
6912 sources. This scenario is considered rare.
6914 This issue was discovered by the BoringSSL project and fixed in their
6920 * X509_to_X509_REQ NULL pointer deref fix
6922 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6923 the certificate key is invalid. This function is rarely used in practice.
6925 This issue was discovered by Brian Carpenter.
6930 * Removed the export ciphers from the DEFAULT ciphers
6934 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6936 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6937 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6938 So far those who have to target multiple platforms would compromise
6939 and argue that binary targeting say ARMv5 would still execute on
6940 ARMv8. "Universal" build resolves this compromise by providing
6941 near-optimal performance even on newer platforms.
6945 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6946 (other platforms pending).
6948 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6950 * Add support for the SignedCertificateTimestampList certificate and
6951 OCSP response extensions from RFC6962.
6955 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6956 for corner cases. (Certain input points at infinity could lead to
6957 bogus results, with non-infinity inputs mapped to infinity too.)
6961 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6962 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6963 common cases are optimized and there still is room for further
6964 improvements. Vector Permutation AES for Altivec is also added.
6968 * Add support for little-endian ppc64 Linux target.
6970 *Marcelo Cerri (IBM)*
6972 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6973 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6974 are optimized and there still is room for further improvements.
6975 Both 32- and 64-bit modes are supported.
6977 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6979 * Improved ARMv7 NEON support.
6983 * Support for SPARC Architecture 2011 crypto extensions, first
6984 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6985 SHA256/512, MD5, GHASH and modular exponentiation.
6987 *Andy Polyakov, David Miller*
6989 * Accelerated modular exponentiation for Intel processors, a.k.a.
6992 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6994 * Support for new and upcoming Intel processors, including AVX2,
6995 BMI and SHA ISA extensions. This includes additional "stitched"
6996 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6999 This work was sponsored by Intel Corp.
7003 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7004 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7005 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7009 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7010 this fixes a limitation in previous versions of OpenSSL.
7014 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7015 MGF1 digest and OAEP label.
7019 * Add EVP support for key wrapping algorithms, to avoid problems with
7020 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7021 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7022 algorithms and include tests cases.
7026 * Add functions to allocate and set the fields of an ECDSA_METHOD
7029 *Douglas E. Engert, Steve Henson*
7031 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7032 difference in days and seconds between two tm or ASN1_TIME structures.
7036 * Add -rev test option to s_server to just reverse order of characters
7037 received by client and send back to server. Also prints an abbreviated
7038 summary of the connection parameters.
7042 * New option -brief for s_client and s_server to print out a brief summary
7043 of connection parameters.
7047 * Add callbacks for arbitrary TLS extensions.
7049 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7051 * New option -crl_download in several openssl utilities to download CRLs
7052 from CRLDP extension in certificates.
7056 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7060 * New function X509_CRL_diff to generate a delta CRL from the difference
7061 of two full CRLs. Add support to "crl" utility.
7065 * New functions to set lookup_crls function and to retrieve
7066 X509_STORE from X509_STORE_CTX.
7070 * Print out deprecated issuer and subject unique ID fields in
7075 * Extend OCSP I/O functions so they can be used for simple general purpose
7076 HTTP as well as OCSP. New wrapper function which can be used to download
7077 CRLs using the OCSP API.
7081 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7085 * `SSL_CONF*` functions. These provide a common framework for application
7086 configuration using configuration files or command lines.
7090 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7091 message callback and prints the results. Needs compile time option
7092 "enable-ssl-trace". New options to s_client and s_server to enable
7097 * New ctrl and macro to retrieve supported points extensions.
7098 Print out extension in s_server and s_client.
7102 * New functions to retrieve certificate signature and signature
7107 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7112 * New Suite B modes for TLS code. These use and enforce the requirements
7113 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7114 only use Suite B curves. The Suite B modes can be set by using the
7115 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7119 * New chain verification flags for Suite B levels of security. Check
7120 algorithms are acceptable when flags are set in X509_verify_cert.
7124 * Make tls1_check_chain return a set of flags indicating checks passed
7125 by a certificate chain. Add additional tests to handle client
7126 certificates: checks for matching certificate type and issuer name
7131 * If an attempt is made to use a signature algorithm not in the peer
7132 preference list abort the handshake. If client has no suitable
7133 signature algorithms in response to a certificate request do not
7134 use the certificate.
7138 * If server EC tmp key is not in client preference list abort handshake.
7142 * Add support for certificate stores in CERT structure. This makes it
7143 possible to have different stores per SSL structure or one store in
7144 the parent SSL_CTX. Include distinct stores for certificate chain
7145 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7146 to build and store a certificate chain in CERT structure: returning
7147 an error if the chain cannot be built: this will allow applications
7148 to test if a chain is correctly configured.
7150 Note: if the CERT based stores are not set then the parent SSL_CTX
7151 store is used to retain compatibility with existing behaviour.
7155 * New function ssl_set_client_disabled to set a ciphersuite disabled
7156 mask based on the current session, check mask when sending client
7157 hello and checking the requested ciphersuite.
7161 * New ctrls to retrieve and set certificate types in a certificate
7162 request message. Print out received values in s_client. If certificate
7163 types is not set with custom values set sensible values based on
7164 supported signature algorithms.
7168 * Support for distinct client and server supported signature algorithms.
7172 * Add certificate callback. If set this is called whenever a certificate
7173 is required by client or server. An application can decide which
7174 certificate chain to present based on arbitrary criteria: for example
7175 supported signature algorithms. Add very simple example to s_server.
7176 This fixes many of the problems and restrictions of the existing client
7177 certificate callback: for example you can now clear an existing
7178 certificate and specify the whole chain.
7182 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7183 the certificate can be used for (if anything). Set valid_flags field
7184 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7185 to have similar checks in it.
7187 Add new "cert_flags" field to CERT structure and include a "strict mode".
7188 This enforces some TLS certificate requirements (such as only permitting
7189 certificate signature algorithms contained in the supported algorithms
7190 extension) which some implementations ignore: this option should be used
7191 with caution as it could cause interoperability issues.
7195 * Update and tidy signature algorithm extension processing. Work out
7196 shared signature algorithms based on preferences and peer algorithms
7197 and print them out in s_client and s_server. Abort handshake if no
7198 shared signature algorithms.
7202 * Add new functions to allow customised supported signature algorithms
7203 for SSL and SSL_CTX structures. Add options to s_client and s_server
7208 * New function SSL_certs_clear() to delete all references to certificates
7209 from an SSL structure. Before this once a certificate had been added
7210 it couldn't be removed.
7214 * Integrate hostname, email address and IP address checking with certificate
7215 verification. New verify options supporting checking in openssl utility.
7219 * Fixes and wildcard matching support to hostname and email checking
7220 functions. Add manual page.
7222 *Florian Weimer (Red Hat Product Security Team)*
7224 * New functions to check a hostname email or IP address against a
7225 certificate. Add options x509 utility to print results of checks against
7230 * Fix OCSP checking.
7232 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7234 * Initial experimental support for explicitly trusted non-root CAs.
7235 OpenSSL still tries to build a complete chain to a root but if an
7236 intermediate CA has a trust setting included that is used. The first
7237 setting is used: whether to trust (e.g., -addtrust option to the x509
7242 * Add -trusted_first option which attempts to find certificates in the
7243 trusted store even if an untrusted chain is also supplied.
7247 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7248 platform support for Linux and Android.
7252 * Support for linux-x32, ILP32 environment in x86_64 framework.
7256 * Experimental multi-implementation support for FIPS capable OpenSSL.
7257 When in FIPS mode the approved implementations are used as normal,
7258 when not in FIPS mode the internal unapproved versions are used instead.
7259 This means that the FIPS capable OpenSSL isn't forced to use the
7260 (often lower performance) FIPS implementations outside FIPS mode.
7264 * Transparently support X9.42 DH parameters when calling
7265 PEM_read_bio_DHparameters. This means existing applications can handle
7266 the new parameter format automatically.
7270 * Initial experimental support for X9.42 DH parameter format: mainly
7271 to support use of 'q' parameter for RFC5114 parameters.
7275 * Add DH parameters from RFC5114 including test data to dhtest.
7279 * Support for automatic EC temporary key parameter selection. If enabled
7280 the most preferred EC parameters are automatically used instead of
7281 hardcoded fixed parameters. Now a server just has to call:
7282 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7283 support ECDH and use the most appropriate parameters.
7287 * Enhance and tidy EC curve and point format TLS extension code. Use
7288 static structures instead of allocation if default values are used.
7289 New ctrls to set curves we wish to support and to retrieve shared curves.
7290 Print out shared curves in s_server. New options to s_server and s_client
7291 to set list of supported curves.
7295 * New ctrls to retrieve supported signature algorithms and
7296 supported curve values as an array of NIDs. Extend openssl utility
7297 to print out received values.
7301 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7302 between NIDs and the more common NIST names such as "P-256". Enhance
7303 ecparam utility and ECC method to recognise the NIST names for curves.
7307 * Enhance SSL/TLS certificate chain handling to support different
7308 chains for each certificate instead of one chain in the parent SSL_CTX.
7312 * Support for fixed DH ciphersuite client authentication: where both
7313 server and client use DH certificates with common parameters.
7317 * Support for fixed DH ciphersuites: those requiring DH server
7322 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7324 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7325 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7326 X509_CINF_get_signature were reverted post internal team review.
7331 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7333 * OCSP Status Request extension unbounded memory growth
7335 A malicious client can send an excessively large OCSP Status Request
7336 extension. If that client continually requests renegotiation, sending a
7337 large OCSP Status Request extension each time, then there will be unbounded
7338 memory growth on the server. This will eventually lead to a Denial Of
7339 Service attack through memory exhaustion. Servers with a default
7340 configuration are vulnerable even if they do not support OCSP. Builds using
7341 the "no-ocsp" build time option are not affected.
7343 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7348 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7351 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7357 * OOB write in MDC2_Update()
7359 An overflow can occur in MDC2_Update() either if called directly or
7360 through the EVP_DigestUpdate() function using MDC2. If an attacker
7361 is able to supply very large amounts of input data after a previous
7362 call to EVP_EncryptUpdate() with a partial block then a length check
7363 can overflow resulting in a heap corruption.
7365 The amount of data needed is comparable to SIZE_MAX which is impractical
7368 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7373 * Malformed SHA512 ticket DoS
7375 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7376 DoS attack where a malformed ticket will result in an OOB read which will
7379 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7380 a custom server callback and ticket lookup mechanism.
7382 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7387 * OOB write in BN_bn2dec()
7389 The function BN_bn2dec() does not check the return value of BN_div_word().
7390 This can cause an OOB write if an application uses this function with an
7391 overly large BIGNUM. This could be a problem if an overly large certificate
7392 or CRL is printed out from an untrusted source. TLS is not affected because
7393 record limits will reject an oversized certificate before it is parsed.
7395 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7400 * OOB read in TS_OBJ_print_bio()
7402 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7403 the total length the OID text representation would use and not the amount
7404 of data written. This will result in OOB reads when large OIDs are
7407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7412 * Pointer arithmetic undefined behaviour
7414 Avoid some undefined pointer arithmetic
7416 A common idiom in the codebase is to check limits in the following manner:
7419 Where "p" points to some malloc'd data of SIZE bytes and
7422 "len" here could be from some externally supplied data (e.g. from a TLS
7425 The rules of C pointer arithmetic are such that "p + len" is only well
7426 defined where len <= SIZE. Therefore, the above idiom is actually
7427 undefined behaviour.
7429 For example this could cause problems if some malloc implementation
7430 provides an address for "p" such that "p + len" actually overflows for
7431 values of len that are too big and therefore p + len < limit.
7433 This issue was reported to OpenSSL by Guido Vranken
7438 * Constant time flag not preserved in DSA signing
7440 Operations in the DSA signing algorithm should run in constant time in
7441 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7442 implementation means that a non-constant time codepath is followed for
7443 certain operations. This has been demonstrated through a cache-timing
7444 attack to be sufficient for an attacker to recover the private DSA key.
7446 This issue was reported by César Pereida (Aalto University), Billy Brumley
7447 (Tampere University of Technology), and Yuval Yarom (The University of
7448 Adelaide and NICTA).
7453 * DTLS buffered message DoS
7455 In a DTLS connection where handshake messages are delivered out-of-order
7456 those messages that OpenSSL is not yet ready to process will be buffered
7457 for later use. Under certain circumstances, a flaw in the logic means that
7458 those messages do not get removed from the buffer even though the handshake
7459 has been completed. An attacker could force up to approx. 15 messages to
7460 remain in the buffer when they are no longer required. These messages will
7461 be cleared when the DTLS connection is closed. The default maximum size for
7462 a message is 100k. Therefore, the attacker could force an additional 1500k
7463 to be consumed per connection. By opening many simultaneous connections an
7464 attacker could cause a DoS attack through memory exhaustion.
7466 This issue was reported to OpenSSL by Quan Luo.
7471 * DTLS replay protection DoS
7473 A flaw in the DTLS replay attack protection mechanism means that records
7474 that arrive for future epochs update the replay protection "window" before
7475 the MAC for the record has been validated. This could be exploited by an
7476 attacker by sending a record for the next epoch (which does not have to
7477 decrypt or have a valid MAC), with a very large sequence number. This means
7478 that all subsequent legitimate packets are dropped causing a denial of
7479 service for a specific DTLS connection.
7481 This issue was reported to OpenSSL by the OCAP audit team.
7486 * Certificate message OOB reads
7488 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7489 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7490 theoretical DoS risk but this has not been observed in practice on common
7493 The messages affected are client certificate, client certificate request
7494 and server certificate. As a result the attack can only be performed
7495 against a client or a server which enables client authentication.
7497 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7502 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7504 * Prevent padding oracle in AES-NI CBC MAC check
7506 A MITM attacker can use a padding oracle attack to decrypt traffic
7507 when the connection uses an AES CBC cipher and the server support
7510 This issue was introduced as part of the fix for Lucky 13 padding
7511 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7512 constant time by making sure that always the same bytes are read and
7513 compared against either the MAC or padding bytes. But it no longer
7514 checked that there was enough data to have both the MAC and padding
7517 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7522 * Fix EVP_EncodeUpdate overflow
7524 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7525 Base64 encoding of binary data. If an attacker is able to supply very large
7526 amounts of input data then a length check can overflow resulting in a heap
7529 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7530 the `PEM_write_bio*` family of functions. These are mainly used within the
7531 OpenSSL command line applications, so any application which processes data
7532 from an untrusted source and outputs it as a PEM file should be considered
7533 vulnerable to this issue. User applications that call these APIs directly
7534 with large amounts of untrusted data may also be vulnerable.
7536 This issue was reported by Guido Vranken.
7541 * Fix EVP_EncryptUpdate overflow
7543 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7544 is able to supply very large amounts of input data after a previous call to
7545 EVP_EncryptUpdate() with a partial block then a length check can overflow
7546 resulting in a heap corruption. Following an analysis of all OpenSSL
7547 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7548 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7549 the first called function after an EVP_EncryptInit(), and therefore that
7550 specific call must be safe. The second form is where the length passed to
7551 EVP_EncryptUpdate() can be seen from the code to be some small value and
7552 therefore there is no possibility of an overflow. Since all instances are
7553 one of these two forms, it is believed that there can be no overflows in
7554 internal code due to this problem. It should be noted that
7555 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7556 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7557 of these calls have also been analysed too and it is believed there are no
7558 instances in internal usage where an overflow could occur.
7560 This issue was reported by Guido Vranken.
7565 * Prevent ASN.1 BIO excessive memory allocation
7567 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7568 a short invalid encoding can casuse allocation of large amounts of memory
7569 potentially consuming excessive resources or exhausting memory.
7571 Any application parsing untrusted data through d2i BIO functions is
7572 affected. The memory based functions such as d2i_X509() are *not* affected.
7573 Since the memory based functions are used by the TLS library, TLS
7574 applications are not affected.
7576 This issue was reported by Brian Carpenter.
7583 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7584 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7585 in arbitrary stack data being returned in the buffer.
7587 This issue was reported by Guido Vranken.
7592 * Modify behavior of ALPN to invoke callback after SNI/servername
7593 callback, such that updates to the SSL_CTX affect ALPN.
7597 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7602 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7603 methods are enabled and ssl2 is disabled the methods return NULL.
7607 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7609 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7610 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7611 provide any "EXPORT" or "LOW" strength ciphers.
7615 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7616 is by default disabled at build-time. Builds that are not configured with
7617 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7618 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7619 will need to explicitly call either of:
7621 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7623 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7625 as appropriate. Even if either of those is used, or the application
7626 explicitly uses the version-specific SSLv2_method() or its client and
7627 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7628 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7629 ciphers, and SSLv2 56-bit DES are no longer available.
7634 * Fix a double-free in DSA code
7636 A double free bug was discovered when OpenSSL parses malformed DSA private
7637 keys and could lead to a DoS attack or memory corruption for applications
7638 that receive DSA private keys from untrusted sources. This scenario is
7641 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7647 * Disable SRP fake user seed to address a server memory leak.
7649 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7651 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7652 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7653 was changed to ignore the "fake user" SRP seed, even if the seed
7656 Users should use SRP_VBASE_get1_by_user instead. Note that in
7657 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7658 also that even though configuring the SRP seed attempts to hide
7659 invalid usernames by continuing the handshake with fake
7660 credentials, this behaviour is not constant time and no strong
7661 guarantees are made that the handshake is indistinguishable from
7662 that of a valid user.
7667 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7669 In the BN_hex2bn function the number of hex digits is calculated using an
7670 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7671 large values of `i` this can result in `bn_expand` not allocating any
7672 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7673 field as NULL leading to a subsequent NULL ptr deref. For very large values
7674 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7675 In this case memory is allocated to the internal BIGNUM data field, but it
7676 is insufficiently sized leading to heap corruption. A similar issue exists
7677 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7678 is ever called by user applications with very large untrusted hex/dec data.
7679 This is anticipated to be a rare occurrence.
7681 All OpenSSL internal usage of these functions use data that is not expected
7682 to be untrusted, e.g. config file data or application command line
7683 arguments. If user developed applications generate config file data based
7684 on untrusted data then it is possible that this could also lead to security
7685 consequences. This is also anticipated to be rare.
7687 This issue was reported to OpenSSL by Guido Vranken.
7692 * Fix memory issues in `BIO_*printf` functions
7694 The internal `fmtstr` function used in processing a "%s" format string in
7695 the `BIO_*printf` functions could overflow while calculating the length of a
7696 string and cause an OOB read when printing very long strings.
7698 Additionally the internal `doapr_outch` function can attempt to write to an
7699 OOB memory location (at an offset from the NULL pointer) in the event of a
7700 memory allocation failure. In 1.0.2 and below this could be caused where
7701 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7702 could be in processing a very long "%s" format string. Memory leaks can
7705 The first issue may mask the second issue dependent on compiler behaviour.
7706 These problems could enable attacks where large amounts of untrusted data
7707 is passed to the `BIO_*printf` functions. If applications use these functions
7708 in this way then they could be vulnerable. OpenSSL itself uses these
7709 functions when printing out human-readable dumps of ASN.1 data. Therefore
7710 applications that print this data could be vulnerable if the data is from
7711 untrusted sources. OpenSSL command line applications could also be
7712 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7713 as command line arguments.
7715 Libssl is not considered directly vulnerable. Additionally certificates etc
7716 received via remote connections via libssl are also unlikely to be able to
7717 trigger these issues because of message size limits enforced within libssl.
7719 This issue was reported to OpenSSL Guido Vranken.
7724 * Side channel attack on modular exponentiation
7726 A side-channel attack was found which makes use of cache-bank conflicts on
7727 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7728 of RSA keys. The ability to exploit this issue is limited as it relies on
7729 an attacker who has control of code in a thread running on the same
7730 hyper-threaded core as the victim thread which is performing decryptions.
7732 This issue was reported to OpenSSL by Yuval Yarom, The University of
7733 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7734 Nadia Heninger, University of Pennsylvania with more information at
7735 <http://cachebleed.info>.
7740 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7741 if no keysize is specified with default_bits. This fixes an
7742 omission in an earlier change that changed all RSA/DSA key generation
7743 commands to use 2048 bits by default.
7747 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7749 * Protection for DH small subgroup attacks
7751 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7752 switched on by default and cannot be disabled. This could have some
7757 * SSLv2 doesn't block disabled ciphers
7759 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7760 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7761 been disabled, provided that the SSLv2 protocol was not also disabled via
7764 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7765 and Sebastian Schinzel.
7770 * Reject DH handshakes with parameters shorter than 1024 bits.
7774 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7776 * Certificate verify crash with missing PSS parameter
7778 The signature verification routines will crash with a NULL pointer
7779 dereference if presented with an ASN.1 signature using the RSA PSS
7780 algorithm and absent mask generation function parameter. Since these
7781 routines are used to verify certificate signature algorithms this can be
7782 used to crash any certificate verification operation and exploited in a
7783 DoS attack. Any application which performs certificate verification is
7784 vulnerable including OpenSSL clients and servers which enable client
7787 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7792 * X509_ATTRIBUTE memory leak
7794 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7795 memory. This structure is used by the PKCS#7 and CMS routines so any
7796 application which reads PKCS#7 or CMS data from untrusted sources is
7797 affected. SSL/TLS is not affected.
7799 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7805 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7806 This changes the decoding behaviour for some invalid messages,
7807 though the change is mostly in the more lenient direction, and
7808 legacy behaviour is preserved as much as possible.
7812 * In DSA_generate_parameters_ex, if the provided seed is too short,
7813 use a random seed, as already documented.
7815 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7817 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7819 * Alternate chains certificate forgery
7821 During certificate verification, OpenSSL will attempt to find an
7822 alternative certificate chain if the first attempt to build such a chain
7823 fails. An error in the implementation of this logic can mean that an
7824 attacker could cause certain checks on untrusted certificates to be
7825 bypassed, such as the CA flag, enabling them to use a valid leaf
7826 certificate to act as a CA and "issue" an invalid certificate.
7828 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7834 * Race condition handling PSK identify hint
7836 If PSK identity hints are received by a multi-threaded client then
7837 the values are wrongly updated in the parent SSL_CTX structure. This can
7838 result in a race condition potentially leading to a double free of the
7844 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7846 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7847 incompatibility in the handling of HMAC. The previous ABI has now been
7850 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7852 * Malformed ECParameters causes infinite loop
7854 When processing an ECParameters structure OpenSSL enters an infinite loop
7855 if the curve specified is over a specially malformed binary polynomial
7858 This can be used to perform denial of service against any
7859 system which processes public keys, certificate requests or
7860 certificates. This includes TLS clients and TLS servers with
7861 client authentication enabled.
7863 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7868 * Exploitable out-of-bounds read in X509_cmp_time
7870 X509_cmp_time does not properly check the length of the ASN1_TIME
7871 string and can read a few bytes out of bounds. In addition,
7872 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7875 An attacker can use this to craft malformed certificates and CRLs of
7876 various sizes and potentially cause a segmentation fault, resulting in
7877 a DoS on applications that verify certificates or CRLs. TLS clients
7878 that verify CRLs are affected. TLS clients and servers with client
7879 authentication enabled may be affected if they use custom verification
7882 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7883 independently by Hanno Böck.
7888 * PKCS7 crash with missing EnvelopedContent
7890 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7891 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7892 with missing content and trigger a NULL pointer dereference on parsing.
7894 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7895 structures from untrusted sources are affected. OpenSSL clients and
7896 servers are not affected.
7898 This issue was reported to OpenSSL by Michal Zalewski (Google).
7903 * CMS verify infinite loop with unknown hash function
7905 When verifying a signedData message the CMS code can enter an infinite loop
7906 if presented with an unknown hash function OID. This can be used to perform
7907 denial of service against any system which verifies signedData messages using
7909 This issue was reported to OpenSSL by Johannes Bauer.
7914 * Race condition handling NewSessionTicket
7916 If a NewSessionTicket is received by a multi-threaded client when attempting to
7917 reuse a previous ticket then a race condition can occur potentially leading to
7918 a double free of the ticket data.
7923 * Reject DH handshakes with parameters shorter than 768 bits.
7925 *Kurt Roeckx and Emilia Kasper*
7927 * dhparam: generate 2048-bit parameters by default.
7929 *Kurt Roeckx and Emilia Kasper*
7931 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7933 * Segmentation fault in ASN1_TYPE_cmp fix
7935 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7936 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7937 certificate signature algorithm consistency this can be used to crash any
7938 certificate verification operation and exploited in a DoS attack. Any
7939 application which performs certificate verification is vulnerable including
7940 OpenSSL clients and servers which enable client authentication.
7945 * ASN.1 structure reuse memory corruption fix
7947 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7948 memory corruption via an invalid write. Such reuse is and has been
7949 strongly discouraged and is believed to be rare.
7951 Applications that parse structures containing CHOICE or ANY DEFINED BY
7952 components may be affected. Certificate parsing (d2i_X509 and related
7953 functions) are however not affected. OpenSSL clients and servers are
7959 * PKCS7 NULL pointer dereferences fix
7961 The PKCS#7 parsing code does not handle missing outer ContentInfo
7962 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7963 missing content and trigger a NULL pointer dereference on parsing.
7965 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7966 otherwise parse PKCS#7 structures from untrusted sources are
7967 affected. OpenSSL clients and servers are not affected.
7969 This issue was reported to OpenSSL by Michal Zalewski (Google).
7974 * DoS via reachable assert in SSLv2 servers fix
7976 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7977 servers that both support SSLv2 and enable export cipher suites by sending
7978 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7980 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7981 (OpenSSL development team).
7986 * Use After Free following d2i_ECPrivatekey error fix
7988 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7989 could cause a use after free condition. This, in turn, could cause a double
7990 free in several private key parsing functions (such as d2i_PrivateKey
7991 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7992 for applications that receive EC private keys from untrusted
7993 sources. This scenario is considered rare.
7995 This issue was discovered by the BoringSSL project and fixed in their
8001 * X509_to_X509_REQ NULL pointer deref fix
8003 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8004 the certificate key is invalid. This function is rarely used in practice.
8006 This issue was discovered by Brian Carpenter.
8011 * Removed the export ciphers from the DEFAULT ciphers
8015 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8017 * Build fixes for the Windows and OpenVMS platforms
8019 *Matt Caswell and Richard Levitte*
8021 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8023 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8024 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8025 dereference. This could lead to a Denial Of Service attack. Thanks to
8026 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8031 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8032 dtls1_buffer_record function under certain conditions. In particular this
8033 could occur if an attacker sent repeated DTLS records with the same
8034 sequence number but for the next epoch. The memory leak could be exploited
8035 by an attacker in a Denial of Service attack through memory exhaustion.
8036 Thanks to Chris Mueller for reporting this issue.
8041 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8042 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8043 method would be set to NULL which could later result in a NULL pointer
8044 dereference. Thanks to Frank Schmirler for reporting this issue.
8049 * Abort handshake if server key exchange message is omitted for ephemeral
8052 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8053 reporting this issue.
8058 * Remove non-export ephemeral RSA code on client and server. This code
8059 violated the TLS standard by allowing the use of temporary RSA keys in
8060 non-export ciphersuites and could be used by a server to effectively
8061 downgrade the RSA key length used to a value smaller than the server
8062 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8063 INRIA or reporting this issue.
8068 * Fixed issue where DH client certificates are accepted without verification.
8069 An OpenSSL server will accept a DH certificate for client authentication
8070 without the certificate verify message. This effectively allows a client to
8071 authenticate without the use of a private key. This only affects servers
8072 which trust a client certificate authority which issues certificates
8073 containing DH keys: these are extremely rare and hardly ever encountered.
8074 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8080 * Ensure that the session ID context of an SSL is updated when its
8081 SSL_CTX is updated via SSL_set_SSL_CTX.
8083 The session ID context is typically set from the parent SSL_CTX,
8084 and can vary with the CTX.
8088 * Fix various certificate fingerprint issues.
8090 By using non-DER or invalid encodings outside the signed portion of a
8091 certificate the fingerprint can be changed without breaking the signature.
8092 Although no details of the signed portion of the certificate can be changed
8093 this can cause problems with some applications: e.g. those using the
8094 certificate fingerprint for blacklists.
8096 1. Reject signatures with non zero unused bits.
8098 If the BIT STRING containing the signature has non zero unused bits reject
8099 the signature. All current signature algorithms require zero unused bits.
8101 2. Check certificate algorithm consistency.
8103 Check the AlgorithmIdentifier inside TBS matches the one in the
8104 certificate signature. NB: this will result in signature failure
8105 errors for some broken certificates.
8107 Thanks to Konrad Kraszewski from Google for reporting this issue.
8109 3. Check DSA/ECDSA signatures use DER.
8111 Re-encode DSA/ECDSA signatures and compare with the original received
8112 signature. Return an error if there is a mismatch.
8114 This will reject various cases including garbage after signature
8115 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8116 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8117 (negative or with leading zeroes).
8119 Further analysis was conducted and fixes were developed by Stephen Henson
8120 of the OpenSSL core team.
8126 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8127 results on some platforms, including x86_64. This bug occurs at random
8128 with a very low probability, and is not known to be exploitable in any
8129 way, though its exact impact is difficult to determine. Thanks to Pieter
8130 Wuille (Blockstream) who reported this issue and also suggested an initial
8131 fix. Further analysis was conducted by the OpenSSL development team and
8132 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8133 the OpenSSL core team.
8138 * Do not resume sessions on the server if the negotiated protocol
8139 version does not match the session's version. Resuming with a different
8140 version, while not strictly forbidden by the RFC, is of questionable
8141 sanity and breaks all known clients.
8143 *David Benjamin, Emilia Käsper*
8145 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8146 early CCS messages during renegotiation. (Note that because
8147 renegotiation is encrypted, this early CCS was not exploitable.)
8151 * Tighten client-side session ticket handling during renegotiation:
8152 ensure that the client only accepts a session ticket if the server sends
8153 the extension anew in the ServerHello. Previously, a TLS client would
8154 reuse the old extension state and thus accept a session ticket if one was
8155 announced in the initial ServerHello.
8157 Similarly, ensure that the client requires a session ticket if one
8158 was advertised in the ServerHello. Previously, a TLS client would
8159 ignore a missing NewSessionTicket message.
8163 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8167 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8168 sends a carefully crafted handshake message, to cause OpenSSL to fail
8169 to free up to 64k of memory causing a memory leak. This could be
8170 exploited in a Denial Of Service attack. This issue affects OpenSSL
8171 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8172 whether SRTP is used or configured. Implementations of OpenSSL that
8173 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8175 The fix was developed by the OpenSSL team.
8180 * Session Ticket Memory Leak.
8182 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8183 integrity of that ticket is first verified. In the event of a session
8184 ticket integrity check failing, OpenSSL will fail to free memory
8185 causing a memory leak. By sending a large number of invalid session
8186 tickets an attacker could exploit this issue in a Denial Of Service
8192 * Build option no-ssl3 is incomplete.
8194 When OpenSSL is configured with "no-ssl3" as a build option, servers
8195 could accept and complete an SSL 3.0 handshake, and clients could be
8196 configured to send them.
8199 *Akamai and the OpenSSL team*
8201 * Add support for TLS_FALLBACK_SCSV.
8202 Client applications doing fallback retries should call
8203 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8206 *Adam Langley, Bodo Moeller*
8208 * Add additional DigestInfo checks.
8210 Re-encode DigestInto in DER and check against the original when
8211 verifying RSA signature: this will reject any improperly encoded
8212 DigestInfo structures.
8214 Note: this is a precautionary measure and no attacks are currently known.
8218 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8220 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8221 SRP code can be overrun an internal buffer. Add sanity check that
8222 g, A, B < N to SRP code.
8224 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8225 Group for discovering this issue.
8230 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8231 TLS 1.0 instead of higher protocol versions when the ClientHello message
8232 is badly fragmented. This allows a man-in-the-middle attacker to force a
8233 downgrade to TLS 1.0 even if both the server and the client support a
8234 higher protocol version, by modifying the client's TLS records.
8236 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8237 researching this issue.
8242 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8243 to a denial of service attack. A malicious server can crash the client
8244 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8245 ciphersuite and sending carefully crafted handshake messages.
8247 Thanks to Felix Gröbert (Google) for discovering and researching this
8253 * By sending carefully crafted DTLS packets an attacker could cause openssl
8254 to leak memory. This can be exploited through a Denial of Service attack.
8255 Thanks to Adam Langley for discovering and researching this issue.
8260 * An attacker can force openssl to consume large amounts of memory whilst
8261 processing DTLS handshake messages. This can be exploited through a
8262 Denial of Service attack.
8263 Thanks to Adam Langley for discovering and researching this issue.
8268 * An attacker can force an error condition which causes openssl to crash
8269 whilst processing DTLS packets due to memory being freed twice. This
8270 can be exploited through a Denial of Service attack.
8271 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8277 * If a multithreaded client connects to a malicious server using a resumed
8278 session and the server sends an ec point format extension it could write
8279 up to 255 bytes to freed memory.
8281 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8287 * A malicious server can crash an OpenSSL client with a null pointer
8288 dereference (read) by specifying an SRP ciphersuite even though it was not
8289 properly negotiated with the client. This can be exploited through a
8290 Denial of Service attack.
8292 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8293 discovering and researching this issue.
8298 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8299 X509_name_oneline, X509_name_print_ex et al. to leak some information
8300 from the stack. Applications may be affected if they echo pretty printing
8301 output to the attacker.
8303 Thanks to Ivan Fratric (Google) for discovering this issue.
8306 *Emilia Käsper, and Steve Henson*
8308 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8309 for corner cases. (Certain input points at infinity could lead to
8310 bogus results, with non-infinity inputs mapped to infinity too.)
8314 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8316 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8317 handshake can force the use of weak keying material in OpenSSL
8318 SSL/TLS clients and servers.
8320 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8321 researching this issue. ([CVE-2014-0224])
8323 *KIKUCHI Masashi, Steve Henson*
8325 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8326 OpenSSL DTLS client the code can be made to recurse eventually crashing
8329 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8332 *Imre Rad, Steve Henson*
8334 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8335 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8336 client or server. This is potentially exploitable to run arbitrary
8337 code on a vulnerable client or server.
8339 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8341 *Jüri Aedla, Steve Henson*
8343 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8344 are subject to a denial of service attack.
8346 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8347 this issue. ([CVE-2014-3470])
8349 *Felix Gröbert, Ivan Fratric, Steve Henson*
8351 * Harmonize version and its documentation. -f flag is used to display
8354 *mancha <mancha1@zoho.com>*
8356 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8357 in i2d_ECPrivateKey.
8359 *mancha <mancha1@zoho.com>*
8361 * Fix some double frees. These are not thought to be exploitable.
8363 *mancha <mancha1@zoho.com>*
8365 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8367 * A missing bounds check in the handling of the TLS heartbeat extension
8368 can be used to reveal up to 64k of memory to a connected client or
8371 Thanks for Neel Mehta of Google Security for discovering this bug and to
8372 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8373 preparing the fix ([CVE-2014-0160])
8375 *Adam Langley, Bodo Moeller*
8377 * Fix for the attack described in the paper "Recovering OpenSSL
8378 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8379 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8380 <http://eprint.iacr.org/2014/140>
8382 Thanks to Yuval Yarom and Naomi Benger for discovering this
8383 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8385 *Yuval Yarom and Naomi Benger*
8387 * TLS pad extension: draft-agl-tls-padding-03
8389 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8390 TLS client Hello record length value would otherwise be > 255 and
8391 less that 512 pad with a dummy extension containing zeroes so it
8392 is at least 512 bytes long.
8394 *Adam Langley, Steve Henson*
8396 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8398 * Fix for TLS record tampering bug. A carefully crafted invalid
8399 handshake could crash OpenSSL with a NULL pointer exception.
8400 Thanks to Anton Johansson for reporting this issues.
8403 * Keep original DTLS digest and encryption contexts in retransmission
8404 structures so we can use the previous session parameters if they need
8405 to be resent. ([CVE-2013-6450])
8409 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8410 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8411 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8412 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8413 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8414 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8416 *Rob Stradling, Adam Langley*
8418 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8420 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8421 supporting platforms or when small records were transferred.
8423 *Andy Polyakov, Steve Henson*
8425 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8427 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8429 This addresses the flaw in CBC record processing discovered by
8430 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8431 at: <http://www.isg.rhul.ac.uk/tls/>
8433 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8434 Security Group at Royal Holloway, University of London
8435 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8436 Emilia Käsper for the initial patch.
8439 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8441 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8442 ciphersuites which can be exploited in a denial of service attack.
8443 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8444 and detecting this bug and to Wolfgang Ettlinger
8445 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8450 * Return an error when checking OCSP signatures when key is NULL.
8451 This fixes a DoS attack. ([CVE-2013-0166])
8455 * Make openssl verify return errors.
8457 *Chris Palmer <palmer@google.com> and Ben Laurie*
8459 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8460 the right response is stapled. Also change SSL_get_certificate()
8461 so it returns the certificate actually sent.
8462 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8464 *Rob Stradling <rob.stradling@comodo.com>*
8466 * Fix possible deadlock when decoding public keys.
8470 * Don't use TLS 1.0 record version number in initial client hello
8475 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8477 * Sanity check record length before skipping explicit IV in TLS
8478 1.2, 1.1 and DTLS to fix DoS attack.
8480 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8481 fuzzing as a service testing platform.
8486 * Initialise tkeylen properly when encrypting CMS messages.
8487 Thanks to Solar Designer of Openwall for reporting this issue.
8491 * In FIPS mode don't try to use composite ciphers as they are not
8496 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8498 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8499 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8500 mean any application compiled against OpenSSL 1.0.0 headers setting
8501 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8502 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8503 0x10000000L Any application which was previously compiled against
8504 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8505 will need to be recompiled as a result. Letting be results in
8506 inability to disable specifically TLS 1.1 and in client context,
8507 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8511 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8512 disable just protocol X, but all protocols above X *if* there are
8513 protocols *below* X still enabled. In more practical terms it means
8514 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8515 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8516 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8521 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8523 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8524 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8525 in CRYPTO_realloc_clean.
8527 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8528 issue and to Adam Langley <agl@chromium.org> for fixing it.
8531 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8533 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8537 * Workarounds for some broken servers that "hang" if a client hello
8538 record length exceeds 255 bytes.
8540 1. Do not use record version number > TLS 1.0 in initial client
8541 hello: some (but not all) hanging servers will now work.
8542 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8543 the number of ciphers sent in the client hello. This should be
8544 set to an even number, such as 50, for example by passing:
8545 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8546 Most broken servers should now work.
8547 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8548 TLS 1.2 client support entirely.
8552 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8556 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8558 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8559 STRING form instead of a DigestInfo.
8563 * The format used for MDC2 RSA signatures is inconsistent between EVP
8564 and the RSA_sign/RSA_verify functions. This was made more apparent when
8565 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8566 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8567 the correct format in RSA_verify so both forms transparently work.
8571 * Some servers which support TLS 1.0 can choke if we initially indicate
8572 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8573 encrypted premaster secret. As a workaround use the maximum permitted
8574 client version in client hello, this should keep such servers happy
8575 and still work with previous versions of OpenSSL.
8579 * Add support for TLS/DTLS heartbeats.
8581 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8583 * Add support for SCTP.
8585 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8587 * Improved PRNG seeding for VOS.
8589 *Paul Green <Paul.Green@stratus.com>*
8591 * Extensive assembler packs updates, most notably:
8593 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8594 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8595 - x86_64: bit-sliced AES implementation;
8596 - ARM: NEON support, contemporary platforms optimizations;
8597 - s390x: z196 support;
8598 - `*`: GHASH and GF(2^m) multiplication implementations;
8602 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8603 (removal of unnecessary code)
8605 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8607 * Add TLS key material exporter from RFC 5705.
8611 * Add DTLS-SRTP negotiation from RFC 5764.
8615 * Add Next Protocol Negotiation,
8616 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8617 disabled with a no-npn flag to config or Configure. Code donated
8620 *Adam Langley <agl@google.com> and Ben Laurie*
8622 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8623 NIST-P256, NIST-P521, with constant-time single point multiplication on
8624 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8625 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8626 Code made available under Apache License version 2.0.
8628 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8629 line to include this in your build of OpenSSL, and run "make depend" (or
8630 "make update"). This enables the following EC_METHODs:
8632 EC_GFp_nistp224_method()
8633 EC_GFp_nistp256_method()
8634 EC_GFp_nistp521_method()
8636 EC_GROUP_new_by_curve_name() will automatically use these (while
8637 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8640 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8642 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8643 all platforms. Move ssize_t definition from e_os.h to the public
8644 header file e_os2.h as it now appears in public header file cms.h
8648 * New -sigopt option to the ca, req and x509 utilities. Additional
8649 signature parameters can be passed using this option and in
8654 * Add RSA PSS signing function. This will generate and set the
8655 appropriate AlgorithmIdentifiers for PSS based on those in the
8656 corresponding EVP_MD_CTX structure. No application support yet.
8660 * Support for companion algorithm specific ASN1 signing routines.
8661 New function ASN1_item_sign_ctx() signs a pre-initialised
8662 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8663 the appropriate parameters.
8667 * Add new algorithm specific ASN1 verification initialisation function
8668 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8669 handling will be the same no matter what EVP_PKEY_METHOD is used.
8670 Add a PSS handler to support verification of PSS signatures: checked
8671 against a number of sample certificates.
8675 * Add signature printing for PSS. Add PSS OIDs.
8677 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8679 * Add algorithm specific signature printing. An individual ASN1 method
8680 can now print out signatures instead of the standard hex dump.
8682 More complex signatures (e.g. PSS) can print out more meaningful
8683 information. Include DSA version that prints out the signature
8688 * Password based recipient info support for CMS library: implementing
8693 * Split password based encryption into PBES2 and PBKDF2 functions. This
8694 neatly separates the code into cipher and PBE sections and is required
8695 for some algorithms that split PBES2 into separate pieces (such as
8696 password based CMS).
8700 * Session-handling fixes:
8701 - Fix handling of connections that are resuming with a session ID,
8702 but also support Session Tickets.
8703 - Fix a bug that suppressed issuing of a new ticket if the client
8704 presented a ticket with an expired session.
8705 - Try to set the ticket lifetime hint to something reasonable.
8706 - Make tickets shorter by excluding irrelevant information.
8707 - On the client side, don't ignore renewed tickets.
8709 *Adam Langley, Bodo Moeller (Google)*
8711 * Fix PSK session representation.
8715 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8717 This work was sponsored by Intel.
8721 * Add GCM support to TLS library. Some custom code is needed to split
8722 the IV between the fixed (from PRF) and explicit (from TLS record)
8723 portions. This adds all GCM ciphersuites supported by RFC5288 and
8724 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8725 add a special AESGCM string for GCM only.
8729 * Expand range of ctrls for AES GCM. Permit setting invocation
8730 field on decrypt and retrieval of invocation field only on encrypt.
8734 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8735 As required by RFC5289 these ciphersuites cannot be used if for
8736 versions of TLS earlier than 1.2.
8740 * For FIPS capable OpenSSL interpret a NULL default public key method
8741 as unset and return the appropriate default but do *not* set the default.
8742 This means we can return the appropriate method in applications that
8743 switch between FIPS and non-FIPS modes.
8747 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8748 ENGINE is used then we cannot handle that in the FIPS module so we
8749 keep original code iff non-FIPS operations are allowed.
8753 * Add -attime option to openssl utilities.
8755 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8757 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8761 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8762 FIPS EC methods unconditionally for now.
8766 * New build option no-ec2m to disable characteristic 2 code.
8770 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8771 all cases can be covered as some introduce binary incompatibilities.
8775 * Redirect RSA operations to FIPS module including keygen,
8776 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8780 * Add similar low-level API blocking to ciphers.
8784 * low-level digest APIs are not approved in FIPS mode: any attempt
8785 to use these will cause a fatal error. Applications that *really* want
8786 to use them can use the `private_*` version instead.
8790 * Redirect cipher operations to FIPS module for FIPS builds.
8794 * Redirect digest operations to FIPS module for FIPS builds.
8798 * Update build system to add "fips" flag which will link in fipscanister.o
8799 for static and shared library builds embedding a signature if needed.
8803 * Output TLS supported curves in preference order instead of numerical
8804 order. This is currently hardcoded for the highest order curves first.
8805 This should be configurable so applications can judge speed vs strength.
8809 * Add TLS v1.2 server support for client authentication.
8813 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8818 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8819 FIPS modules versions.
8823 * Add TLS v1.2 client side support for client authentication. Keep cache
8824 of handshake records longer as we don't know the hash algorithm to use
8825 until after the certificate request message is received.
8829 * Initial TLS v1.2 client support. Add a default signature algorithms
8830 extension including all the algorithms we support. Parse new signature
8831 format in client key exchange. Relax some ECC signing restrictions for
8832 TLS v1.2 as indicated in RFC5246.
8836 * Add server support for TLS v1.2 signature algorithms extension. Switch
8837 to new signature format when needed using client digest preference.
8838 All server ciphersuites should now work correctly in TLS v1.2. No client
8839 support yet and no support for client certificates.
8843 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8844 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8845 ciphersuites. At present only RSA key exchange ciphersuites work with
8846 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8847 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8848 and version checking.
8852 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8853 with this defined it will not be affected by any changes to ssl internal
8854 structures. Add several utility functions to allow openssl application
8855 to work with OPENSSL_NO_SSL_INTERN defined.
8859 * A long standing patch to add support for SRP from EdelWeb (Peter
8860 Sylvester and Christophe Renou) was integrated.
8861 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8862 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8865 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8869 * Permit abbreviated handshakes when renegotiating using the function
8870 SSL_renegotiate_abbreviated().
8872 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8874 * Add call to ENGINE_register_all_complete() to
8875 ENGINE_load_builtin_engines(), so some implementations get used
8876 automatically instead of needing explicit application support.
8880 * Add support for TLS key exporter as described in RFC5705.
8882 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8884 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8885 a few changes are required:
8887 Add SSL_OP_NO_TLSv1_1 flag.
8888 Add TLSv1_1 methods.
8889 Update version checking logic to handle version 1.1.
8890 Add explicit IV handling (ported from DTLS code).
8891 Add command line options to s_client/s_server.
8898 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8900 * X509_ATTRIBUTE memory leak
8902 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8903 memory. This structure is used by the PKCS#7 and CMS routines so any
8904 application which reads PKCS#7 or CMS data from untrusted sources is
8905 affected. SSL/TLS is not affected.
8907 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8913 * Race condition handling PSK identify hint
8915 If PSK identity hints are received by a multi-threaded client then
8916 the values are wrongly updated in the parent SSL_CTX structure. This can
8917 result in a race condition potentially leading to a double free of the
8923 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8925 * Malformed ECParameters causes infinite loop
8927 When processing an ECParameters structure OpenSSL enters an infinite loop
8928 if the curve specified is over a specially malformed binary polynomial
8931 This can be used to perform denial of service against any
8932 system which processes public keys, certificate requests or
8933 certificates. This includes TLS clients and TLS servers with
8934 client authentication enabled.
8936 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8941 * Exploitable out-of-bounds read in X509_cmp_time
8943 X509_cmp_time does not properly check the length of the ASN1_TIME
8944 string and can read a few bytes out of bounds. In addition,
8945 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8948 An attacker can use this to craft malformed certificates and CRLs of
8949 various sizes and potentially cause a segmentation fault, resulting in
8950 a DoS on applications that verify certificates or CRLs. TLS clients
8951 that verify CRLs are affected. TLS clients and servers with client
8952 authentication enabled may be affected if they use custom verification
8955 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8956 independently by Hanno Böck.
8961 * PKCS7 crash with missing EnvelopedContent
8963 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8964 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8965 with missing content and trigger a NULL pointer dereference on parsing.
8967 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8968 structures from untrusted sources are affected. OpenSSL clients and
8969 servers are not affected.
8971 This issue was reported to OpenSSL by Michal Zalewski (Google).
8976 * CMS verify infinite loop with unknown hash function
8978 When verifying a signedData message the CMS code can enter an infinite loop
8979 if presented with an unknown hash function OID. This can be used to perform
8980 denial of service against any system which verifies signedData messages using
8982 This issue was reported to OpenSSL by Johannes Bauer.
8987 * Race condition handling NewSessionTicket
8989 If a NewSessionTicket is received by a multi-threaded client when attempting to
8990 reuse a previous ticket then a race condition can occur potentially leading to
8991 a double free of the ticket data.
8996 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8998 * Segmentation fault in ASN1_TYPE_cmp fix
9000 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9001 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9002 certificate signature algorithm consistency this can be used to crash any
9003 certificate verification operation and exploited in a DoS attack. Any
9004 application which performs certificate verification is vulnerable including
9005 OpenSSL clients and servers which enable client authentication.
9010 * ASN.1 structure reuse memory corruption fix
9012 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9013 memory corruption via an invalid write. Such reuse is and has been
9014 strongly discouraged and is believed to be rare.
9016 Applications that parse structures containing CHOICE or ANY DEFINED BY
9017 components may be affected. Certificate parsing (d2i_X509 and related
9018 functions) are however not affected. OpenSSL clients and servers are
9024 * PKCS7 NULL pointer dereferences fix
9026 The PKCS#7 parsing code does not handle missing outer ContentInfo
9027 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9028 missing content and trigger a NULL pointer dereference on parsing.
9030 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9031 otherwise parse PKCS#7 structures from untrusted sources are
9032 affected. OpenSSL clients and servers are not affected.
9034 This issue was reported to OpenSSL by Michal Zalewski (Google).
9039 * DoS via reachable assert in SSLv2 servers fix
9041 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9042 servers that both support SSLv2 and enable export cipher suites by sending
9043 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9045 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9046 (OpenSSL development team).
9051 * Use After Free following d2i_ECPrivatekey error fix
9053 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9054 could cause a use after free condition. This, in turn, could cause a double
9055 free in several private key parsing functions (such as d2i_PrivateKey
9056 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9057 for applications that receive EC private keys from untrusted
9058 sources. This scenario is considered rare.
9060 This issue was discovered by the BoringSSL project and fixed in their
9066 * X509_to_X509_REQ NULL pointer deref fix
9068 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9069 the certificate key is invalid. This function is rarely used in practice.
9071 This issue was discovered by Brian Carpenter.
9076 * Removed the export ciphers from the DEFAULT ciphers
9080 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9082 * Build fixes for the Windows and OpenVMS platforms
9084 *Matt Caswell and Richard Levitte*
9086 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9088 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9089 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9090 dereference. This could lead to a Denial Of Service attack. Thanks to
9091 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9096 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9097 dtls1_buffer_record function under certain conditions. In particular this
9098 could occur if an attacker sent repeated DTLS records with the same
9099 sequence number but for the next epoch. The memory leak could be exploited
9100 by an attacker in a Denial of Service attack through memory exhaustion.
9101 Thanks to Chris Mueller for reporting this issue.
9106 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9107 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9108 method would be set to NULL which could later result in a NULL pointer
9109 dereference. Thanks to Frank Schmirler for reporting this issue.
9114 * Abort handshake if server key exchange message is omitted for ephemeral
9117 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9118 reporting this issue.
9123 * Remove non-export ephemeral RSA code on client and server. This code
9124 violated the TLS standard by allowing the use of temporary RSA keys in
9125 non-export ciphersuites and could be used by a server to effectively
9126 downgrade the RSA key length used to a value smaller than the server
9127 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9128 INRIA or reporting this issue.
9133 * Fixed issue where DH client certificates are accepted without verification.
9134 An OpenSSL server will accept a DH certificate for client authentication
9135 without the certificate verify message. This effectively allows a client to
9136 authenticate without the use of a private key. This only affects servers
9137 which trust a client certificate authority which issues certificates
9138 containing DH keys: these are extremely rare and hardly ever encountered.
9139 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9145 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9146 results on some platforms, including x86_64. This bug occurs at random
9147 with a very low probability, and is not known to be exploitable in any
9148 way, though its exact impact is difficult to determine. Thanks to Pieter
9149 Wuille (Blockstream) who reported this issue and also suggested an initial
9150 fix. Further analysis was conducted by the OpenSSL development team and
9151 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9152 the OpenSSL core team.
9157 * Fix various certificate fingerprint issues.
9159 By using non-DER or invalid encodings outside the signed portion of a
9160 certificate the fingerprint can be changed without breaking the signature.
9161 Although no details of the signed portion of the certificate can be changed
9162 this can cause problems with some applications: e.g. those using the
9163 certificate fingerprint for blacklists.
9165 1. Reject signatures with non zero unused bits.
9167 If the BIT STRING containing the signature has non zero unused bits reject
9168 the signature. All current signature algorithms require zero unused bits.
9170 2. Check certificate algorithm consistency.
9172 Check the AlgorithmIdentifier inside TBS matches the one in the
9173 certificate signature. NB: this will result in signature failure
9174 errors for some broken certificates.
9176 Thanks to Konrad Kraszewski from Google for reporting this issue.
9178 3. Check DSA/ECDSA signatures use DER.
9180 Re-encode DSA/ECDSA signatures and compare with the original received
9181 signature. Return an error if there is a mismatch.
9183 This will reject various cases including garbage after signature
9184 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9185 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9186 (negative or with leading zeroes).
9188 Further analysis was conducted and fixes were developed by Stephen Henson
9189 of the OpenSSL core team.
9195 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9197 * Session Ticket Memory Leak.
9199 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9200 integrity of that ticket is first verified. In the event of a session
9201 ticket integrity check failing, OpenSSL will fail to free memory
9202 causing a memory leak. By sending a large number of invalid session
9203 tickets an attacker could exploit this issue in a Denial Of Service
9209 * Build option no-ssl3 is incomplete.
9211 When OpenSSL is configured with "no-ssl3" as a build option, servers
9212 could accept and complete an SSL 3.0 handshake, and clients could be
9213 configured to send them.
9216 *Akamai and the OpenSSL team*
9218 * Add support for TLS_FALLBACK_SCSV.
9219 Client applications doing fallback retries should call
9220 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9223 *Adam Langley, Bodo Moeller*
9225 * Add additional DigestInfo checks.
9227 Re-encode DigestInto in DER and check against the original when
9228 verifying RSA signature: this will reject any improperly encoded
9229 DigestInfo structures.
9231 Note: this is a precautionary measure and no attacks are currently known.
9235 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9237 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9238 to a denial of service attack. A malicious server can crash the client
9239 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9240 ciphersuite and sending carefully crafted handshake messages.
9242 Thanks to Felix Gröbert (Google) for discovering and researching this
9248 * By sending carefully crafted DTLS packets an attacker could cause openssl
9249 to leak memory. This can be exploited through a Denial of Service attack.
9250 Thanks to Adam Langley for discovering and researching this issue.
9255 * An attacker can force openssl to consume large amounts of memory whilst
9256 processing DTLS handshake messages. This can be exploited through a
9257 Denial of Service attack.
9258 Thanks to Adam Langley for discovering and researching this issue.
9263 * An attacker can force an error condition which causes openssl to crash
9264 whilst processing DTLS packets due to memory being freed twice. This
9265 can be exploited through a Denial of Service attack.
9266 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9272 * If a multithreaded client connects to a malicious server using a resumed
9273 session and the server sends an ec point format extension it could write
9274 up to 255 bytes to freed memory.
9276 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9282 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9283 X509_name_oneline, X509_name_print_ex et al. to leak some information
9284 from the stack. Applications may be affected if they echo pretty printing
9285 output to the attacker.
9287 Thanks to Ivan Fratric (Google) for discovering this issue.
9290 *Emilia Käsper, and Steve Henson*
9292 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9293 for corner cases. (Certain input points at infinity could lead to
9294 bogus results, with non-infinity inputs mapped to infinity too.)
9298 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9300 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9301 handshake can force the use of weak keying material in OpenSSL
9302 SSL/TLS clients and servers.
9304 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9305 researching this issue. ([CVE-2014-0224])
9307 *KIKUCHI Masashi, Steve Henson*
9309 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9310 OpenSSL DTLS client the code can be made to recurse eventually crashing
9313 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9316 *Imre Rad, Steve Henson*
9318 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9319 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9320 client or server. This is potentially exploitable to run arbitrary
9321 code on a vulnerable client or server.
9323 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9325 *Jüri Aedla, Steve Henson*
9327 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9328 are subject to a denial of service attack.
9330 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9331 this issue. ([CVE-2014-3470])
9333 *Felix Gröbert, Ivan Fratric, Steve Henson*
9335 * Harmonize version and its documentation. -f flag is used to display
9338 *mancha <mancha1@zoho.com>*
9340 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9341 in i2d_ECPrivateKey.
9343 *mancha <mancha1@zoho.com>*
9345 * Fix some double frees. These are not thought to be exploitable.
9347 *mancha <mancha1@zoho.com>*
9349 * Fix for the attack described in the paper "Recovering OpenSSL
9350 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9351 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9352 <http://eprint.iacr.org/2014/140>
9354 Thanks to Yuval Yarom and Naomi Benger for discovering this
9355 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9357 *Yuval Yarom and Naomi Benger*
9359 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9361 * Keep original DTLS digest and encryption contexts in retransmission
9362 structures so we can use the previous session parameters if they need
9363 to be resent. ([CVE-2013-6450])
9367 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9368 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9369 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9370 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9371 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9372 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9374 *Rob Stradling, Adam Langley*
9376 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9378 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9380 This addresses the flaw in CBC record processing discovered by
9381 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9382 at: <http://www.isg.rhul.ac.uk/tls/>
9384 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9385 Security Group at Royal Holloway, University of London
9386 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9387 Emilia Käsper for the initial patch.
9390 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9392 * Return an error when checking OCSP signatures when key is NULL.
9393 This fixes a DoS attack. ([CVE-2013-0166])
9397 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9398 the right response is stapled. Also change SSL_get_certificate()
9399 so it returns the certificate actually sent.
9400 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9401 (This is a backport)
9403 *Rob Stradling <rob.stradling@comodo.com>*
9405 * Fix possible deadlock when decoding public keys.
9409 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9411 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9414 * Sanity check record length before skipping explicit IV in DTLS
9417 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9418 fuzzing as a service testing platform.
9423 * Initialise tkeylen properly when encrypting CMS messages.
9424 Thanks to Solar Designer of Openwall for reporting this issue.
9428 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9430 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9431 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9432 in CRYPTO_realloc_clean.
9434 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9435 issue and to Adam Langley <agl@chromium.org> for fixing it.
9438 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9440 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9442 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9443 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9444 content decryption and always return the same error. Note: this attack
9445 needs on average 2^20 messages so it only affects automated senders. The
9446 old behaviour can be re-enabled in the CMS code by setting the
9447 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9448 an MMA defence is not necessary.
9449 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9450 this issue. ([CVE-2012-0884])
9454 * Fix CVE-2011-4619: make sure we really are receiving a
9455 client hello before rejecting multiple SGC restarts. Thanks to
9456 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9460 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9462 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9463 Thanks to Antonio Martin, Enterprise Secure Access Research and
9464 Development, Cisco Systems, Inc. for discovering this bug and
9465 preparing a fix. ([CVE-2012-0050])
9469 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9471 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9472 of the Vaudenay padding oracle attack on CBC mode encryption
9473 which enables an efficient plaintext recovery attack against
9474 the OpenSSL implementation of DTLS. Their attack exploits timing
9475 differences arising during decryption processing. A research
9476 paper describing this attack can be found at:
9477 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9478 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9479 Security Group at Royal Holloway, University of London
9480 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9481 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9482 for preparing the fix. ([CVE-2011-4108])
9484 *Robin Seggelmann, Michael Tuexen*
9486 * Clear bytes used for block padding of SSL 3.0 records.
9489 *Adam Langley (Google)*
9491 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9492 Kadianakis <desnacked@gmail.com> for discovering this issue and
9493 Adam Langley for preparing the fix. ([CVE-2011-4619])
9495 *Adam Langley (Google)*
9497 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9499 *Andrey Kulikov <amdeich@gmail.com>*
9501 * Prevent malformed RFC3779 data triggering an assertion failure.
9502 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9503 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9505 *Rob Austein <sra@hactrn.net>*
9507 * Improved PRNG seeding for VOS.
9509 *Paul Green <Paul.Green@stratus.com>*
9511 * Fix ssl_ciph.c set-up race.
9513 *Adam Langley (Google)*
9515 * Fix spurious failures in ecdsatest.c.
9517 *Emilia Käsper (Google)*
9519 * Fix the BIO_f_buffer() implementation (which was mixing different
9520 interpretations of the `..._len` fields).
9522 *Adam Langley (Google)*
9524 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9525 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9526 threads won't reuse the same blinding coefficients.
9528 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9529 lock to call BN_BLINDING_invert_ex, and avoids one use of
9530 BN_BLINDING_update for each BN_BLINDING structure (previously,
9531 the last update always remained unused).
9533 *Emilia Käsper (Google)*
9535 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9537 *Bob Buckholz (Google)*
9539 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9541 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9542 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9544 *Kaspar Brand <ossl@velox.ch>*
9546 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9547 for multi-threaded use of ECDH. ([CVE-2011-3210])
9549 *Adam Langley (Google)*
9551 * Fix x509_name_ex_d2i memory leak on bad inputs.
9555 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9556 signature public key algorithm by using OID xref utilities instead.
9557 Before this you could only use some ECC ciphersuites with SHA1 only.
9561 * Add protection against ECDSA timing attacks as mentioned in the paper
9562 by Billy Bob Brumley and Nicola Tuveri, see:
9563 <http://eprint.iacr.org/2011/232.pdf>
9565 *Billy Bob Brumley and Nicola Tuveri*
9567 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9569 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9571 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9573 * Fix bug in string printing code: if *any* escaping is enabled we must
9574 escape the escape character (backslash) or the resulting string is
9579 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9581 * Disable code workaround for ancient and obsolete Netscape browsers
9582 and servers: an attacker can use it in a ciphersuite downgrade attack.
9583 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9587 * Fixed J-PAKE implementation error, originally discovered by
9588 Sebastien Martini, further info and confirmation from Stefan
9589 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9593 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9595 * Fix extension code to avoid race conditions which can result in a buffer
9596 overrun vulnerability: resumed sessions must not be modified as they can
9597 be shared by multiple threads. CVE-2010-3864
9601 * Fix WIN32 build system to correctly link an ENGINE directory into
9606 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9608 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9611 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9613 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9615 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9616 context. The operation can be customised via the ctrl mechanism in
9617 case ENGINEs want to include additional functionality.
9621 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9625 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9626 output hashes compatible with older versions of OpenSSL.
9628 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9630 * Fix compression algorithm handling: if resuming a session use the
9631 compression algorithm of the resumed session instead of determining
9632 it from client hello again. Don't allow server to change algorithm.
9636 * Add load_crls() function to commands tidying load_certs() too. Add option
9637 to verify utility to allow additional CRLs to be included.
9641 * Update OCSP request code to permit adding custom headers to the request:
9642 some responders need this.
9646 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9649 *Julia Lawall <julia@diku.dk>*
9651 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9652 needlessly dereferenced structures, used obsolete functions and
9653 didn't handle all updated verify codes correctly.
9657 * Disable MD2 in the default configuration.
9661 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9662 indicate the initial BIO being pushed or popped. This makes it possible
9663 to determine whether the BIO is the one explicitly called or as a result
9664 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9665 it handles reference counts correctly and doesn't zero out the I/O bio
9666 when it is not being explicitly popped. WARNING: applications which
9667 included workarounds for the old buggy behaviour will need to be modified
9668 or they could free up already freed BIOs.
9672 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9673 renaming to all platforms (within the 0.9.8 branch, this was
9674 done conditionally on Netware platforms to avoid a name clash).
9676 *Guenter <lists@gknw.net>*
9678 * Add ECDHE and PSK support to DTLS.
9680 *Michael Tuexen <tuexen@fh-muenster.de>*
9682 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9687 * Add "missing" function EVP_MD_flags() (without this the only way to
9688 retrieve a digest flags is by accessing the structure directly. Update
9689 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9690 or cipher is registered as in the "from" argument. Print out all
9691 registered digests in the dgst usage message instead of manually
9692 attempting to work them out.
9696 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9697 this allows the use of compression and extensions. Change default cipher
9698 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9699 by default unless an application cipher string requests it.
9703 * Alter match criteria in PKCS12_parse(). It used to try to use local
9704 key ids to find matching certificates and keys but some PKCS#12 files
9705 don't follow the (somewhat unwritten) rules and this strategy fails.
9706 Now just gather all certificates together and the first private key
9707 then look for the first certificate that matches the key.
9711 * Support use of registered digest and cipher names for dgst and cipher
9712 commands instead of having to add each one as a special case. So now
9719 openssl dgst -sha256 foo
9721 and this works for ENGINE based algorithms too.
9725 * Update Gost ENGINE to support parameter files.
9727 *Victor B. Wagner <vitus@cryptocom.ru>*
9729 * Support GeneralizedTime in ca utility.
9731 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9733 * Enhance the hash format used for certificate directory links. The new
9734 form uses the canonical encoding (meaning equivalent names will work
9735 even if they aren't identical) and uses SHA1 instead of MD5. This form
9736 is incompatible with the older format and as a result c_rehash should
9737 be used to rebuild symbolic links.
9741 * Make PKCS#8 the default write format for private keys, replacing the
9742 traditional format. This form is standardised, more secure and doesn't
9743 include an implicit MD5 dependency.
9747 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9748 committed to OpenSSL should pass this lot as a minimum.
9752 * Add session ticket override functionality for use by EAP-FAST.
9754 *Jouni Malinen <j@w1.fi>*
9756 * Modify HMAC functions to return a value. Since these can be implemented
9757 in an ENGINE errors can occur.
9761 * Type-checked OBJ_bsearch_ex.
9765 * Type-checked OBJ_bsearch. Also some constification necessitated
9766 by type-checking. Still to come: TXT_DB, bsearch(?),
9767 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9772 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9773 seconds to a tm structure directly, instead of going through OS
9774 specific date routines. This avoids any issues with OS routines such
9775 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9776 and X509_time_adj_ex() to cover the extended range. The existing
9777 X509_time_adj() is still usable and will no longer have any date issues.
9781 * Delta CRL support. New use deltas option which will attempt to locate
9782 and search any appropriate delta CRLs available.
9784 This work was sponsored by Google.
9788 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9789 code and add additional score elements. Validate alternate CRL paths
9790 as part of the CRL checking and indicate a new error "CRL path validation
9791 error" in this case. Applications wanting additional details can use
9792 the verify callback and check the new "parent" field. If this is not
9793 NULL CRL path validation is taking place. Existing applications won't
9794 see this because it requires extended CRL support which is off by
9797 This work was sponsored by Google.
9801 * Support for freshest CRL extension.
9803 This work was sponsored by Google.
9807 * Initial indirect CRL support. Currently only supported in the CRLs
9808 passed directly and not via lookup. Process certificate issuer
9809 CRL entry extension and lookup CRL entries by bother issuer name
9810 and serial number. Check and process CRL issuer entry in IDP extension.
9812 This work was sponsored by Google.
9816 * Add support for distinct certificate and CRL paths. The CRL issuer
9817 certificate is validated separately in this case. Only enabled if
9818 an extended CRL support flag is set: this flag will enable additional
9819 CRL functionality in future.
9821 This work was sponsored by Google.
9825 * Add support for policy mappings extension.
9827 This work was sponsored by Google.
9831 * Fixes to pathlength constraint, self issued certificate handling,
9832 policy processing to align with RFC3280 and PKITS tests.
9834 This work was sponsored by Google.
9838 * Support for name constraints certificate extension. DN, email, DNS
9839 and URI types are currently supported.
9841 This work was sponsored by Google.
9845 * To cater for systems that provide a pointer-based thread ID rather
9846 than numeric, deprecate the current numeric thread ID mechanism and
9847 replace it with a structure and associated callback type. This
9848 mechanism allows a numeric "hash" to be extracted from a thread ID in
9849 either case, and on platforms where pointers are larger than 'long',
9850 mixing is done to help ensure the numeric 'hash' is usable even if it
9851 can't be guaranteed unique. The default mechanism is to use "&errno"
9852 as a pointer-based thread ID to distinguish between threads.
9854 Applications that want to provide their own thread IDs should now use
9855 CRYPTO_THREADID_set_callback() to register a callback that will call
9856 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9858 Note that ERR_remove_state() is now deprecated, because it is tied
9859 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9860 to free the current thread's error state should be replaced by
9861 ERR_remove_thread_state(NULL).
9863 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9864 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9865 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9866 application was previously providing a numeric thread callback that
9867 was inappropriate for distinguishing threads, then uniqueness might
9868 have been obtained with &errno that happened immediately in the
9869 intermediate development versions of OpenSSL; this is no longer the
9870 case, the numeric thread callback will now override the automatic use
9873 *Geoff Thorpe, with help from Bodo Moeller*
9875 * Initial support for different CRL issuing certificates. This covers a
9876 simple case where the self issued certificates in the chain exist and
9877 the real CRL issuer is higher in the existing chain.
9879 This work was sponsored by Google.
9883 * Removed effectively defunct crypto/store from the build.
9887 * Revamp of STACK to provide stronger type-checking. Still to come:
9888 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9889 ASN1_STRING, CONF_VALUE.
9893 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9894 RAM on SSL connections. This option can save about 34k per idle SSL.
9898 * Revamp of LHASH to provide stronger type-checking. Still to come:
9899 STACK, TXT_DB, bsearch, qsort.
9903 * Initial support for Cryptographic Message Syntax (aka CMS) based
9904 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9905 support for data, signedData, compressedData, digestedData and
9906 encryptedData, envelopedData types included. Scripts to check against
9907 RFC4134 examples draft and interop and consistency checks of many
9908 content types and variants.
9912 * Add options to enc utility to support use of zlib compression BIO.
9916 * Extend mk1mf to support importing of options and assembly language
9917 files from Configure script, currently only included in VC-WIN32.
9918 The assembly language rules can now optionally generate the source
9919 files from the associated perl scripts.
9923 * Implement remaining functionality needed to support GOST ciphersuites.
9924 Interop testing has been performed using CryptoPro implementations.
9926 *Victor B. Wagner <vitus@cryptocom.ru>*
9928 * s390x assembler pack.
9932 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9937 * Implement Opaque PRF Input TLS extension as specified in
9938 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9939 official specification yet and no extension type assignment by
9940 IANA exists, this extension (for now) will have to be explicitly
9941 enabled when building OpenSSL by providing the extension number
9942 to use. For example, specify an option
9944 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9946 to the "config" or "Configure" script to enable the extension,
9947 assuming extension number 0x9527 (which is a completely arbitrary
9948 and unofficial assignment based on the MD5 hash of the Internet
9949 Draft). Note that by doing so, you potentially lose
9950 interoperability with other TLS implementations since these might
9951 be using the same extension number for other purposes.
9953 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9954 opaque PRF input value to use in the handshake. This will create
9955 an internal copy of the length-'len' string at 'src', and will
9956 return non-zero for success.
9958 To get more control and flexibility, provide a callback function
9961 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9962 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9966 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9969 Callback function 'cb' will be called in handshakes, and is
9970 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9971 Argument 'arg' is for application purposes (the value as given to
9972 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9973 be provided to the callback function). The callback function
9974 has to return non-zero to report success: usually 1 to use opaque
9975 PRF input just if possible, or 2 to enforce use of the opaque PRF
9976 input. In the latter case, the library will abort the handshake
9977 if opaque PRF input is not successfully negotiated.
9979 Arguments 'peerinput' and 'len' given to the callback function
9980 will always be NULL and 0 in the case of a client. A server will
9981 see the client's opaque PRF input through these variables if
9982 available (NULL and 0 otherwise). Note that if the server
9983 provides an opaque PRF input, the length must be the same as the
9984 length of the client's opaque PRF input.
9986 Note that the callback function will only be called when creating
9987 a new session (session resumption can resume whatever was
9988 previously negotiated), and will not be called in SSL 2.0
9989 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9990 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9991 for applications that need to enforce opaque PRF input.
9995 * Update ssl code to support digests other than SHA1+MD5 for handshake
9998 *Victor B. Wagner <vitus@cryptocom.ru>*
10000 * Add RFC4507 support to OpenSSL. This includes the corrections in
10001 RFC4507bis. The encrypted ticket format is an encrypted encoded
10002 SSL_SESSION structure, that way new session features are automatically
10005 If a client application caches session in an SSL_SESSION structure
10006 support is transparent because tickets are now stored in the encoded
10009 The SSL_CTX structure automatically generates keys for ticket
10010 protection in servers so again support should be possible
10011 with no application modification.
10013 If a client or server wishes to disable RFC4507 support then the option
10014 SSL_OP_NO_TICKET can be set.
10016 Add a TLS extension debugging callback to allow the contents of any client
10017 or server extensions to be examined.
10019 This work was sponsored by Google.
10023 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10024 OpenSSL should now compile cleanly on gcc 4.2
10026 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10028 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10029 support including streaming MAC support: this is required for GOST
10030 ciphersuite support.
10032 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10034 * Add option -stream to use PKCS#7 streaming in smime utility. New
10035 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10036 to output in BER and PEM format.
10040 * Experimental support for use of HMAC via EVP_PKEY interface. This
10041 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10042 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10043 ENGINE support for HMAC keys which are unextractable. New -mac and
10044 -macopt options to dgst utility.
10048 * New option -sigopt to dgst utility. Update dgst to use
10049 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10050 alternative signing parameters such as X9.31 or PSS in the dgst
10055 * Change ssl_cipher_apply_rule(), the internal function that does
10056 the work each time a ciphersuite string requests enabling
10057 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10058 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10059 the order of disabled ciphersuites such that those ciphersuites
10060 that most recently went from enabled to disabled not only stay
10061 in order with respect to each other, but also have higher priority
10062 than other disabled ciphersuites the next time ciphersuites are
10065 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10066 the same ciphersuites as with "HIGH" alone, but in a specific
10067 order where the PSK ciphersuites come first (since they are the
10068 most recently disabled ciphersuites when "HIGH" is parsed).
10070 Also, change ssl_create_cipher_list() (using this new
10071 functionality) such that between otherwise identical
10072 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10077 * Change ssl_create_cipher_list() so that it automatically
10078 arranges the ciphersuites in reasonable order before starting
10079 to process the rule string. Thus, the definition for "DEFAULT"
10080 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10081 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10082 This makes it much easier to arrive at a reasonable default order
10083 in applications for which anonymous ciphers are OK (meaning
10084 that you can't actually use DEFAULT).
10086 *Bodo Moeller; suggested by Victor Duchovni*
10088 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10089 processing) into multiple integers instead of setting
10090 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10091 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10092 (These masks as well as the individual bit definitions are hidden
10093 away into the non-exported interface ssl/ssl_locl.h, so this
10094 change to the definition of the SSL_CIPHER structure shouldn't
10095 affect applications.) This give us more bits for each of these
10096 categories, so there is no longer a need to coagulate AES128 and
10097 AES256 into a single algorithm bit, and to coagulate Camellia128
10098 and Camellia256 into a single algorithm bit, which has led to all
10101 Thus, among other things, the kludge introduced in 0.9.7m and
10102 0.9.8e for masking out AES256 independently of AES128 or masking
10103 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10105 With the change, we also introduce new ciphersuite aliases that
10106 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10111 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10112 Use the leftmost N bytes of the signature input if the input is
10113 larger than the prime q (with N being the size in bytes of q).
10117 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10118 it yet and it is largely untested.
10122 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10126 * Initial incomplete changes to avoid need for function casts in OpenSSL
10127 some compilers (gcc 4.2 and later) reject their use. Safestack is
10128 reimplemented. Update ASN1 to avoid use of legacy functions.
10132 * Win32/64 targets are linked with Winsock2.
10136 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10137 to external functions. This can be used to increase CRL handling
10138 efficiency especially when CRLs are very large by (for example) storing
10139 the CRL revoked certificates in a database.
10143 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10144 new CRLs added to a directory can be used. New command line option
10145 -verify_return_error to s_client and s_server. This causes real errors
10146 to be returned by the verify callback instead of carrying on no matter
10147 what. This reflects the way a "real world" verify callback would behave.
10151 * GOST engine, supporting several GOST algorithms and public key formats.
10152 Kindly donated by Cryptocom.
10156 * Partial support for Issuing Distribution Point CRL extension. CRLs
10157 partitioned by DP are handled but no indirect CRL or reason partitioning
10158 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10159 selected via a scoring technique which handles IDP and AKID in CRLs.
10163 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10164 will ultimately be used for all verify operations: this will remove the
10165 X509_STORE dependency on certificate verification and allow alternative
10166 lookup methods. X509_STORE based implementations of these two callbacks.
10170 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10171 Modify get_crl() to find a valid (unexpired) CRL if possible.
10175 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10176 this would be called X509_CRL_cmp() but that name is already used by
10177 a function that just compares CRL issuer names. Cache several CRL
10178 extensions in X509_CRL structure and cache CRLDP in X509.
10182 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10183 this maps equivalent X509_NAME structures into a consistent structure.
10184 Name comparison can then be performed rapidly using memcmp().
10188 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10193 * Allow digests to supply their own micalg string for S/MIME type using
10194 the ctrl EVP_MD_CTRL_MICALG.
10198 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10199 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10200 ctrl. It can then customise the structure before and/or after signing
10205 * New function OBJ_add_sigid() to allow application defined signature OIDs
10206 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10207 to free up any added signature OIDs.
10211 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10212 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10213 digest and cipher tables. New options added to openssl utility:
10214 list-message-digest-algorithms and list-cipher-algorithms.
10218 * Change the array representation of binary polynomials: the list
10219 of degrees of non-zero coefficients is now terminated with -1.
10220 Previously it was terminated with 0, which was also part of the
10221 value; thus, the array representation was not applicable to
10222 polynomials where t^0 has coefficient zero. This change makes
10223 the array representation useful in a more general context.
10227 * Various modifications and fixes to SSL/TLS cipher string
10228 handling. For ECC, the code now distinguishes between fixed ECDH
10229 with RSA certificates on the one hand and with ECDSA certificates
10230 on the other hand, since these are separate ciphersuites. The
10231 unused code for Fortezza ciphersuites has been removed.
10233 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10234 (not "ECDHE"). For consistency with the code for DH
10235 certificates, use of ECDH certificates is now considered ECDH
10236 authentication, not RSA or ECDSA authentication (the latter is
10237 merely the CA's signing algorithm and not actively used in the
10240 The temporary ciphersuite alias "ECCdraft" is no longer
10241 available, and ECC ciphersuites are no longer excluded from "ALL"
10242 and "DEFAULT". The following aliases now exist for RFC 4492
10243 ciphersuites, most of these by analogy with the DH case:
10245 kECDHr - ECDH cert, signed with RSA
10246 kECDHe - ECDH cert, signed with ECDSA
10247 kECDH - ECDH cert (signed with either RSA or ECDSA)
10248 kEECDH - ephemeral ECDH
10249 ECDH - ECDH cert or ephemeral ECDH
10252 aECDSA - ECDSA cert
10255 AECDH - anonymous ECDH
10256 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10260 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10261 Use correct micalg parameters depending on digest(s) in signed message.
10265 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10266 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10270 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10271 an engine to register a method. Add ENGINE lookups for methods and
10272 functional reference processing.
10276 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10277 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10282 * New -resign option to smime utility. This adds one or more signers
10283 to an existing PKCS#7 signedData structure. Also -md option to use an
10284 alternative message digest algorithm for signing.
10288 * Tidy up PKCS#7 routines and add new functions to make it easier to
10289 create PKCS7 structures containing multiple signers. Update smime
10290 application to support multiple signers.
10294 * New -macalg option to pkcs12 utility to allow setting of an alternative
10299 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10300 Reorganize PBE internals to lookup from a static table using NIDs,
10301 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10302 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10303 PRF which will be automatically used with PBES2.
10307 * Replace the algorithm specific calls to generate keys in "req" with the
10312 * Update PKCS#7 enveloped data routines to use new API. This is now
10313 supported by any public key method supporting the encrypt operation. A
10314 ctrl is added to allow the public key algorithm to examine or modify
10315 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10320 * Add a ctrl to asn1 method to allow a public key algorithm to express
10321 a default digest type to use. In most cases this will be SHA1 but some
10322 algorithms (such as GOST) need to specify an alternative digest. The
10323 return value indicates how strong the preference is 1 means optional and
10324 2 is mandatory (that is it is the only supported type). Modify
10325 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10326 use the default md. Update openssl utilities to use the default digest
10327 type for signing if it is not explicitly indicated.
10331 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10332 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10333 signing method from the key type. This effectively removes the link
10334 between digests and public key types.
10338 * Add an OID cross reference table and utility functions. Its purpose is to
10339 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10340 rsaEncryption. This will allow some of the algorithm specific hackery
10341 needed to use the correct OID to be removed.
10345 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10346 structures for PKCS7_sign(). They are now set up by the relevant public
10351 * Add provisional EC pkey method with support for ECDSA and ECDH.
10355 * Add support for key derivation (agreement) in the API, DH method and
10360 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10361 public and private key formats. As a side effect these add additional
10362 command line functionality not previously available: DSA signatures can be
10363 generated and verified using pkeyutl and DH key support and generation in
10370 *Oliver Tappe <zooey@hirschkaefer.de>*
10372 * New make target "install_html_docs" installs HTML renditions of the
10375 *Oliver Tappe <zooey@hirschkaefer.de>*
10377 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10378 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10379 support key and parameter generation and add initial key generation
10380 functionality for RSA.
10384 * Add functions for main EVP_PKEY_method operations. The undocumented
10385 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10386 `EVP_PKEY_{encrypt,decrypt}_old`.
10390 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10391 key API, doesn't do much yet.
10395 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10396 public key algorithms. New option to openssl utility:
10397 "list-public-key-algorithms" to print out info.
10401 * Implement the Supported Elliptic Curves Extension for
10402 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10406 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10407 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10411 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10412 utilities such as rsa, dsa, dsaparam etc except they process any key
10417 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10418 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10419 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10424 * Initial support for pluggable public key ASN1.
10425 De-spaghettify the public key ASN1 handling. Move public and private
10426 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10427 algorithm specific handling to a single module within the relevant
10428 algorithm directory. Add functions to allow (near) opaque processing
10429 of public and private key structures.
10433 * Implement the Supported Point Formats Extension for
10434 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10438 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10439 for the psk identity [hint] and the psk callback functions to the
10440 SSL_SESSION, SSL and SSL_CTX structure.
10443 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10447 SSL_CTX_use_psk_identity_hint
10448 SSL_get_psk_identity_hint
10449 SSL_get_psk_identity
10450 SSL_use_psk_identity_hint
10452 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10454 * Add RFC 3161 compliant time stamp request creation, response generation
10455 and response verification functionality.
10457 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10459 * Add initial support for TLS extensions, specifically for the server_name
10460 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10461 have new members for a hostname. The SSL data structure has an
10462 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10463 stored in that context to allow for session resumption, even after the
10464 SSL has been switched to a new SSL_CTX in reaction to a client's
10465 server_name extension.
10467 New functions (subject to change):
10469 SSL_get_servername()
10470 SSL_get_servername_type()
10473 New CTRL codes and macros (subject to change):
10475 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10476 - SSL_CTX_set_tlsext_servername_callback()
10477 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10478 - SSL_CTX_set_tlsext_servername_arg()
10479 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10481 openssl s_client has a new '-servername ...' option.
10483 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10484 '-key2 ...', '-servername_fatal' (subject to change). This allows
10485 testing the HostName extension for a specific single hostname ('-cert'
10486 and '-key' remain fallbacks for handshakes without HostName
10487 negotiation). If the unrecognized_name alert has to be sent, this by
10488 default is a warning; it becomes fatal with the '-servername_fatal'
10491 *Peter Sylvester, Remy Allais, Christophe Renou*
10493 * Whirlpool hash implementation is added.
10497 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10498 bn(64,32). Because of instruction set limitations it doesn't have
10499 any negative impact on performance. This was done mostly in order
10500 to make it possible to share assembler modules, such as bn_mul_mont
10501 implementations, between 32- and 64-bit builds without hassle.
10505 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10506 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10511 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10512 dedicated Montgomery multiplication procedure, is introduced.
10513 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10514 "64-bit" performance on certain 32-bit targets.
10518 * New option SSL_OP_NO_COMP to disable use of compression selectively
10519 in SSL structures. New SSL ctrl to set maximum send fragment size.
10520 Save memory by setting the I/O buffer sizes dynamically instead of
10521 using the maximum available value.
10525 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10526 in addition to the text details.
10530 * Very, very preliminary EXPERIMENTAL support for printing of general
10531 ASN1 structures. This currently produces rather ugly output and doesn't
10532 handle several customised structures at all.
10536 * Integrated support for PVK file format and some related formats such
10537 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10538 these in the 'rsa' and 'dsa' utilities.
10542 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10546 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10547 place for the (very old) "NETSCAPE" format certificates which are now
10548 handled using new ASN1 code equivalents.
10552 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10553 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10554 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10558 * Modify CRL distribution points extension code to print out previously
10559 unsupported fields. Enhance extension setting code to allow setting of
10564 * Add print and set support for Issuing Distribution Point CRL extension.
10568 * Change 'Configure' script to enable Camellia by default.
10575 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10577 * When rejecting SSL/TLS records due to an incorrect version number, never
10578 update s->server with a new major version number. As of
10579 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10580 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10581 the previous behavior could result in a read attempt at NULL when
10582 receiving specific incorrect SSL/TLS records once record payload
10583 protection is active. ([CVE-2010-0740])
10585 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10587 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10588 could be crashed if the relevant tables were not present (e.g. chrooted).
10590 *Tomas Hoger <thoger@redhat.com>*
10592 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10594 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10596 *Martin Olsson, Neel Mehta*
10598 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10599 accommodate for stack sorting, always a write lock!).
10603 * On some versions of WIN32 Heap32Next is very slow. This can cause
10604 excessive delays in the RAND_poll(): over a minute. As a workaround
10605 include a time check in the inner Heap32Next loop too.
10609 * The code that handled flushing of data in SSL/TLS originally used the
10610 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10611 the problem outlined in PR#1949. The fix suggested there however can
10612 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10613 of Apache). So instead simplify the code to flush unconditionally.
10614 This should be fine since flushing with no data to flush is a no op.
10618 * Handle TLS versions 2.0 and later properly and correctly use the
10619 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10620 off ancient servers have a habit of sticking around for a while...
10624 * Modify compression code so it frees up structures without using the
10625 ex_data callbacks. This works around a problem where some applications
10626 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10627 restarting) then use compression (e.g. SSL with compression) later.
10628 This results in significant per-connection memory leaks and
10629 has caused some security issues including CVE-2008-1678 and
10634 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10635 change when encrypting or decrypting.
10639 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10640 connect and renegotiate with servers which do not support RI.
10641 Until RI is more widely deployed this option is enabled by default.
10645 * Add "missing" ssl ctrls to clear options and mode.
10649 * If client attempts to renegotiate and doesn't support RI respond with
10650 a no_renegotiation alert as required by RFC5746. Some renegotiating
10651 TLS clients will continue a connection gracefully when they receive
10652 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10653 waiting for a server hello which it will never receive. Now we treat a
10654 received no_renegotiation alert as a fatal error. This is because
10655 applications requesting a renegotiation might well expect it to succeed
10656 and would have no code in place to handle the server denying it so the
10657 only safe thing to do is to terminate the connection.
10661 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10662 peer supports secure renegotiation and 0 otherwise. Print out peer
10663 renegotiation support in s_client/s_server.
10667 * Replace the highly broken and deprecated SPKAC certification method with
10668 the updated NID creation version. This should correctly handle UTF8.
10672 * Implement RFC5746. Re-enable renegotiation but require the extension
10673 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10674 turns out to be a bad idea. It has been replaced by
10675 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10676 SSL_CTX_set_options(). This is really not recommended unless you
10677 know what you are doing.
10679 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10681 * Fixes to stateless session resumption handling. Use initial_ctx when
10682 issuing and attempting to decrypt tickets in case it has changed during
10683 servername handling. Use a non-zero length session ID when attempting
10684 stateless session resumption: this makes it possible to determine if
10685 a resumption has occurred immediately after receiving server hello
10686 (several places in OpenSSL subtly assume this) instead of later in
10691 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10692 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10693 fixes for a few places where the return code is not checked
10696 *Julia Lawall <julia@diku.dk>*
10698 * Add --strict-warnings option to Configure script to include devteam
10699 warnings in other configurations.
10703 * Add support for --libdir option and LIBDIR variable in makefiles. This
10704 makes it possible to install openssl libraries in locations which
10705 have names other than "lib", for example "/usr/lib64" which some
10708 *Steve Henson, based on patch from Jeremy Utley*
10710 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10711 X690 8.9.12 and can produce some misleading textual output of OIDs.
10713 *Steve Henson, reported by Dan Kaminsky*
10715 * Delete MD2 from algorithm tables. This follows the recommendation in
10716 several standards that it is not used in new applications due to
10717 several cryptographic weaknesses. For binary compatibility reasons
10718 the MD2 API is still compiled in by default.
10722 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10727 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10728 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10731 *Guenter <lists@gknw.net>*
10733 * Fix the server certificate chain building code to use X509_verify_cert(),
10734 it used to have an ad-hoc builder which was unable to cope with anything
10735 other than a simple chain.
10737 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10739 * Don't check self signed certificate signatures in X509_verify_cert()
10740 by default (a flag can override this): it just wastes time without
10741 adding any security. As a useful side effect self signed root CAs
10742 with non-FIPS digests are now usable in FIPS mode.
10746 * In dtls1_process_out_of_seq_message() the check if the current message
10747 is already buffered was missing. For every new message was memory
10748 allocated, allowing an attacker to perform an denial of service attack
10749 with sending out of seq handshake messages until there is no memory
10750 left. Additionally every future message was buffered, even if the
10751 sequence number made no sense and would be part of another handshake.
10752 So only messages with sequence numbers less than 10 in advance will be
10753 buffered. ([CVE-2009-1378])
10755 *Robin Seggelmann, discovered by Daniel Mentz*
10757 * Records are buffered if they arrive with a future epoch to be
10758 processed after finishing the corresponding handshake. There is
10759 currently no limitation to this buffer allowing an attacker to perform
10760 a DOS attack with sending records with future epochs until there is no
10761 memory left. This patch adds the pqueue_size() function to determine
10762 the size of a buffer and limits the record buffer to 100 entries.
10765 *Robin Seggelmann, discovered by Daniel Mentz*
10767 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10768 parent structure is freed. ([CVE-2009-1379])
10772 * Handle non-blocking I/O properly in SSL_shutdown() call.
10774 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10776 * Add `2.5.4.*` OIDs
10778 *Ilya O. <vrghost@gmail.com>*
10780 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10782 * Disable renegotiation completely - this fixes a severe security
10783 problem ([CVE-2009-3555]) at the cost of breaking all
10784 renegotiation. Renegotiation can be re-enabled by setting
10785 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10786 run-time. This is really not recommended unless you know what
10791 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10793 * Don't set val to NULL when freeing up structures, it is freed up by
10794 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10795 zeroing past the valid field. ([CVE-2009-0789])
10797 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10799 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10800 checked correctly. This would allow some invalid signed attributes to
10801 appear to verify correctly. ([CVE-2009-0591])
10803 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10805 * Reject UniversalString and BMPString types with invalid lengths. This
10806 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10807 a legal length. ([CVE-2009-0590])
10811 * Set S/MIME signing as the default purpose rather than setting it
10812 unconditionally. This allows applications to override it at the store
10817 * Permit restricted recursion of ASN1 strings. This is needed in practice
10818 to handle some structures.
10822 * Improve efficiency of mem_gets: don't search whole buffer each time
10825 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10827 * New -hex option for openssl rand.
10831 * Print out UTF8String and NumericString when parsing ASN1.
10835 * Support NumericString type for name components.
10839 * Allow CC in the environment to override the automatically chosen
10840 compiler. Note that nothing is done to ensure flags work with the
10845 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10847 * Properly check EVP_VerifyFinal() and similar return values
10850 *Ben Laurie, Bodo Moeller, Google Security Team*
10852 * Enable TLS extensions by default.
10856 * Allow the CHIL engine to be loaded, whether the application is
10857 multithreaded or not. (This does not release the developer from the
10858 obligation to set up the dynamic locking callbacks.)
10860 *Sander Temme <sander@temme.net>*
10862 * Use correct exit code if there is an error in dgst command.
10864 *Steve Henson; problem pointed out by Roland Dirlewanger*
10866 * Tweak Configure so that you need to say "experimental-jpake" to enable
10867 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10871 * Add experimental JPAKE support, including demo authentication in
10872 s_client and s_server.
10876 * Set the comparison function in v3_addr_canonize().
10878 *Rob Austein <sra@hactrn.net>*
10880 * Add support for XMPP STARTTLS in s_client.
10882 *Philip Paeps <philip@freebsd.org>*
10884 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10885 to ensure that even with this option, only ciphersuites in the
10886 server's preference list will be accepted. (Note that the option
10887 applies only when resuming a session, so the earlier behavior was
10888 just about the algorithm choice for symmetric cryptography.)
10892 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10894 * Fix NULL pointer dereference if a DTLS server received
10895 ChangeCipherSpec as first record ([CVE-2009-1386]).
10899 * Fix a state transition in s3_srvr.c and d1_srvr.c
10900 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10902 *Nagendra Modadugu*
10904 * The fix in 0.9.8c that supposedly got rid of unsafe
10905 double-checked locking was incomplete for RSA blinding,
10906 addressing just one layer of what turns out to have been
10907 doubly unsafe triple-checked locking.
10909 So now fix this for real by retiring the MONT_HELPER macro
10910 in crypto/rsa/rsa_eay.c.
10912 *Bodo Moeller; problem pointed out by Marius Schilder*
10914 * Various precautionary measures:
10916 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10918 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10919 (NB: This would require knowledge of the secret session ticket key
10920 to exploit, in which case you'd be SOL either way.)
10922 - Change bn_nist.c so that it will properly handle input BIGNUMs
10923 outside the expected range.
10925 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10928 *Neel Mehta, Bodo Moeller*
10930 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10931 the load fails. Useful for distros.
10933 *Ben Laurie and the FreeBSD team*
10935 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10939 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10943 * Expand ENGINE to support engine supplied SSL client certificate functions.
10945 This work was sponsored by Logica.
10949 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10950 keystores. Support for SSL/TLS client authentication too.
10951 Not compiled unless enable-capieng specified to Configure.
10953 This work was sponsored by Logica.
10957 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10958 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10959 attribute creation routines such as certificate requests and PKCS#12
10964 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10966 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10967 handshake which could lead to a client crash as found using the
10968 Codenomicon TLS test suite ([CVE-2008-1672])
10970 *Steve Henson, Mark Cox*
10972 * Fix double free in TLS server name extensions which could lead to
10973 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10977 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10979 Clear the error queue to ensure that error entries left from
10980 older function calls do not interfere with the correct operation.
10982 *Lutz Jaenicke, Erik de Castro Lopo*
10984 * Remove root CA certificates of commercial CAs:
10986 The OpenSSL project does not recommend any specific CA and does not
10987 have any policy with respect to including or excluding any CA.
10988 Therefore, it does not make any sense to ship an arbitrary selection
10989 of root CA certificates with the OpenSSL software.
10993 * RSA OAEP patches to fix two separate invalid memory reads.
10994 The first one involves inputs when 'lzero' is greater than
10995 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10996 before the beginning of from). The second one involves inputs where
10997 the 'db' section contains nothing but zeroes (there is a one-byte
10998 invalid read after the end of 'db').
11000 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11002 * Partial backport from 0.9.9-dev:
11004 Introduce bn_mul_mont (dedicated Montgomery multiplication
11005 procedure) as a candidate for BIGNUM assembler implementation.
11006 While 0.9.9-dev uses assembler for various architectures, only
11007 x86_64 is available by default here in the 0.9.8 branch, and
11008 32-bit x86 is available through a compile-time setting.
11010 To try the 32-bit x86 assembler implementation, use Configure
11011 option "enable-montasm" (which exists only for this backport).
11013 As "enable-montasm" for 32-bit x86 disclaims code stability
11014 anyway, in this constellation we activate additional code
11015 backported from 0.9.9-dev for further performance improvements,
11016 namely BN_from_montgomery_word. (To enable this otherwise,
11017 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11019 *Andy Polyakov (backport partially by Bodo Moeller)*
11021 * Add TLS session ticket callback. This allows an application to set
11022 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11023 values. This is useful for key rollover for example where several key
11024 sets may exist with different names.
11028 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11029 This was broken until now in 0.9.8 releases, such that the only way
11030 a registered ENGINE could be used (assuming it initialises
11031 successfully on the host) was to explicitly set it as the default
11032 for the relevant algorithms. This is in contradiction with 0.9.7
11033 behaviour and the documentation. With this fix, when an ENGINE is
11034 registered into a given algorithm's table of implementations, the
11035 'uptodate' flag is reset so that auto-discovery will be used next
11036 time a new context for that algorithm attempts to select an
11039 *Ian Lister (tweaked by Geoff Thorpe)*
11041 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11042 implementation in the following ways:
11044 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11047 Lack of BER streaming support means one pass streaming processing is
11048 only supported if data is detached: setting the streaming flag is
11049 ignored for embedded content.
11051 CMS support is disabled by default and must be explicitly enabled
11052 with the enable-cms configuration option.
11056 * Update the GMP engine glue to do direct copies between BIGNUM and
11057 mpz_t when openssl and GMP use the same limb size. Otherwise the
11058 existing "conversion via a text string export" trick is still used.
11060 *Paul Sheer <paulsheer@gmail.com>*
11062 * Zlib compression BIO. This is a filter BIO which compressed and
11063 uncompresses any data passed through it.
11067 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11068 RFC3394 compatible AES key wrapping.
11072 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11073 sets string data without copying. X509_ALGOR_set0() and
11074 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11075 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11076 from an X509_ATTRIBUTE structure optionally checking it occurs only
11077 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11082 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11083 to get the expected BN_FLG_CONSTTIME behavior.
11085 *Bodo Moeller (Google)*
11089 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11090 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11091 - added some more tests to do_tests.pl
11092 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11093 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11094 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11095 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11096 - various changes to netware.pl to enable gcc-cross builds on Win32
11098 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11099 - various changes to fix missing prototype warnings
11100 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11101 - added AES, WHIRLPOOL and CPUID assembler code to build files
11102 - added missing AES assembler make rules to mk1mf.pl
11103 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11105 *Guenter Knauf <eflash@gmx.net>*
11107 * Implement certificate status request TLS extension defined in RFC3546.
11108 A client can set the appropriate parameters and receive the encoded
11109 OCSP response via a callback. A server can query the supplied parameters
11110 and set the encoded OCSP response in the callback. Add simplified examples
11111 to s_client and s_server.
11115 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11117 * Fix various bugs:
11118 + Binary incompatibility of ssl_ctx_st structure
11119 + DTLS interoperation with non-compliant servers
11120 + Don't call get_session_cb() without proposed session
11121 + Fix ia64 assembler code
11123 *Andy Polyakov, Steve Henson*
11125 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11127 * DTLS Handshake overhaul. There were longstanding issues with
11128 OpenSSL DTLS implementation, which were making it impossible for
11129 RFC 4347 compliant client to communicate with OpenSSL server.
11130 Unfortunately just fixing these incompatibilities would "cut off"
11131 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11132 server keeps tolerating non RFC compliant syntax. The opposite is
11133 not true, 0.9.8f client can not communicate with earlier server.
11134 This update even addresses CVE-2007-4995.
11138 * Changes to avoid need for function casts in OpenSSL: some compilers
11139 (gcc 4.2 and later) reject their use.
11140 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11143 * Add RFC4507 support to OpenSSL. This includes the corrections in
11144 RFC4507bis. The encrypted ticket format is an encrypted encoded
11145 SSL_SESSION structure, that way new session features are automatically
11148 If a client application caches session in an SSL_SESSION structure
11149 support is transparent because tickets are now stored in the encoded
11152 The SSL_CTX structure automatically generates keys for ticket
11153 protection in servers so again support should be possible
11154 with no application modification.
11156 If a client or server wishes to disable RFC4507 support then the option
11157 SSL_OP_NO_TICKET can be set.
11159 Add a TLS extension debugging callback to allow the contents of any client
11160 or server extensions to be examined.
11162 This work was sponsored by Google.
11166 * Add initial support for TLS extensions, specifically for the server_name
11167 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11168 have new members for a hostname. The SSL data structure has an
11169 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11170 stored in that context to allow for session resumption, even after the
11171 SSL has been switched to a new SSL_CTX in reaction to a client's
11172 server_name extension.
11174 New functions (subject to change):
11176 SSL_get_servername()
11177 SSL_get_servername_type()
11180 New CTRL codes and macros (subject to change):
11182 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11183 - SSL_CTX_set_tlsext_servername_callback()
11184 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11185 - SSL_CTX_set_tlsext_servername_arg()
11186 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11188 openssl s_client has a new '-servername ...' option.
11190 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11191 '-key2 ...', '-servername_fatal' (subject to change). This allows
11192 testing the HostName extension for a specific single hostname ('-cert'
11193 and '-key' remain fallbacks for handshakes without HostName
11194 negotiation). If the unrecognized_name alert has to be sent, this by
11195 default is a warning; it becomes fatal with the '-servername_fatal'
11198 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11200 * Add AES and SSE2 assembly language support to VC++ build.
11204 * Mitigate attack on final subtraction in Montgomery reduction.
11208 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11209 (which previously caused an internal error).
11213 * Squeeze another 10% out of IGE mode when in != out.
11217 * AES IGE mode speedup.
11219 *Dean Gaudet (Google)*
11221 * Add the Korean symmetric 128-bit cipher SEED (see
11222 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11223 add SEED ciphersuites from RFC 4162:
11225 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11226 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11227 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11228 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11230 To minimize changes between patchlevels in the OpenSSL 0.9.8
11231 series, SEED remains excluded from compilation unless OpenSSL
11232 is configured with 'enable-seed'.
11234 *KISA, Bodo Moeller*
11236 * Mitigate branch prediction attacks, which can be practical if a
11237 single processor is shared, allowing a spy process to extract
11238 information. For detailed background information, see
11239 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11240 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11241 and Necessary Software Countermeasures"). The core of the change
11242 are new versions BN_div_no_branch() and
11243 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11244 respectively, which are slower, but avoid the security-relevant
11245 conditional branches. These are automatically called by BN_div()
11246 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11247 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11248 remove a conditional branch.
11250 BN_FLG_CONSTTIME is the new name for the previous
11251 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11252 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11253 in the exponent causes BN_mod_exp_mont() to use the alternative
11254 implementation in BN_mod_exp_mont_consttime().) The old name
11255 remains as a deprecated alias.
11257 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11258 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11259 constant-time implementations for more than just exponentiation.
11260 Here too the old name is kept as a deprecated alias.
11262 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11263 the BN_BLINDING structure gets an independent copy of the
11264 modulus. This means that the previous `BIGNUM *m` argument to
11265 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11266 essentially becomes `const BIGNUM *m`, although we can't actually
11267 change this in the header file before 0.9.9. It allows
11268 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11269 enable BN_FLG_CONSTTIME.
11271 *Matthew D Wood (Intel Corp)*
11273 * In the SSL/TLS server implementation, be strict about session ID
11274 context matching (which matters if an application uses a single
11275 external cache for different purposes). Previously,
11276 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11277 set. This did ensure strict client verification, but meant that,
11278 with applications using a single external cache for quite
11279 different requirements, clients could circumvent ciphersuite
11280 restrictions for a given session ID context by starting a session
11281 in a different context.
11285 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11286 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11287 authentication-only ciphersuites.
11291 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11292 not complete and could lead to a possible single byte overflow
11293 ([CVE-2007-5135]) [Ben Laurie]
11295 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11297 * Since AES128 and AES256 (and similarly Camellia128 and
11298 Camellia256) share a single mask bit in the logic of
11299 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11300 kludge to work properly if AES128 is available and AES256 isn't
11301 (or if Camellia128 is available and Camellia256 isn't).
11305 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11306 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11307 When a point or a seed is encoded in a BIT STRING, we need to
11308 prevent the removal of trailing zero bits to get the proper DER
11309 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11310 of a NamedBitList, for which trailing 0 bits need to be removed.)
11314 * Have SSL/TLS server implementation tolerate "mismatched" record
11315 protocol version while receiving ClientHello even if the
11316 ClientHello is fragmented. (The server can't insist on the
11317 particular protocol version it has chosen before the ServerHello
11318 message has informed the client about his choice.)
11322 * Add RFC 3779 support.
11324 *Rob Austein for ARIN, Ben Laurie*
11326 * Load error codes if they are not already present instead of using a
11327 static variable. This allows them to be cleanly unloaded and reloaded.
11328 Improve header file function name parsing.
11332 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11333 or CAPABILITY handshake as required by RFCs.
11335 *Goetz Babin-Ebell*
11337 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11339 * Introduce limits to prevent malicious keys being able to
11340 cause a denial of service. ([CVE-2006-2940])
11342 *Steve Henson, Bodo Moeller*
11344 * Fix ASN.1 parsing of certain invalid structures that can result
11345 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11347 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11348 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11350 * Fix SSL client code which could crash if connecting to a
11351 malicious SSLv2 server. ([CVE-2006-4343])
11353 *Tavis Ormandy and Will Drewry, Google Security Team*
11355 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11356 match only those. Before that, "AES256-SHA" would be interpreted
11357 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11358 the same strength classification in 0.9.7h) as we currently only
11359 have a single AES bit in the ciphersuite description bitmap.
11360 That change, however, also applied to ciphersuite strings such as
11361 "RC4-MD5" that intentionally matched multiple ciphersuites --
11362 namely, SSL 2.0 ciphersuites in addition to the more common ones
11363 from SSL 3.0/TLS 1.0.
11365 So we change the selection algorithm again: Naming an explicit
11366 ciphersuite selects this one ciphersuite, and any other similar
11367 ciphersuite (same bitmap) from *other* protocol versions.
11368 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11369 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11371 Since SSL 2.0 does not have any ciphersuites for which the
11372 128/256 bit distinction would be relevant, this works for now.
11373 The proper fix will be to use different bits for AES128 and
11374 AES256, which would have avoided the problems from the beginning;
11375 however, bits are scarce, so we can only do this in a new release
11376 (not just a patchlevel) when we can change the SSL_CIPHER
11377 definition to split the single 'unsigned long mask' bitmap into
11378 multiple values to extend the available space.
11382 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11384 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11385 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11387 * Add AES IGE and biIGE modes.
11391 * Change the Unix randomness entropy gathering to use poll() when
11392 possible instead of select(), since the latter has some
11393 undesirable limitations.
11395 *Darryl Miles via Richard Levitte and Bodo Moeller*
11397 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11398 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11399 cannot be implicitly activated as part of, e.g., the "AES" alias.
11400 However, please upgrade to OpenSSL 0.9.9[-dev] for
11401 non-experimental use of the ECC ciphersuites to get TLS extension
11402 support, which is required for curve and point format negotiation
11403 to avoid potential handshake problems.
11407 * Disable rogue ciphersuites:
11409 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11410 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11411 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11413 The latter two were purportedly from
11414 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11417 Also deactivate the remaining ciphersuites from
11418 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11419 unofficial, and the ID has long expired.
11423 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11424 dual-core machines) and other potential thread-safety issues.
11428 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11429 versions), which is now available for royalty-free use
11430 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11431 Also, add Camellia TLS ciphersuites from RFC 4132.
11433 To minimize changes between patchlevels in the OpenSSL 0.9.8
11434 series, Camellia remains excluded from compilation unless OpenSSL
11435 is configured with 'enable-camellia'.
11439 * Disable the padding bug check when compression is in use. The padding
11440 bug check assumes the first packet is of even length, this is not
11441 necessarily true if compression is enabled and can result in false
11442 positives causing handshake failure. The actual bug test is ancient
11443 code so it is hoped that implementations will either have fixed it by
11444 now or any which still have the bug do not support compression.
11448 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11450 * When applying a cipher rule check to see if string match is an explicit
11451 cipher suite and only match that one cipher suite if it is.
11455 * Link in manifests for VC++ if needed.
11457 *Austin Ziegler <halostatue@gmail.com>*
11459 * Update support for ECC-based TLS ciphersuites according to
11460 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11461 TLS extensions, which are supported starting with the 0.9.9
11462 branch, not in the OpenSSL 0.9.8 branch).
11466 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11467 opaque EVP_CIPHER_CTX handling.
11471 * Fixes and enhancements to zlib compression code. We now only use
11472 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11473 to conform with the standards mentioned here:
11474 <http://www.zlib.net/DLL_FAQ.txt>
11475 Static zlib linking now works on Windows and the new --with-zlib-include
11476 --with-zlib-lib options to Configure can be used to supply the location
11477 of the headers and library. Gracefully handle case where zlib library
11482 * Several fixes and enhancements to the OID generation code. The old code
11483 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11484 handle numbers larger than ULONG_MAX, truncated printing and had a
11485 non standard OBJ_obj2txt() behaviour.
11489 * Add support for building of engines under engine/ as shared libraries
11490 under VC++ build system.
11494 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11495 Hopefully, we will not see any false combination of paths any more.
11499 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11501 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11502 (part of SSL_OP_ALL). This option used to disable the
11503 countermeasure against man-in-the-middle protocol-version
11504 rollback in the SSL 2.0 server implementation, which is a bad
11505 idea. ([CVE-2005-2969])
11507 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11508 for Information Security, National Institute of Advanced Industrial
11509 Science and Technology [AIST], Japan)*
11511 * Add two function to clear and return the verify parameter flags.
11515 * Keep cipherlists sorted in the source instead of sorting them at
11516 runtime, thus removing the need for a lock.
11520 * Avoid some small subgroup attacks in Diffie-Hellman.
11522 *Nick Mathewson and Ben Laurie*
11524 * Add functions for well-known primes.
11528 * Extended Windows CE support.
11530 *Satoshi Nakamura and Andy Polyakov*
11532 * Initialize SSL_METHOD structures at compile time instead of during
11533 runtime, thus removing the need for a lock.
11537 * Make PKCS7_decrypt() work even if no certificate is supplied by
11538 attempting to decrypt each encrypted key in turn. Add support to
11543 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11545 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11548 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11552 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11553 key into the same file any more.
11557 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11561 * Add -utf8 command line and config file option to 'ca'.
11563 *Stefan <stf@udoma.org*
11565 * Removed the macro des_crypt(), as it seems to conflict with some
11566 libraries. Use DES_crypt().
11570 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11571 involves renaming the source and generated shared-libs for
11572 both. The engines will accept the corrected or legacy ids
11573 ('ncipher' and '4758_cca' respectively) when binding. NB,
11574 this only applies when building 'shared'.
11576 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11578 * Add attribute functions to EVP_PKEY structure. Modify
11579 PKCS12_create() to recognize a CSP name attribute and
11580 use it. Make -CSP option work again in pkcs12 utility.
11584 * Add new functionality to the bn blinding code:
11585 - automatic re-creation of the BN_BLINDING parameters after
11586 a fixed number of uses (currently 32)
11587 - add new function for parameter creation
11588 - introduce flags to control the update behaviour of the
11589 BN_BLINDING parameters
11590 - hide BN_BLINDING structure
11591 Add a second BN_BLINDING slot to the RSA structure to improve
11592 performance when a single RSA object is shared among several
11597 * Add support for DTLS.
11599 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11601 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11602 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11606 * Remove buggy and incomplete DH cert support from
11607 ssl/ssl_rsa.c and ssl/s3_both.c
11611 * Use SHA-1 instead of MD5 as the default digest algorithm for
11612 the `apps/openssl` commands.
11616 * Compile clean with "-Wall -Wmissing-prototypes
11617 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11618 DEBUG_SAFESTACK must also be set.
11622 * Change ./Configure so that certain algorithms can be disabled by default.
11623 The new counterpiece to "no-xxx" is "enable-xxx".
11625 The patented RC5 and MDC2 algorithms will now be disabled unless
11626 "enable-rc5" and "enable-mdc2", respectively, are specified.
11628 (IDEA remains enabled despite being patented. This is because IDEA
11629 is frequently required for interoperability, and there is no license
11630 fee for non-commercial use. As before, "no-idea" can be used to
11631 avoid this algorithm.)
11635 * Add processing of proxy certificates (see RFC 3820). This work was
11636 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11637 EGEE (Enabling Grids for E-science in Europe).
11641 * RC4 performance overhaul on modern architectures/implementations, such
11642 as Intel P4, IA-64 and AMD64.
11646 * New utility extract-section.pl. This can be used specify an alternative
11647 section number in a pod file instead of having to treat each file as
11648 a separate case in Makefile. This can be done by adding two lines to the
11651 =for comment openssl_section:XXX
11653 The blank line is mandatory.
11657 * New arguments -certform, -keyform and -pass for s_client and s_server
11658 to allow alternative format key and certificate files and passphrase
11663 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11664 update associated structures and add various utility functions.
11666 Add new policy related verify parameters, include policy checking in
11667 standard verify code. Enhance 'smime' application with extra parameters
11668 to support policy checking and print out.
11672 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11673 Nehemiah processors. These extensions support AES encryption in hardware
11674 as well as RNG (though RNG support is currently disabled).
11676 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11678 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11682 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11684 *Andy Polyakov and a number of other people*
11686 * Improved PowerPC platform support. Most notably BIGNUM assembler
11687 implementation contributed by IBM.
11689 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11691 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11692 exponent rather than 'unsigned long'. There is a corresponding change to
11693 the new 'rsa_keygen' element of the RSA_METHOD structure.
11695 *Jelte Jansen, Geoff Thorpe*
11697 * Functionality for creating the initial serial number file is now
11698 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11700 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11701 number file to 1, which is bound to cause problems. To avoid
11702 the problems while respecting compatibility between different 0.9.7
11703 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11704 CA.pl for serial number initialization. With the new release 0.9.8,
11705 we can fix the problem directly in the 'ca' utility.)
11709 * Reduced header interdependencies by declaring more opaque objects in
11710 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11711 give fewer recursive includes, which could break lazy source code - so
11712 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11713 developers should define this symbol when building and using openssl to
11714 ensure they track the recommended behaviour, interfaces, [etc], but
11715 backwards-compatible behaviour prevails when this isn't defined.
11719 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11723 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11724 This will generate a random key of the appropriate length based on the
11725 cipher context. The EVP_CIPHER can provide its own random key generation
11726 routine to support keys of a specific form. This is used in the des and
11727 3des routines to generate a key of the correct parity. Update S/MIME
11728 code to use new functions and hence generate correct parity DES keys.
11729 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11730 valid (weak or incorrect parity).
11734 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11735 as looking them up. This is useful when the verified structure may contain
11736 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11737 present unless the new PKCS7_NO_CRL flag is asserted.
11741 * Extend ASN1 oid configuration module. It now additionally accepts the
11744 shortName = some long name, 1.2.3.4
11748 * Reimplemented the BN_CTX implementation. There is now no more static
11749 limitation on the number of variables it can handle nor the depth of the
11750 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11751 information can now expand as required, and rather than having a single
11752 static array of bignums, BN_CTX now uses a linked-list of such arrays
11753 allowing it to expand on demand whilst maintaining the usefulness of
11754 BN_CTX's "bundling".
11758 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11759 to allow all RSA operations to function using a single BN_CTX.
11763 * Preliminary support for certificate policy evaluation and checking. This
11764 is initially intended to pass the tests outlined in "Conformance Testing
11765 of Relying Party Client Certificate Path Processing Logic" v1.07.
11769 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11770 remained unused and not that useful. A variety of other little bignum
11771 tweaks and fixes have also been made continuing on from the audit (see
11776 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11777 associated ASN1, EVP and SSL functions and old ASN1 macros.
11781 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11782 and this should never fail. So the return value from the use of
11783 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11784 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11788 * BN_CTX_get() should return zero-valued bignums, providing the same
11789 initialised value as BN_new().
11791 *Geoff Thorpe, suggested by Ulf Möller*
11793 * Support for inhibitAnyPolicy certificate extension.
11797 * An audit of the BIGNUM code is underway, for which debugging code is
11798 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11799 is considered valid when processing BIGNUMs, and causes execution to
11800 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11801 further steps are taken to deliberately pollute unused data in BIGNUM
11802 structures to try and expose faulty code further on. For now, openssl will
11803 (in its default mode of operation) continue to tolerate the inconsistent
11804 forms that it has tolerated in the past, but authors and packagers should
11805 consider trying openssl and their own applications when compiled with
11806 these debugging symbols defined. It will help highlight potential bugs in
11807 their own code, and will improve the test coverage for OpenSSL itself. At
11808 some point, these tighter rules will become openssl's default to improve
11809 maintainability, though the assert()s and other overheads will remain only
11810 in debugging configurations. See bn.h for more details.
11812 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11814 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11815 that can only be obtained through BN_CTX_new() (which implicitly
11816 initialises it). The presence of this function only made it possible
11817 to overwrite an existing structure (and cause memory leaks).
11821 * Because of the callback-based approach for implementing LHASH as a
11822 template type, lh_insert() adds opaque objects to hash-tables and
11823 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11824 to clean up those corresponding objects before destroying the hash table
11825 (and losing the object pointers). So some over-zealous constifications in
11826 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11827 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11828 prototyped to have "const" restrictions on the object pointers they are
11829 given (and so aren't required to cast them away any more).
11833 * The tmdiff.h API was so ugly and minimal that our own timing utility
11834 (speed) prefers to use its own implementation. The two implementations
11835 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11836 its object type properly exposed (MS_TM) instead of casting to/from
11837 `char *`. This may still change yet if someone realises MS_TM and
11839 aren't necessarily the greatest nomenclatures - but this is what was used
11840 internally to the implementation so I've used that for now.
11844 * Ensure that deprecated functions do not get compiled when
11845 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11846 the self-tests were still using deprecated key-generation functions so
11847 these have been updated also.
11851 * Reorganise PKCS#7 code to separate the digest location functionality
11852 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11853 New function PKCS7_set_digest() to set the digest type for PKCS#7
11854 digestedData type. Add additional code to correctly generate the
11855 digestedData type and add support for this type in PKCS7 initialization
11860 * New function PKCS7_set0_type_other() this initializes a PKCS7
11861 structure of type "other".
11865 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11866 sure the loop does correctly stop and breaking ("division by zero")
11867 modulus operations are not performed. The (pre-generated) prime
11868 table crypto/bn/bn_prime.h was already correct, but it could not be
11869 re-generated on some platforms because of the "division by zero"
11870 situation in the script.
11872 *Ralf S. Engelschall*
11874 * Update support for ECC-based TLS ciphersuites according to
11875 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11876 SHA-1 now is only used for "small" curves (where the
11877 representation of a field element takes up to 24 bytes); for
11878 larger curves, the field element resulting from ECDH is directly
11879 used as premaster secret.
11881 *Douglas Stebila (Sun Microsystems Laboratories)*
11883 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11884 curve secp160r1 to the tests.
11886 *Douglas Stebila (Sun Microsystems Laboratories)*
11888 * Add the possibility to load symbols globally with DSO.
11890 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11892 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11893 control of the error stack.
11897 * Add support for STORE in ENGINE.
11901 * Add the STORE type. The intention is to provide a common interface
11902 to certificate and key stores, be they simple file-based stores, or
11903 HSM-type store, or LDAP stores, or...
11904 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11908 * Add a generic structure called OPENSSL_ITEM. This can be used to
11909 pass a list of arguments to any function as well as provide a way
11910 for a function to pass data back to the caller.
11914 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11915 works like BUF_strdup() but can be used to duplicate a portion of
11916 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11921 * Add the function sk_find_ex() which works like sk_find(), but will
11922 return an index to an element even if an exact match couldn't be
11923 found. The index is guaranteed to point at the element where the
11924 searched-for key would be inserted to preserve sorting order.
11928 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11929 takes an extra flags argument for optional functionality. Currently,
11930 the following flags are defined:
11932 OBJ_BSEARCH_VALUE_ON_NOMATCH
11933 This one gets OBJ_bsearch_ex() to return a pointer to the first
11934 element where the comparing function returns a negative or zero
11937 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11938 This one gets OBJ_bsearch_ex() to return a pointer to the first
11939 element where the comparing function returns zero. This is useful
11940 if there are more than one element where the comparing function
11945 * Make it possible to create self-signed certificates with 'openssl ca'
11946 in such a way that the self-signed certificate becomes part of the
11947 CA database and uses the same mechanisms for serial number generation
11948 as all other certificate signing. The new flag '-selfsign' enables
11949 this functionality. Adapt CA.sh and CA.pl.in.
11953 * Add functionality to check the public key of a certificate request
11954 against a given private. This is useful to check that a certificate
11955 request can be signed by that key (self-signing).
11959 * Make it possible to have multiple active certificates with the same
11960 subject in the CA index file. This is done only if the keyword
11961 'unique_subject' is set to 'no' in the main CA section (default
11962 if 'CA_default') of the configuration file. The value is saved
11963 with the database itself in a separate index attribute file,
11964 named like the index file with '.attr' appended to the name.
11968 * Generate multi-valued AVAs using '+' notation in config files for
11973 * Support for nameConstraints certificate extension.
11977 * Support for policyConstraints certificate extension.
11981 * Support for policyMappings certificate extension.
11985 * Make sure the default DSA_METHOD implementation only uses its
11986 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11987 and change its own handlers to be NULL so as to remove unnecessary
11988 indirection. This lets alternative implementations fallback to the
11989 default implementation more easily.
11993 * Support for directoryName in GeneralName related extensions
11998 * Make it possible to link applications using Makefile.shared.
11999 Make that possible even when linking against static libraries!
12003 * Support for single pass processing for S/MIME signing. This now
12004 means that S/MIME signing can be done from a pipe, in addition
12005 cleartext signing (multipart/signed type) is effectively streaming
12006 and the signed data does not need to be all held in memory.
12008 This is done with a new flag PKCS7_STREAM. When this flag is set
12009 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12010 is done after the data is output (and digests calculated) in
12011 SMIME_write_PKCS7().
12015 * Add full support for -rpath/-R, both in shared libraries and
12016 applications, at least on the platforms where it's known how
12021 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12022 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12023 will now compute a table of multiples of the generator that
12024 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12025 faster (notably in the case of a single point multiplication,
12026 scalar * generator).
12028 *Nils Larsch, Bodo Moeller*
12030 * IPv6 support for certificate extensions. The various extensions
12031 which use the IP:a.b.c.d can now take IPv6 addresses using the
12032 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12037 * Added an ENGINE that implements RSA by performing private key
12038 exponentiations with the GMP library. The conversions to and from
12039 GMP's mpz_t format aren't optimised nor are any montgomery forms
12040 cached, and on x86 it appears OpenSSL's own performance has caught up.
12041 However there are likely to be other architectures where GMP could
12042 provide a boost. This ENGINE is not built in by default, but it can be
12043 specified at Configure time and should be accompanied by the necessary
12044 linker additions, eg;
12045 ./config -DOPENSSL_USE_GMP -lgmp
12049 * "openssl engine" will not display ENGINE/DSO load failure errors when
12050 testing availability of engines with "-t" - the old behaviour is
12051 produced by increasing the feature's verbosity with "-tt".
12055 * ECDSA routines: under certain error conditions uninitialized BN objects
12056 could be freed. Solution: make sure initialization is performed early
12057 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12062 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12063 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12064 software implementations. For DSA and DH, parameter generation can
12065 also be overridden by providing the appropriate method callbacks.
12069 * Change the "progress" mechanism used in key-generation and
12070 primality testing to functions that take a new BN_GENCB pointer in
12071 place of callback/argument pairs. The new API functions have `_ex`
12072 postfixes and the older functions are reimplemented as wrappers for
12073 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12074 declarations of the old functions to help (graceful) attempts to
12075 migrate to the new functions. Also, the new key-generation API
12076 functions operate on a caller-supplied key-structure and return
12077 success/failure rather than returning a key or NULL - this is to
12078 help make "keygen" another member function of RSA_METHOD etc.
12080 Example for using the new callback interface:
12082 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12083 void *my_arg = ...;
12086 BN_GENCB_set(&my_cb, my_callback, my_arg);
12088 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12089 /* For the meaning of a, b in calls to my_callback(), see the
12090 * documentation of the function that calls the callback.
12091 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12092 * my_callback should return 1 if it wants BN_is_prime_ex()
12093 * to continue, or 0 to stop.
12098 * Change the ZLIB compression method to be stateful, and make it
12099 available to TLS with the number defined in
12100 draft-ietf-tls-compression-04.txt.
12104 * Add the ASN.1 structures and functions for CertificatePair, which
12105 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12107 CertificatePair ::= SEQUENCE {
12108 forward [0] Certificate OPTIONAL,
12109 reverse [1] Certificate OPTIONAL,
12110 -- at least one of the pair shall be present -- }
12112 Also implement the PEM functions to read and write certificate
12113 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12115 This needed to be defined, mostly for the sake of the LDAP
12116 attribute crossCertificatePair, but may prove useful elsewhere as
12121 * Make it possible to inhibit symlinking of shared libraries in
12122 Makefile.shared, for Cygwin's sake.
12126 * Extend the BIGNUM API by creating a function
12127 void BN_set_negative(BIGNUM *a, int neg);
12128 and a macro that behave like
12129 int BN_is_negative(const BIGNUM *a);
12131 to avoid the need to access 'a->neg' directly in applications.
12135 * Implement fast modular reduction for pseudo-Mersenne primes
12136 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12137 EC_GROUP_new_curve_GFp() will now automatically use this
12140 *Nils Larsch <nla@trustcenter.de>*
12142 * Add new lock type (CRYPTO_LOCK_BN).
12146 * Change the ENGINE framework to automatically load engines
12147 dynamically from specific directories unless they could be
12148 found to already be built in or loaded. Move all the
12149 current engines except for the cryptodev one to a new
12150 directory engines/.
12151 The engines in engines/ are built as shared libraries if
12152 the "shared" options was given to ./Configure or ./config.
12153 Otherwise, they are inserted in libcrypto.a.
12154 /usr/local/ssl/engines is the default directory for dynamic
12155 engines, but that can be overridden at configure time through
12156 the usual use of --prefix and/or --openssldir, and at run
12157 time with the environment variable OPENSSL_ENGINES.
12159 *Geoff Thorpe and Richard Levitte*
12161 * Add Makefile.shared, a helper makefile to build shared
12162 libraries. Adapt Makefile.org.
12166 * Add version info to Win32 DLLs.
12168 *Peter 'Luna' Runestig" <peter@runestig.com>*
12170 * Add new 'medium level' PKCS#12 API. Certificates and keys
12171 can be added using this API to created arbitrary PKCS#12
12172 files while avoiding the low-level API.
12174 New options to PKCS12_create(), key or cert can be NULL and
12175 will then be omitted from the output file. The encryption
12176 algorithm NIDs can be set to -1 for no encryption, the mac
12177 iteration count can be set to 0 to omit the mac.
12179 Enhance pkcs12 utility by making the -nokeys and -nocerts
12180 options work when creating a PKCS#12 file. New option -nomac
12181 to omit the mac, NONE can be set for an encryption algorithm.
12182 New code is modified to use the enhanced PKCS12_create()
12183 instead of the low-level API.
12187 * Extend ASN1 encoder to support indefinite length constructed
12188 encoding. This can output sequences tags and octet strings in
12189 this form. Modify pk7_asn1.c to support indefinite length
12190 encoding. This is experimental and needs additional code to
12191 be useful, such as an ASN1 bio and some enhanced streaming
12194 Extend template encode functionality so that tagging is passed
12195 down to the template encoder.
12199 * Let 'openssl req' fail if an argument to '-newkey' is not
12200 recognized instead of using RSA as a default.
12204 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12205 As these are not official, they are not included in "ALL";
12206 the "ECCdraft" ciphersuite group alias can be used to select them.
12208 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12210 * Add ECDH engine support.
12212 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12214 * Add ECDH in new directory crypto/ecdh/.
12216 *Douglas Stebila (Sun Microsystems Laboratories)*
12218 * Let BN_rand_range() abort with an error after 100 iterations
12219 without success (which indicates a broken PRNG).
12223 * Change BN_mod_sqrt() so that it verifies that the input value
12224 is really the square of the return value. (Previously,
12225 BN_mod_sqrt would show GIGO behaviour.)
12229 * Add named elliptic curves over binary fields from X9.62, SECG,
12230 and WAP/WTLS; add OIDs that were still missing.
12232 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12234 * Extend the EC library for elliptic curves over binary fields
12235 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12238 EC_GF2m_simple_method
12242 EC_GROUP_new_curve_GF2m
12243 EC_GROUP_set_curve_GF2m
12244 EC_GROUP_get_curve_GF2m
12245 EC_POINT_set_affine_coordinates_GF2m
12246 EC_POINT_get_affine_coordinates_GF2m
12247 EC_POINT_set_compressed_coordinates_GF2m
12249 Point compression for binary fields is disabled by default for
12250 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12253 As binary polynomials are represented as BIGNUMs, various members
12254 of the EC_GROUP and EC_POINT data structures can be shared
12255 between the implementations for prime fields and binary fields;
12256 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12257 are essentially identical to their `..._GFp` counterparts.
12258 (For simplicity, the `..._GFp` prefix has been dropped from
12259 various internal method names.)
12261 An internal 'field_div' method (similar to 'field_mul' and
12262 'field_sqr') has been added; this is used only for binary fields.
12264 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12266 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12267 through methods ('mul', 'precompute_mult').
12269 The generic implementations (now internally called 'ec_wNAF_mul'
12270 and 'ec_wNAF_precomputed_mult') remain the default if these
12271 methods are undefined.
12273 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12275 * New function EC_GROUP_get_degree, which is defined through
12276 EC_METHOD. For curves over prime fields, this returns the bit
12277 length of the modulus.
12279 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12281 * New functions EC_GROUP_dup, EC_POINT_dup.
12282 (These simply call ..._new and ..._copy).
12284 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12286 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12287 Polynomials are represented as BIGNUMs (where the sign bit is not
12288 used) in the following functions [macros]:
12291 BN_GF2m_sub [= BN_GF2m_add]
12292 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12293 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12294 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12296 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12297 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12298 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12299 BN_GF2m_cmp [= BN_ucmp]
12301 (Note that only the 'mod' functions are actually for fields GF(2^m).
12302 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12304 For some functions, an the irreducible polynomial defining a
12305 field can be given as an 'unsigned int[]' with strictly
12306 decreasing elements giving the indices of those bits that are set;
12307 i.e., p[] represents the polynomial
12308 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12310 p[0] > p[1] > ... > p[k] = 0.
12311 This applies to the following functions:
12314 BN_GF2m_mod_mul_arr
12315 BN_GF2m_mod_sqr_arr
12316 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12317 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12318 BN_GF2m_mod_exp_arr
12319 BN_GF2m_mod_sqrt_arr
12320 BN_GF2m_mod_solve_quad_arr
12324 Conversion can be performed by the following functions:
12329 bntest.c has additional tests for binary polynomial arithmetic.
12331 Two implementations for BN_GF2m_mod_div() are available.
12332 The default algorithm simply uses BN_GF2m_mod_inv() and
12333 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12334 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12335 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12337 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12339 * Add new error code 'ERR_R_DISABLED' that can be used when some
12340 functionality is disabled at compile-time.
12342 *Douglas Stebila <douglas.stebila@sun.com>*
12344 * Change default behaviour of 'openssl asn1parse' so that more
12345 information is visible when viewing, e.g., a certificate:
12347 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12348 mode the content of non-printable OCTET STRINGs is output in a
12349 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12350 avoid the appearance of a printable string.
12352 *Nils Larsch <nla@trustcenter.de>*
12354 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12356 EC_GROUP_set_asn1_flag()
12357 EC_GROUP_get_asn1_flag()
12358 EC_GROUP_set_point_conversion_form()
12359 EC_GROUP_get_point_conversion_form()
12360 These control ASN1 encoding details:
12361 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12362 has been set to OPENSSL_EC_NAMED_CURVE.
12363 - Points are encoded in uncompressed form by default; options for
12364 asn1_for are as for point2oct, namely
12365 POINT_CONVERSION_COMPRESSED
12366 POINT_CONVERSION_UNCOMPRESSED
12367 POINT_CONVERSION_HYBRID
12369 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12371 EC_GROUP_set_seed()
12372 EC_GROUP_get0_seed()
12373 EC_GROUP_get_seed_len()
12374 This is used only for ASN1 purposes (so far).
12376 *Nils Larsch <nla@trustcenter.de>*
12378 * Add 'field_type' member to EC_METHOD, which holds the NID
12379 of the appropriate field type OID. The new function
12380 EC_METHOD_get_field_type() returns this value.
12382 *Nils Larsch <nla@trustcenter.de>*
12385 EC_POINT_point2bn()
12386 EC_POINT_bn2point()
12387 EC_POINT_point2hex()
12388 EC_POINT_hex2point()
12389 providing useful interfaces to EC_POINT_point2oct() and
12390 EC_POINT_oct2point().
12392 *Nils Larsch <nla@trustcenter.de>*
12394 * Change internals of the EC library so that the functions
12395 EC_GROUP_set_generator()
12396 EC_GROUP_get_generator()
12397 EC_GROUP_get_order()
12398 EC_GROUP_get_cofactor()
12399 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12400 to methods, which would lead to unnecessary code duplication when
12401 adding different types of curves.
12403 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12405 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12406 arithmetic, and such that modified wNAFs are generated
12407 (which avoid length expansion in many cases).
12411 * Add a function EC_GROUP_check_discriminant() (defined via
12412 EC_METHOD) that verifies that the curve discriminant is non-zero.
12414 Add a function EC_GROUP_check() that makes some sanity tests
12415 on a EC_GROUP, its generator and order. This includes
12416 EC_GROUP_check_discriminant().
12418 *Nils Larsch <nla@trustcenter.de>*
12420 * Add ECDSA in new directory crypto/ecdsa/.
12422 Add applications 'openssl ecparam' and 'openssl ecdsa'
12423 (these are based on 'openssl dsaparam' and 'openssl dsa').
12425 ECDSA support is also included in various other files across the
12426 library. Most notably,
12427 - 'openssl req' now has a '-newkey ecdsa:file' option;
12428 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12429 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12430 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12431 them suitable for ECDSA where domain parameters must be
12432 extracted before the specific public key;
12433 - ECDSA engine support has been added.
12435 *Nils Larsch <nla@trustcenter.de>*
12437 * Include some named elliptic curves, and add OIDs from X9.62,
12438 SECG, and WAP/WTLS. Each curve can be obtained from the new
12440 EC_GROUP_new_by_curve_name(),
12441 and the list of available named curves can be obtained with
12442 EC_get_builtin_curves().
12443 Also add a 'curve_name' member to EC_GROUP objects, which can be
12445 EC_GROUP_set_curve_name()
12446 EC_GROUP_get_curve_name()
12448 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12450 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12451 was actually never needed) and in BN_mul(). The removal in BN_mul()
12452 required a small change in bn_mul_part_recursive() and the addition
12453 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12454 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12455 bn_sub_words() and bn_add_words() except they take arrays with
12460 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12462 * Cleanse PEM buffers before freeing them since they may contain
12465 *Benjamin Bennett <ben@psc.edu>*
12467 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12468 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12469 authentication-only ciphersuites.
12473 * Since AES128 and AES256 share a single mask bit in the logic of
12474 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12475 kludge to work properly if AES128 is available and AES256 isn't.
12479 * Expand security boundary to match 1.1.1 module.
12483 * Remove redundant features: hash file source, editing of test vectors
12484 modify fipsld to use external fips_premain.c signature.
12488 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12489 run algorithm test programs.
12493 * Make algorithm test programs more tolerant of whitespace.
12497 * Have SSL/TLS server implementation tolerate "mismatched" record
12498 protocol version while receiving ClientHello even if the
12499 ClientHello is fragmented. (The server can't insist on the
12500 particular protocol version it has chosen before the ServerHello
12501 message has informed the client about his choice.)
12505 * Load error codes if they are not already present instead of using a
12506 static variable. This allows them to be cleanly unloaded and reloaded.
12510 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12512 * Introduce limits to prevent malicious keys being able to
12513 cause a denial of service. ([CVE-2006-2940])
12515 *Steve Henson, Bodo Moeller*
12517 * Fix ASN.1 parsing of certain invalid structures that can result
12518 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12520 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12521 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12523 * Fix SSL client code which could crash if connecting to a
12524 malicious SSLv2 server. ([CVE-2006-4343])
12526 *Tavis Ormandy and Will Drewry, Google Security Team*
12528 * Change ciphersuite string processing so that an explicit
12529 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12530 will no longer include "AES128-SHA"), and any other similar
12531 ciphersuite (same bitmap) from *other* protocol versions (so that
12532 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12533 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12534 changes from 0.9.8b and 0.9.8d.
12538 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12540 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12541 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12543 * Change the Unix randomness entropy gathering to use poll() when
12544 possible instead of select(), since the latter has some
12545 undesirable limitations.
12547 *Darryl Miles via Richard Levitte and Bodo Moeller*
12549 * Disable rogue ciphersuites:
12551 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12552 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12553 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12555 The latter two were purportedly from
12556 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12559 Also deactivate the remaining ciphersuites from
12560 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12561 unofficial, and the ID has long expired.
12565 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12566 dual-core machines) and other potential thread-safety issues.
12570 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12572 * Adapt fipsld and the build system to link against the validated FIPS
12573 module in FIPS mode.
12577 * Fixes for VC++ 2005 build under Windows.
12581 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12582 from a Windows bash shell such as MSYS. It is autodetected from the
12583 "config" script when run from a VC++ environment. Modify standard VC++
12584 build to use fipscanister.o from the GNU make build.
12588 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12590 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12591 The value now differs depending on if you build for FIPS or not.
12592 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12593 safely run with a non-FIPSed libcrypto, as it may crash because of
12594 the difference induced by this change.
12598 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12600 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12601 (part of SSL_OP_ALL). This option used to disable the
12602 countermeasure against man-in-the-middle protocol-version
12603 rollback in the SSL 2.0 server implementation, which is a bad
12604 idea. ([CVE-2005-2969])
12606 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12607 for Information Security, National Institute of Advanced Industrial
12608 Science and Technology [AIST, Japan)]*
12610 * Minimal support for X9.31 signatures and PSS padding modes. This is
12611 mainly for FIPS compliance and not fully integrated at this stage.
12615 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12616 the exponentiation using a fixed-length exponent. (Otherwise,
12617 the information leaked through timing could expose the secret key
12618 after many signatures; cf. Bleichenbacher's attack on DSA with
12623 * Make a new fixed-window mod_exp implementation the default for
12624 RSA, DSA, and DH private-key operations so that the sequence of
12625 squares and multiplies and the memory access pattern are
12626 independent of the particular secret key. This will mitigate
12627 cache-timing and potential related attacks.
12629 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12630 and this is automatically used by BN_mod_exp_mont() if the new flag
12631 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12632 will use this BN flag for private exponents unless the flag
12633 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12634 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12636 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12638 * Change the client implementation for SSLv23_method() and
12639 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12640 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12641 (Previously, the SSL 2.0 backwards compatible Client Hello
12642 message format would be used even with SSL_OP_NO_SSLv2.)
12646 * Add support for smime-type MIME parameter in S/MIME messages which some
12651 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12652 a threadsafe manner. Modify rsa code to use new function and add calls
12653 to dsa and dh code (which had race conditions before).
12657 * Include the fixed error library code in the C error file definitions
12658 instead of fixing them up at runtime. This keeps the error code
12659 structures constant.
12663 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12665 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12668 * Fixes for newer kerberos headers. NB: the casts are needed because
12669 the 'length' field is signed on one version and unsigned on another
12670 with no (?) obvious way to tell the difference, without these VC++
12671 complains. Also the "definition" of FAR (blank) is no longer included
12672 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12673 some needed definitions.
12677 * Undo Cygwin change.
12681 * Added support for proxy certificates according to RFC 3820.
12682 Because they may be a security thread to unaware applications,
12683 they must be explicitly allowed in run-time. See
12684 docs/HOWTO/proxy_certificates.txt for further information.
12688 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12690 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12691 server and client random values. Previously
12692 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12693 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12695 This change has negligible security impact because:
12697 1. Server and client random values still have 24 bytes of pseudo random
12700 2. Server and client random values are sent in the clear in the initial
12703 3. The master secret is derived using the premaster secret (48 bytes in
12704 size for static RSA ciphersuites) as well as client server and random
12707 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12710 *Stephen Henson, reported by UK NISCC*
12712 * Use Windows randomness collection on Cygwin.
12716 * Fix hang in EGD/PRNGD query when communication socket is closed
12717 prematurely by EGD/PRNGD.
12719 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12721 * Prompt for pass phrases when appropriate for PKCS12 input format.
12725 * Back-port of selected performance improvements from development
12726 branch, as well as improved support for PowerPC platforms.
12730 * Add lots of checks for memory allocation failure, error codes to indicate
12731 failure and freeing up memory if a failure occurs.
12733 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12735 * Add new -passin argument to dgst.
12739 * Perform some character comparisons of different types in X509_NAME_cmp:
12740 this is needed for some certificates that re-encode DNs into UTF8Strings
12741 (in violation of RFC3280) and can't or won't issue name rollover
12746 * Make an explicit check during certificate validation to see that
12747 the CA setting in each certificate on the chain is correct. As a
12748 side effect always do the following basic checks on extensions,
12749 not just when there's an associated purpose to the check:
12751 - if there is an unhandled critical extension (unless the user
12752 has chosen to ignore this fault)
12753 - if the path length has been exceeded (if one is set at all)
12754 - that certain extensions fit the associated purpose (if one has
12759 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12761 * Avoid a race condition when CRLs are checked in a multi threaded
12762 environment. This would happen due to the reordering of the revoked
12763 entries during signature checking and serial number lookup. Now the
12764 encoding is cached and the serial number sort performed under a lock.
12765 Add new STACK function sk_is_sorted().
12769 * Add Delta CRL to the extension code.
12773 * Various fixes to s3_pkt.c so alerts are sent properly.
12775 *David Holmes <d.holmes@f5.com>*
12777 * Reduce the chances of duplicate issuer name and serial numbers (in
12778 violation of RFC3280) using the OpenSSL certificate creation utilities.
12779 This is done by creating a random 64 bit value for the initial serial
12780 number when a serial number file is created or when a self signed
12781 certificate is created using 'openssl req -x509'. The initial serial
12782 number file is created using 'openssl x509 -next_serial' in CA.pl
12783 rather than being initialized to 1.
12787 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12789 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12790 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12792 *Joe Orton, Steve Henson*
12794 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12797 *Joe Orton, Steve Henson*
12799 * Make it possible to have multiple active certificates with the same
12800 subject in the CA index file. This is done only if the keyword
12801 'unique_subject' is set to 'no' in the main CA section (default
12802 if 'CA_default') of the configuration file. The value is saved
12803 with the database itself in a separate index attribute file,
12804 named like the index file with '.attr' appended to the name.
12808 * X509 verify fixes. Disable broken certificate workarounds when
12809 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12810 keyUsage extension present. Don't accept CRLs with unhandled critical
12811 extensions: since verify currently doesn't process CRL extensions this
12812 rejects a CRL with *any* critical extensions. Add new verify error codes
12817 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12818 A clarification of RFC2560 will require the use of OCTET STRINGs and
12819 some implementations cannot handle the current raw format. Since OpenSSL
12820 copies and compares OCSP nonces as opaque blobs without any attempt at
12821 parsing them this should not create any compatibility issues.
12825 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12826 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12827 this HMAC (and other) operations are several times slower than OpenSSL
12832 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12834 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12836 * Use the correct content when signing type "other".
12840 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12842 * Fix various bugs revealed by running the NISCC test suite:
12844 Stop out of bounds reads in the ASN1 code when presented with
12845 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12847 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12849 If verify callback ignores invalid public key errors don't try to check
12850 certificate signature with the NULL public key.
12854 * New -ignore_err option in ocsp application to stop the server
12855 exiting on the first error in a request.
12859 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12860 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12865 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12866 extra data after the compression methods not only for TLS 1.0
12867 but also for SSL 3.0 (as required by the specification).
12869 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12871 * Change X509_certificate_type() to mark the key as exported/exportable
12872 when it's 512 *bits* long, not 512 bytes.
12876 * Change AES_cbc_encrypt() so it outputs exact multiple of
12877 blocks during encryption.
12881 * Various fixes to base64 BIO and non blocking I/O. On write
12882 flushes were not handled properly if the BIO retried. On read
12883 data was not being buffered properly and had various logic bugs.
12884 This also affects blocking I/O when the data being decoded is a
12889 * Various S/MIME bugfixes and compatibility changes:
12890 output correct application/pkcs7 MIME type if
12891 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12892 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12893 of files as .eml work). Correctly handle very long lines in MIME
12898 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12900 * Countermeasure against the Klima-Pokorny-Rosa extension of
12901 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12902 a protocol version number mismatch like a decryption error
12903 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12907 * Turn on RSA blinding by default in the default implementation
12908 to avoid a timing attack. Applications that don't want it can call
12909 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12910 They would be ill-advised to do so in most cases.
12912 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12914 * Change RSA blinding code so that it works when the PRNG is not
12915 seeded (in this case, the secret RSA exponent is abused as
12916 an unpredictable seed -- if it is not unpredictable, there
12917 is no point in blinding anyway). Make RSA blinding thread-safe
12918 by remembering the creator's thread ID in rsa->blinding and
12919 having all other threads use local one-time blinding factors
12920 (this requires more computation than sharing rsa->blinding, but
12921 avoids excessive locking; and if an RSA object is not shared
12922 between threads, blinding will still be very fast).
12926 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12927 ENGINE as defaults for all supported algorithms irrespective of
12928 the 'flags' parameter. 'flags' is now honoured, so applications
12929 should make sure they are passing it correctly.
12933 * Target "mingw" now allows native Windows code to be generated in
12934 the Cygwin environment as well as with the MinGW compiler.
12938 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12940 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12941 via timing by performing a MAC computation even if incorrect
12942 block cipher padding has been found. This is a countermeasure
12943 against active attacks where the attacker has to distinguish
12944 between bad padding and a MAC verification error. ([CVE-2003-0078])
12946 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12947 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12948 Martin Vuagnoux (EPFL, Ilion)*
12950 * Make the no-err option work as intended. The intention with no-err
12951 is not to have the whole error stack handling routines removed from
12952 libcrypto, it's only intended to remove all the function name and
12953 reason texts, thereby removing some of the footprint that may not
12954 be interesting if those errors aren't displayed anyway.
12956 NOTE: it's still possible for any application or module to have its
12957 own set of error texts inserted. The routines are there, just not
12958 used by default when no-err is given.
12962 * Add support for FreeBSD on IA64.
12964 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12966 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12967 Kerberos function mit_des_cbc_cksum(). Before this change,
12968 the value returned by DES_cbc_cksum() was like the one from
12969 mit_des_cbc_cksum(), except the bytes were swapped.
12971 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12973 * Allow an application to disable the automatic SSL chain building.
12974 Before this a rather primitive chain build was always performed in
12975 ssl3_output_cert_chain(): an application had no way to send the
12976 correct chain if the automatic operation produced an incorrect result.
12978 Now the chain builder is disabled if either:
12980 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12982 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12984 The reasoning behind this is that an application would not want the
12985 auto chain building to take place if extra chain certificates are
12986 present and it might also want a means of sending no additional
12987 certificates (for example the chain has two certificates and the
12992 * Add the possibility to build without the ENGINE framework.
12994 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12996 * Under Win32 gmtime() can return NULL: check return value in
12997 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13001 * DSA routines: under certain error conditions uninitialized BN objects
13002 could be freed. Solution: make sure initialization is performed early
13003 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13004 Nils Larsch <nla@trustcenter.de> via PR#459)
13008 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13009 checked on reconnect on the client side, therefore session resumption
13010 could still fail with a "ssl session id is different" error. This
13011 behaviour is masked when SSL_OP_ALL is used due to
13012 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13013 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13014 followup to PR #377.
13018 * IA-32 assembler support enhancements: unified ELF targets, support
13019 for SCO/Caldera platforms, fix for Cygwin shared build.
13023 * Add support for FreeBSD on sparc64. As a consequence, support for
13024 FreeBSD on non-x86 processors is separate from x86 processors on
13025 the config script, much like the NetBSD support.
13027 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13029 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13031 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13034 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13035 code (06) was taken as the first octet of the session ID and the last
13036 octet was ignored consequently. As a result SSLv2 client side session
13037 caching could not have worked due to the session ID mismatch between
13039 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13044 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13045 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13050 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13051 seems that in spite of existing for more than a year, many application
13052 author have done nothing to provide the necessary callbacks, which
13053 means that this particular engine will not work properly anywhere.
13054 This is a very unfortunate situation which forces us, in the name
13055 of usability, to give the hw_ncipher.c a static lock, which is part
13057 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13058 appear in 0.9.8 or later. We EXPECT application authors to have
13059 dealt properly with this when 0.9.8 is released (unless we actually
13060 make such changes in the libcrypto locking code that changes will
13061 have to be made anyway).
13065 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13066 octets have been read, EOF or an error occurs. Without this change
13067 some truncated ASN1 structures will not produce an error.
13071 * Disable Heimdal support, since it hasn't been fully implemented.
13072 Still give the possibility to force the use of Heimdal, but with
13073 warnings and a request that patches get sent to openssl-dev.
13077 * Add the VC-CE target, introduce the WINCE sysname, and add
13078 INSTALL.WCE and appropriate conditionals to make it build.
13080 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13082 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13083 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13084 edit numbers of the version.
13086 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13088 * Introduce safe string copy and catenation functions
13089 (BUF_strlcpy() and BUF_strlcat()).
13091 *Ben Laurie (CHATS) and Richard Levitte*
13093 * Avoid using fixed-size buffers for one-line DNs.
13095 *Ben Laurie (CHATS)*
13097 * Add BUF_MEM_grow_clean() to avoid information leakage when
13098 resizing buffers containing secrets, and use where appropriate.
13100 *Ben Laurie (CHATS)*
13102 * Avoid using fixed size buffers for configuration file location.
13104 *Ben Laurie (CHATS)*
13106 * Avoid filename truncation for various CA files.
13108 *Ben Laurie (CHATS)*
13110 * Use sizeof in preference to magic numbers.
13112 *Ben Laurie (CHATS)*
13114 * Avoid filename truncation in cert requests.
13116 *Ben Laurie (CHATS)*
13118 * Add assertions to check for (supposedly impossible) buffer
13121 *Ben Laurie (CHATS)*
13123 * Don't cache truncated DNS entries in the local cache (this could
13124 potentially lead to a spoofing attack).
13126 *Ben Laurie (CHATS)*
13128 * Fix various buffers to be large enough for hex/decimal
13129 representations in a platform independent manner.
13131 *Ben Laurie (CHATS)*
13133 * Add CRYPTO_realloc_clean() to avoid information leakage when
13134 resizing buffers containing secrets, and use where appropriate.
13136 *Ben Laurie (CHATS)*
13138 * Add BIO_indent() to avoid much slightly worrying code to do
13141 *Ben Laurie (CHATS)*
13143 * Convert sprintf()/BIO_puts() to BIO_printf().
13145 *Ben Laurie (CHATS)*
13147 * buffer_gets() could terminate with the buffer only half
13150 *Ben Laurie (CHATS)*
13152 * Add assertions to prevent user-supplied crypto functions from
13153 overflowing internal buffers by having large block sizes, etc.
13155 *Ben Laurie (CHATS)*
13157 * New OPENSSL_assert() macro (similar to assert(), but enabled
13160 *Ben Laurie (CHATS)*
13162 * Eliminate unused copy of key in RC4.
13164 *Ben Laurie (CHATS)*
13166 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13168 *Ben Laurie (CHATS)*
13170 * Fix off-by-one error in EGD path.
13172 *Ben Laurie (CHATS)*
13174 * If RANDFILE path is too long, ignore instead of truncating.
13176 *Ben Laurie (CHATS)*
13178 * Eliminate unused and incorrectly sized X.509 structure
13181 *Ben Laurie (CHATS)*
13183 * Eliminate unused and dangerous function knumber().
13185 *Ben Laurie (CHATS)*
13187 * Eliminate unused and dangerous structure, KSSL_ERR.
13189 *Ben Laurie (CHATS)*
13191 * Protect against overlong session ID context length in an encoded
13192 session object. Since these are local, this does not appear to be
13195 *Ben Laurie (CHATS)*
13197 * Change from security patch (see 0.9.6e below) that did not affect
13198 the 0.9.6 release series:
13200 Remote buffer overflow in SSL3 protocol - an attacker could
13201 supply an oversized master key in Kerberos-enabled versions.
13204 *Ben Laurie (CHATS)*
13206 * Change the SSL kerb5 codes to match RFC 2712.
13210 * Make -nameopt work fully for req and add -reqopt switch.
13212 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13214 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13216 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13218 * Make sure tests can be performed even if the corresponding algorithms
13219 have been removed entirely. This was also the last step to make
13220 OpenSSL compilable with DJGPP under all reasonable conditions.
13222 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13224 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13225 to allow version independent disabling of normally unselected ciphers,
13226 which may be activated as a side-effect of selecting a single cipher.
13228 (E.g., cipher list string "RSA" enables ciphersuites that are left
13229 out of "ALL" because they do not provide symmetric encryption.
13230 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13232 *Lutz Jaenicke, Bodo Moeller*
13234 * Add appropriate support for separate platform-dependent build
13235 directories. The recommended way to make a platform-dependent
13236 build directory is the following (tested on Linux), maybe with
13239 # Place yourself outside of the OpenSSL source tree. In
13240 # this example, the environment variable OPENSSL_SOURCE
13241 # is assumed to contain the absolute OpenSSL source directory.
13242 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13243 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13244 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13245 mkdir -p `dirname $F`
13246 ln -s $OPENSSL_SOURCE/$F $F
13249 To be absolutely sure not to disturb the source tree, a "make clean"
13250 is a good thing. If it isn't successful, don't worry about it,
13251 it probably means the source directory is very clean.
13255 * Make sure any ENGINE control commands make local copies of string
13256 pointers passed to them whenever necessary. Otherwise it is possible
13257 the caller may have overwritten (or deallocated) the original string
13258 data when a later ENGINE operation tries to use the stored values.
13260 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13262 * Improve diagnostics in file reading and command-line digests.
13264 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13266 * Add AES modes CFB and OFB to the object database. Correct an
13267 error in AES-CFB decryption.
13271 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13272 allows existing EVP_CIPHER_CTX structures to be reused after
13273 calling `EVP_*Final()`. This behaviour is used by encryption
13274 BIOs and some applications. This has the side effect that
13275 applications must explicitly clean up cipher contexts with
13276 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13280 * Check the values of dna and dnb in bn_mul_recursive before calling
13281 bn_mul_comba (a non zero value means the a or b arrays do not contain
13282 n2 elements) and fallback to bn_mul_normal if either is not zero.
13286 * Fix escaping of non-ASCII characters when using the -subj option
13287 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13291 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13292 form for "surname", serialNumber has no short form.
13293 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13294 therefore remove "mail" short name for "internet 7".
13295 The OID for unique identifiers in X509 certificates is
13296 x500UniqueIdentifier, not uniqueIdentifier.
13297 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13301 * Add an "init" command to the ENGINE config module and auto initialize
13302 ENGINEs. Without any "init" command the ENGINE will be initialized
13303 after all ctrl commands have been executed on it. If init=1 the
13304 ENGINE is initialized at that point (ctrls before that point are run
13305 on the uninitialized ENGINE and after on the initialized one). If
13306 init=0 then the ENGINE will not be initialized at all.
13310 * Fix the 'app_verify_callback' interface so that the user-defined
13311 argument is actually passed to the callback: In the
13312 SSL_CTX_set_cert_verify_callback() prototype, the callback
13313 declaration has been changed from
13316 int (*cb)(X509_STORE_CTX *,void *);
13317 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13318 i=s->ctx->app_verify_callback(&ctx)
13319 has been changed into
13320 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13322 To update applications using SSL_CTX_set_cert_verify_callback(),
13323 a dummy argument can be added to their callback functions.
13325 *D. K. Smetters <smetters@parc.xerox.com>*
13327 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13329 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13331 * Add and OPENSSL_LOAD_CONF define which will cause
13332 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13333 This allows older applications to transparently support certain
13334 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13335 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13336 load the config file and OPENSSL_add_all_algorithms_conf() which will
13337 always load it have also been added.
13341 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13342 Adjust NIDs and EVP layer.
13344 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13346 * Config modules support in openssl utility.
13348 Most commands now load modules from the config file,
13349 though in a few (such as version) this isn't done
13350 because it couldn't be used for anything.
13352 In the case of ca and req the config file used is
13353 the same as the utility itself: that is the -config
13354 command line option can be used to specify an
13359 * Move default behaviour from OPENSSL_config(). If appname is NULL
13360 use "openssl_conf" if filename is NULL use default openssl config file.
13364 * Add an argument to OPENSSL_config() to allow the use of an alternative
13365 config section name. Add a new flag to tolerate a missing config file
13366 and move code to CONF_modules_load_file().
13370 * Support for crypto accelerator cards from Accelerated Encryption
13371 Processing, www.aep.ie. (Use engine 'aep')
13372 The support was copied from 0.9.6c [engine] and adapted/corrected
13373 to work with the new engine framework.
13375 *AEP Inc. and Richard Levitte*
13377 * Support for SureWare crypto accelerator cards from Baltimore
13378 Technologies. (Use engine 'sureware')
13379 The support was copied from 0.9.6c [engine] and adapted
13380 to work with the new engine framework.
13384 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13385 make the newer ENGINE framework commands for the CHIL engine work.
13387 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13389 * Make it possible to produce shared libraries on ReliantUNIX.
13391 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13393 * Add the configuration target debug-linux-ppro.
13394 Make 'openssl rsa' use the general key loading routines
13395 implemented in `apps.c`, and make those routines able to
13396 handle the key format FORMAT_NETSCAPE and the variant
13399 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13401 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13403 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13405 * Add -keyform to rsautl, and document -engine.
13407 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13409 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13410 BIO_R_NO_SUCH_FILE error code rather than the generic
13411 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13415 * Add new functions
13416 ERR_peek_last_error
13417 ERR_peek_last_error_line
13418 ERR_peek_last_error_line_data.
13419 These are similar to
13421 ERR_peek_error_line
13422 ERR_peek_error_line_data,
13423 but report on the latest error recorded rather than the first one
13424 still in the error queue.
13426 *Ben Laurie, Bodo Moeller*
13428 * default_algorithms option in ENGINE config module. This allows things
13430 default_algorithms = ALL
13431 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13435 * Preliminary ENGINE config module.
13439 * New experimental application configuration code.
13443 * Change the AES code to follow the same name structure as all other
13444 symmetric ciphers, and behave the same way. Move everything to
13445 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13447 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13449 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13451 *Ben Laurie and Theo de Raadt*
13453 * Add option to output public keys in req command.
13455 *Massimiliano Pala madwolf@openca.org*
13457 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13458 (up to about 10% better than before for P-192 and P-224).
13462 * New functions/macros
13464 SSL_CTX_set_msg_callback(ctx, cb)
13465 SSL_CTX_set_msg_callback_arg(ctx, arg)
13466 SSL_set_msg_callback(ssl, cb)
13467 SSL_set_msg_callback_arg(ssl, arg)
13469 to request calling a callback function
13471 void cb(int write_p, int version, int content_type,
13472 const void *buf, size_t len, SSL *ssl, void *arg)
13474 whenever a protocol message has been completely received
13475 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13476 protocol version according to which the SSL library interprets
13477 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13478 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13479 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13480 specification (change_cipher_spec(20), alert(21), handshake(22)).
13481 'buf' and 'len' point to the actual message, 'ssl' to the
13482 SSL object, and 'arg' is the application-defined value set by
13483 SSL[_CTX]_set_msg_callback_arg().
13485 'openssl s_client' and 'openssl s_server' have new '-msg' options
13486 to enable a callback that displays all protocol messages.
13490 * Change the shared library support so shared libraries are built as
13491 soon as the corresponding static library is finished, and thereby get
13492 openssl and the test programs linked against the shared library.
13493 This still only happens when the keyword "shard" has been given to
13494 the configuration scripts.
13496 NOTE: shared library support is still an experimental thing, and
13497 backward binary compatibility is still not guaranteed.
13499 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13501 * Add support for Subject Information Access extension.
13503 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13505 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13506 additional bytes when new memory had to be allocated, not just
13507 when reusing an existing buffer.
13511 * New command line and configuration option 'utf8' for the req command.
13512 This allows field values to be specified as UTF8 strings.
13516 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13517 runs for the former and machine-readable output for the latter.
13521 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13522 of the e-mail address in the DN (i.e., it will go into a certificate
13523 extension only). The new configuration file option 'email_in_dn = no'
13524 has the same effect.
13526 *Massimiliano Pala madwolf@openca.org*
13528 * Change all functions with names starting with `des_` to be starting
13529 with `DES_` instead. Add wrappers that are compatible with libdes,
13530 but are named `_ossl_old_des_*`. Finally, add macros that map the
13531 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13532 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13533 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13536 Since we provide two compatibility mappings, the user needs to
13537 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13538 compatibility is desired. The default (i.e., when that macro
13539 isn't defined) is OpenSSL 0.9.6c compatibility.
13541 There are also macros that enable and disable the support of old
13542 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13543 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13544 are defined, the default will apply: to support the old des routines.
13546 In either case, one must include openssl/des.h to get the correct
13547 definitions. Do not try to just include openssl/des_old.h, that
13550 NOTE: This is a major break of an old API into a new one. Software
13551 authors are encouraged to switch to the `DES_` style functions. Some
13552 time in the future, des_old.h and the libdes compatibility functions
13553 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13554 default), and then completely removed.
13558 * Test for certificates which contain unsupported critical extensions.
13559 If such a certificate is found during a verify operation it is
13560 rejected by default: this behaviour can be overridden by either
13561 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13562 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13563 X509_supported_extension() has also been added which returns 1 if a
13564 particular extension is supported.
13568 * Modify the behaviour of EVP cipher functions in similar way to digests
13569 to retain compatibility with existing code.
13573 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13574 compatibility with existing code. In particular the 'ctx' parameter does
13575 not have to be to be initialized before the call to EVP_DigestInit() and
13576 it is tidied up after a call to EVP_DigestFinal(). New function
13577 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13578 EVP_MD_CTX_copy() changed to not require the destination to be
13579 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13580 requires the destination to be valid.
13582 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13583 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13587 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13588 so that complete 'Handshake' protocol structures are kept in memory
13589 instead of overwriting 'msg_type' and 'length' with 'body' data.
13593 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13595 *Massimo Santin via Richard Levitte*
13597 * Major restructuring to the underlying ENGINE code. This includes
13598 reduction of linker bloat, separation of pure "ENGINE" manipulation
13599 (initialisation, etc) from functionality dealing with implementations
13600 of specific crypto interfaces. This change also introduces integrated
13601 support for symmetric ciphers and digest implementations - so ENGINEs
13602 can now accelerate these by providing EVP_CIPHER and EVP_MD
13603 implementations of their own. This is detailed in
13604 [crypto/engine/README.md](crypto/engine/README.md)
13605 as it couldn't be adequately described here. However, there are a few
13606 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13607 were changed in the original introduction of ENGINE code have now
13608 reverted back - the hooking from this code to ENGINE is now a good
13609 deal more passive and at run-time, operations deal directly with
13610 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13611 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13612 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13613 they were not being used by the framework as there is no concept of a
13614 BIGNUM_METHOD and they could not be generalised to the new
13615 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13616 ENGINE_cpy() has been removed as it cannot be consistently defined in
13621 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13625 * Change mkdef.pl to sort symbols that get the same entry number,
13626 and make sure the automatically generated functions `ERR_load_*`
13627 become part of libeay.num as well.
13631 * New function SSL_renegotiate_pending(). This returns true once
13632 renegotiation has been requested (either SSL_renegotiate() call
13633 or HelloRequest/ClientHello received from the peer) and becomes
13634 false once a handshake has been completed.
13635 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13636 sends a HelloRequest, but does not ensure that a handshake takes
13637 place. SSL_renegotiate_pending() is useful for checking if the
13638 client has followed the request.)
13642 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13643 By default, clients may request session resumption even during
13644 renegotiation (if session ID contexts permit); with this option,
13645 session resumption is possible only in the first handshake.
13647 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13648 more bits available for options that should not be part of
13649 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13653 * Add some demos for certificate and certificate request creation.
13657 * Make maximum certificate chain size accepted from the peer application
13658 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13659 "Douglas E. Engert" <deengert@anl.gov>.
13663 * Add support for shared libraries for Unixware-7
13664 (Boyd Lynn Gerber <gerberb@zenez.com>).
13668 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13669 be done prior to destruction. Use this to unload error strings from
13670 ENGINEs that load their own error strings. NB: This adds two new API
13671 functions to "get" and "set" this destroy handler in an ENGINE.
13675 * Alter all existing ENGINE implementations (except "openssl" and
13676 "openbsd") to dynamically instantiate their own error strings. This
13677 makes them more flexible to be built both as statically-linked ENGINEs
13678 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13679 Also, add stub code to each that makes building them as self-contained
13680 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13684 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13685 implementations into applications that are completely implemented in
13686 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13687 commands that can be used to configure what shared-library to load and
13688 to control aspects of the way it is handled. Also, made an update to
13689 the [README-Engine.md](README-Engine.md) file
13690 that brings its information up-to-date and
13691 provides some information and instructions on the "dynamic" ENGINE
13692 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13696 * Make it possible to unload ranges of ERR strings with a new
13697 "ERR_unload_strings" function.
13701 * Add a copy() function to EVP_MD.
13705 * Make EVP_MD routines take a context pointer instead of just the
13706 md_data void pointer.
13710 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13711 that the digest can only process a single chunk of data
13712 (typically because it is provided by a piece of
13713 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13714 is only going to provide a single chunk of data, and hence the
13715 framework needn't accumulate the data for oneshot drivers.
13719 * As with "ERR", make it possible to replace the underlying "ex_data"
13720 functions. This change also alters the storage and management of global
13721 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13722 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13723 index counters. The API functions that use this state have been changed
13724 to take a "class_index" rather than pointers to the class's local STACK
13725 and counter, and there is now an API function to dynamically create new
13726 classes. This centralisation allows us to (a) plug a lot of the
13727 thread-safety problems that existed, and (b) makes it possible to clean
13728 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13729 such data would previously have always leaked in application code and
13730 workarounds were in place to make the memory debugging turn a blind eye
13731 to it. Application code that doesn't use this new function will still
13732 leak as before, but their memory debugging output will announce it now
13733 rather than letting it slide.
13735 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13736 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13737 has a return value to indicate success or failure.
13741 * Make it possible to replace the underlying "ERR" functions such that the
13742 global state (2 LHASH tables and 2 locks) is only used by the "default"
13743 implementation. This change also adds two functions to "get" and "set"
13744 the implementation prior to it being automatically set the first time
13745 any other ERR function takes place. Ie. an application can call "get",
13746 pass the return value to a module it has just loaded, and that module
13747 can call its own "set" function using that value. This means the
13748 module's "ERR" operations will use (and modify) the error state in the
13749 application and not in its own statically linked copy of OpenSSL code.
13753 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13754 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13755 the operation, and provides a more encapsulated way for external code
13756 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13757 to use these functions rather than manually incrementing the counts.
13759 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13763 * Add EVP test program.
13767 * Add symmetric cipher support to ENGINE. Expect the API to change!
13771 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13772 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13773 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13774 These allow a CRL to be built without having to access X509_CRL fields
13775 directly. Modify 'ca' application to use new functions.
13779 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13780 bug workarounds. Rollback attack detection is a security feature.
13781 The problem will only arise on OpenSSL servers when TLSv1 is not
13782 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13783 Software authors not wanting to support TLSv1 will have special reasons
13784 for their choice and can explicitly enable this option.
13786 *Bodo Moeller, Lutz Jaenicke*
13788 * Rationalise EVP so it can be extended: don't include a union of
13789 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13790 (similar to those existing for EVP_CIPHER_CTX).
13795 EVP_MD_CTX_init(&md); /* new function call */
13796 EVP_DigestInit(&md, EVP_sha1());
13797 EVP_DigestUpdate(&md, in, len);
13798 EVP_DigestFinal(&md, out, NULL);
13799 EVP_MD_CTX_cleanup(&md); /* new function call */
13803 * Make DES key schedule conform to the usual scheme, as well as
13804 correcting its structure. This means that calls to DES functions
13805 now have to pass a pointer to a des_key_schedule instead of a
13806 plain des_key_schedule (which was actually always a pointer
13809 des_key_schedule ks;
13811 des_set_key_checked(..., &ks);
13812 des_ncbc_encrypt(..., &ks, ...);
13814 (Note that a later change renames 'des_...' into 'DES_...'.)
13818 * Initial reduction of linker bloat: the use of some functions, such as
13819 PEM causes large amounts of unused functions to be linked in due to
13820 poor organisation. For example pem_all.c contains every PEM function
13821 which has a knock on effect of linking in large amounts of (unused)
13822 ASN1 code. Grouping together similar functions and splitting unrelated
13823 functions prevents this.
13827 * Cleanup of EVP macros.
13831 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13832 correct `_ecb suffix`.
13836 * Add initial OCSP responder support to ocsp application. The
13837 revocation information is handled using the text based index
13838 use by the ca application. The responder can either handle
13839 requests generated internally, supplied in files (for example
13840 via a CGI script) or using an internal minimal server.
13844 * Add configuration choices to get zlib compression for TLS.
13848 * Changes to Kerberos SSL for RFC 2712 compliance:
13849 1. Implemented real KerberosWrapper, instead of just using
13850 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13851 2. Implemented optional authenticator field of KerberosWrapper.
13853 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13854 and authenticator structs; see crypto/krb5/.
13856 Generalized Kerberos calls to support multiple Kerberos libraries.
13857 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13858 via Richard Levitte*
13860 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13861 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13862 values for each of the key sizes rather than having just
13863 parameters (and 'speed' generating keys each time).
13867 * Speed up EVP routines.
13870 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13871 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13872 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13873 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13875 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13876 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13877 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13880 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13882 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13886 * Added the OS2-EMX target.
13888 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13890 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13891 New functions to support `NCONF` routines in extension code.
13892 New function `CONF_set_nconf()`
13893 to allow functions which take an `NCONF` to also handle the old `LHASH`
13894 structure: this means that the old `CONF` compatible routines can be
13895 retained (in particular w.rt. extensions) without having to duplicate the
13896 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13900 * Enhance the general user interface with mechanisms for inner control
13901 and with possibilities to have yes/no kind of prompts.
13905 * Change all calls to low-level digest routines in the library and
13906 applications to use EVP. Add missing calls to HMAC_cleanup() and
13907 don't assume HMAC_CTX can be copied using memcpy().
13909 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13911 * Add the possibility to control engines through control names but with
13912 arbitrary arguments instead of just a string.
13913 Change the key loaders to take a UI_METHOD instead of a callback
13914 function pointer. NOTE: this breaks binary compatibility with earlier
13915 versions of OpenSSL [engine].
13916 Adapt the nCipher code for these new conditions and add a card insertion
13921 * Enhance the general user interface with mechanisms to better support
13922 dialog box interfaces, application-defined prompts, the possibility
13923 to use defaults (for example default passwords from somewhere else)
13924 and interrupts/cancellations.
13928 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13929 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13933 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13934 tidy up some unnecessarily weird code in 'sk_new()').
13936 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13938 * Change the key loading routines for ENGINEs to use the same kind
13939 callback (pem_password_cb) as all other routines that need this
13944 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13945 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13946 than this minimum value is recommended.
13950 * New random seeder for OpenVMS, using the system process statistics
13951 that are easily reachable.
13955 * Windows apparently can't transparently handle global
13956 variables defined in DLLs. Initialisations such as:
13958 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13960 won't compile. This is used by the any applications that need to
13961 declare their own ASN1 modules. This was fixed by adding the option
13962 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13963 needed for static libraries under Win32.
13967 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13968 setting of purpose and trust fields. New X509_STORE trust and
13969 purpose functions and tidy up setting in other SSL functions.
13973 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13974 structure. These are inherited by X509_STORE_CTX when it is
13975 initialised. This allows various defaults to be set in the
13976 X509_STORE structure (such as flags for CRL checking and custom
13977 purpose or trust settings) for functions which only use X509_STORE_CTX
13978 internally such as S/MIME.
13980 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13981 trust settings if they are not set in X509_STORE. This allows X509_STORE
13982 purposes and trust (in S/MIME for example) to override any set by default.
13984 Add command line options for CRL checking to smime, s_client and s_server
13989 * Initial CRL based revocation checking. If the CRL checking flag(s)
13990 are set then the CRL is looked up in the X509_STORE structure and
13991 its validity and signature checked, then if the certificate is found
13992 in the CRL the verify fails with a revoked error.
13994 Various new CRL related callbacks added to X509_STORE_CTX structure.
13996 Command line options added to 'verify' application to support this.
13998 This needs some additional work, such as being able to handle multiple
13999 CRLs with different times, extension based lookup (rather than just
14000 by subject name) and ultimately more complete V2 CRL extension
14005 * Add a general user interface API (crypto/ui/). This is designed
14006 to replace things like des_read_password and friends (backward
14007 compatibility functions using this new API are provided).
14008 The purpose is to remove prompting functions from the DES code
14009 section as well as provide for prompting through dialog boxes in
14010 a window system and the like.
14014 * Add "ex_data" support to ENGINE so implementations can add state at a
14015 per-structure level rather than having to store it globally.
14019 * Make it possible for ENGINE structures to be copied when retrieved by
14020 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14021 This causes the "original" ENGINE structure to act like a template,
14022 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14023 operational state can be localised to each ENGINE structure, despite the
14024 fact they all share the same "methods". New ENGINE structures returned in
14025 this case have no functional references and the return value is the single
14026 structural reference. This matches the single structural reference returned
14027 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14032 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14033 needs to match any other type at all we need to manually clear the
14038 * Changes to the "openssl engine" utility to include;
14039 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14040 about an ENGINE's available control commands.
14041 - executing control commands from command line arguments using the
14042 '-pre' and '-post' switches. '-post' is only used if '-t' is
14043 specified and the ENGINE is successfully initialised. The syntax for
14044 the individual commands are colon-separated, for example;
14045 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14049 * New dynamic control command support for ENGINEs. ENGINEs can now
14050 declare their own commands (numbers), names (strings), descriptions,
14051 and input types for run-time discovery by calling applications. A
14052 subset of these commands are implicitly classed as "executable"
14053 depending on their input type, and only these can be invoked through
14054 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14055 can be based on user input, config files, etc). The distinction is
14056 that "executable" commands cannot return anything other than a boolean
14057 result and can only support numeric or string input, whereas some
14058 discoverable commands may only be for direct use through
14059 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14060 pointers, or other custom uses. The "executable" commands are to
14061 support parameterisations of ENGINE behaviour that can be
14062 unambiguously defined by ENGINEs and used consistently across any
14063 OpenSSL-based application. Commands have been added to all the
14064 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14065 control over shared-library paths without source code alterations.
14069 * Changed all ENGINE implementations to dynamically allocate their
14070 ENGINEs rather than declaring them statically. Apart from this being
14071 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14072 this also allows the implementations to compile without using the
14073 internal engine_int.h header.
14077 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14078 'const' value. Any code that should be able to modify a RAND_METHOD
14079 should already have non-const pointers to it (ie. they should only
14080 modify their own ones).
14084 * Made a variety of little tweaks to the ENGINE code.
14085 - "atalla" and "ubsec" string definitions were moved from header files
14086 to C code. "nuron" string definitions were placed in variables
14087 rather than hard-coded - allowing parameterisation of these values
14088 later on via ctrl() commands.
14089 - Removed unused "#if 0"'d code.
14090 - Fixed engine list iteration code so it uses ENGINE_free() to release
14091 structural references.
14092 - Constified the RAND_METHOD element of ENGINE structures.
14093 - Constified various get/set functions as appropriate and added
14094 missing functions (including a catch-all ENGINE_cpy that duplicates
14095 all ENGINE values onto a new ENGINE except reference counts/state).
14096 - Removed NULL parameter checks in get/set functions. Setting a method
14097 or function to NULL is a way of cancelling out a previously set
14098 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14099 and doesn't justify the extra error symbols and code.
14100 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14101 flags from engine_int.h to engine.h.
14102 - Changed prototypes for ENGINE handler functions (init(), finish(),
14103 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14107 * Implement binary inversion algorithm for BN_mod_inverse in addition
14108 to the algorithm using long division. The binary algorithm can be
14109 used only if the modulus is odd. On 32-bit systems, it is faster
14110 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14111 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14112 up to 450 bits. In 64-bit environments, the binary algorithm
14113 appears to be advantageous for much longer moduli; here we use it
14114 for moduli up to 2048 bits.
14118 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14119 could not support the combine flag in choice fields.
14123 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14124 extensions from a certificate request to the certificate.
14128 * Allow multiple 'certopt' and 'nameopt' options to be separated
14129 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14130 file: this allows the display of the certificate about to be
14131 signed to be customised, to allow certain fields to be included
14132 or excluded and extension details. The old system didn't display
14133 multicharacter strings properly, omitted fields not in the policy
14134 and couldn't display additional details such as extensions.
14138 * Function EC_POINTs_mul for multiple scalar multiplication
14139 of an arbitrary number of elliptic curve points
14140 \sum scalars[i]*points[i],
14141 optionally including the generator defined for the EC_GROUP:
14142 scalar*generator + \sum scalars[i]*points[i].
14144 EC_POINT_mul is a simple wrapper function for the typical case
14145 that the point list has just one item (besides the optional
14150 * First EC_METHODs for curves over GF(p):
14152 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14153 operations and provides various method functions that can also
14154 operate with faster implementations of modular arithmetic.
14156 EC_GFp_mont_method() reuses most functions that are part of
14157 EC_GFp_simple_method, but uses Montgomery arithmetic.
14159 *Bodo Moeller; point addition and point doubling
14160 implementation directly derived from source code provided by
14161 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14163 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14164 crypto/ec/ec_lib.c):
14166 Curves are EC_GROUP objects (with an optional group generator)
14167 based on EC_METHODs that are built into the library.
14169 Points are EC_POINT objects based on EC_GROUP objects.
14171 Most of the framework would be able to handle curves over arbitrary
14172 finite fields, but as there are no obvious types for fields other
14173 than GF(p), some functions are limited to that for now.
14177 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14178 that the file contains a complete HTTP response.
14182 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14183 change the def and num file printf format specifier from "%-40sXXX"
14184 to "%-39s XXX". The latter will always guarantee a space after the
14185 field while the former will cause them to run together if the field
14186 is 40 of more characters long.
14190 * Constify the cipher and digest 'method' functions and structures
14191 and modify related functions to take constant EVP_MD and EVP_CIPHER
14196 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14197 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14201 * Modify `EVP_Digest*()` routines so they now return values. Although the
14202 internal software routines can never fail additional hardware versions
14207 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14209 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14210 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14213 ERR_R_NESTED_ASN1_ERROR
14215 ERR_R_MISSING_ASN1_EOS
14216 were 4 .. 9, conflicting with
14217 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14219 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14220 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14222 Add new error code 'ERR_R_INTERNAL_ERROR'.
14226 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14231 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14232 sets the subject name for a new request or supersedes the
14233 subject name in a given request. Formats that can be parsed are
14234 'CN=Some Name, OU=myOU, C=IT'
14236 'CN=Some Name/OU=myOU/C=IT'.
14238 Add options '-batch' and '-verbose' to 'openssl req'.
14240 *Massimiliano Pala <madwolf@hackmasters.net>*
14242 * Introduce the possibility to access global variables through
14243 functions on platform were that's the best way to handle exporting
14244 global variables in shared libraries. To enable this functionality,
14245 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14246 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14247 is normally done by Configure or something similar).
14249 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14250 in the source file (foo.c) like this:
14252 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14253 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14255 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14256 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14258 OPENSSL_DECLARE_GLOBAL(int,foo);
14259 #define foo OPENSSL_GLOBAL_REF(foo)
14260 OPENSSL_DECLARE_GLOBAL(double,bar);
14261 #define bar OPENSSL_GLOBAL_REF(bar)
14263 The #defines are very important, and therefore so is including the
14264 header file everywhere where the defined globals are used.
14266 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14267 of ASN.1 items, but that structure is a bit different.
14269 The largest change is in util/mkdef.pl which has been enhanced with
14270 better and easier to understand logic to choose which symbols should
14271 go into the Windows .def files as well as a number of fixes and code
14272 cleanup (among others, algorithm keywords are now sorted
14273 lexicographically to avoid constant rewrites).
14277 * In BN_div() keep a copy of the sign of 'num' before writing the
14278 result to 'rm' because if rm==num the value will be overwritten
14279 and produce the wrong result if 'num' is negative: this caused
14280 problems with BN_mod() and BN_nnmod().
14284 * Function OCSP_request_verify(). This checks the signature on an
14285 OCSP request and verifies the signer certificate. The signer
14286 certificate is just checked for a generic purpose and OCSP request
14291 * Add OCSP_check_validity() function to check the validity of OCSP
14292 responses. OCSP responses are prepared in real time and may only
14293 be a few seconds old. Simply checking that the current time lies
14294 between thisUpdate and nextUpdate max reject otherwise valid responses
14295 caused by either OCSP responder or client clock inaccuracy. Instead
14296 we allow thisUpdate and nextUpdate to fall within a certain period of
14297 the current time. The age of the response can also optionally be
14298 checked. Two new options -validity_period and -status_age added to
14303 * If signature or public key algorithm is unrecognized print out its
14304 OID rather that just UNKNOWN.
14308 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14309 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14310 ID to be generated from the issuer certificate alone which can then be
14311 passed to OCSP_id_issuer_cmp().
14315 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14316 ASN1 modules to export functions returning ASN1_ITEM pointers
14317 instead of the ASN1_ITEM structures themselves. This adds several
14318 new macros which allow the underlying ASN1 function/structure to
14319 be accessed transparently. As a result code should not use ASN1_ITEM
14320 references directly (such as &X509_it) but instead use the relevant
14321 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14322 use of the new ASN1 code on platforms where exporting structures
14323 is problematical (for example in shared libraries) but exporting
14324 functions returning pointers to structures is not.
14328 * Add support for overriding the generation of SSL/TLS session IDs.
14329 These callbacks can be registered either in an SSL_CTX or per SSL.
14330 The purpose of this is to allow applications to control, if they wish,
14331 the arbitrary values chosen for use as session IDs, particularly as it
14332 can be useful for session caching in multiple-server environments. A
14333 command-line switch for testing this (and any client code that wishes
14334 to use such a feature) has been added to "s_server".
14336 *Geoff Thorpe, Lutz Jaenicke*
14338 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14339 of the form `#if defined(...) || defined(...) || ...` and
14340 `#if !defined(...) && !defined(...) && ...`. This also avoids
14341 the growing number of special cases it was previously handling.
14345 * Make all configuration macros available for application by making
14346 sure they are available in opensslconf.h, by giving them names starting
14347 with `OPENSSL_` to avoid conflicts with other packages and by making
14348 sure e_os2.h will cover all platform-specific cases together with
14350 Additionally, it is now possible to define configuration/platform-
14351 specific names (called "system identities"). In the C code, these
14352 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14353 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14354 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14359 * New option -set_serial to 'req' and 'x509' this allows the serial
14360 number to use to be specified on the command line. Previously self
14361 signed certificates were hard coded with serial number 0 and the
14362 CA options of 'x509' had to use a serial number in a file which was
14367 * New options to 'ca' utility to support V2 CRL entry extensions.
14368 Currently CRL reason, invalidity date and hold instruction are
14369 supported. Add new CRL extensions to V3 code and some new objects.
14373 * New function EVP_CIPHER_CTX_set_padding() this is used to
14374 disable standard block padding (aka PKCS#5 padding) in the EVP
14375 API, which was previously mandatory. This means that the data is
14376 not padded in any way and so the total length much be a multiple
14377 of the block size, otherwise an error occurs.
14381 * Initial (incomplete) OCSP SSL support.
14385 * New function OCSP_parse_url(). This splits up a URL into its host,
14386 port and path components: primarily to parse OCSP URLs. New -url
14387 option to ocsp utility.
14391 * New nonce behavior. The return value of OCSP_check_nonce() now
14392 reflects the various checks performed. Applications can decide
14393 whether to tolerate certain situations such as an absent nonce
14394 in a response when one was present in a request: the ocsp application
14395 just prints out a warning. New function OCSP_add1_basic_nonce()
14396 this is to allow responders to include a nonce in a response even if
14397 the request is nonce-less.
14401 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14402 skipped when using openssl x509 multiple times on a single input file,
14403 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14407 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14408 set string type: to handle setting ASN1_TIME structures. Fix ca
14409 utility to correctly initialize revocation date of CRLs.
14413 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14414 the clients preferred ciphersuites and rather use its own preferences.
14415 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14416 Internet Explorer by ensuring unchanged hash method during stepup.
14417 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14421 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14422 to aes and add a new 'exist' option to print out symbols that don't
14427 * Additional options to ocsp utility to allow flags to be set and
14428 additional certificates supplied.
14432 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14433 OCSP client a number of certificate to only verify the response
14438 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14439 handle the new API. Currently only ECB, CBC modes supported. Add new
14442 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14443 Encryption Standard (AES) Ciphersuites for Transport Layer
14444 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14445 not enabled by default and were not part of the "ALL" ciphersuite
14446 alias because they were not yet official; they could be
14447 explicitly requested by specifying the "AESdraft" ciphersuite
14448 group alias. In the final release of OpenSSL 0.9.7, the group
14449 alias is called "AES" and is part of "ALL".)
14451 *Ben Laurie, Steve Henson, Bodo Moeller*
14453 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14454 request to response.
14458 * Functions for OCSP responders. OCSP_request_onereq_count(),
14459 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14460 extract information from a certificate request. OCSP_response_create()
14461 creates a response and optionally adds a basic response structure.
14462 OCSP_basic_add1_status() adds a complete single response to a basic
14463 response and returns the OCSP_SINGLERESP structure just added (to allow
14464 extensions to be included for example). OCSP_basic_add1_cert() adds a
14465 certificate to a basic response and OCSP_basic_sign() signs a basic
14466 response with various flags. New helper functions ASN1_TIME_check()
14467 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14468 (converts ASN1_TIME to GeneralizedTime).
14472 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14473 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14474 structure from a certificate. X509_pubkey_digest() digests the public_key
14475 contents: this is used in various key identifiers.
14479 * Make sk_sort() tolerate a NULL argument.
14481 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14483 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14484 passed by the function are trusted implicitly. If any of them signed the
14485 response then it is assumed to be valid and is not verified.
14489 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14490 to data. This was previously part of the PKCS7 ASN1 code. This
14491 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14492 *Steve Henson, reported by Kenneth R. Robinette
14493 <support@securenetterm.com>*
14495 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14496 routines: without these tracing memory leaks is very painful.
14497 Fix leaks in PKCS12 and PKCS7 routines.
14501 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14502 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14503 effectively meant GeneralizedTime would never be used. Now it
14504 is initialised to -1 but X509_time_adj() now has to check the value
14505 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14506 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14507 *Steve Henson, reported by Kenneth R. Robinette
14508 <support@securenetterm.com>*
14510 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14511 result in a zero length in the ASN1_INTEGER structure which was
14512 not consistent with the structure when d2i_ASN1_INTEGER() was used
14513 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14514 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14515 where it did not print out a minus for negative ASN1_INTEGER.
14519 * Add summary printout to ocsp utility. The various functions which
14520 convert status values to strings have been renamed to:
14521 OCSP_response_status_str(), OCSP_cert_status_str() and
14522 OCSP_crl_reason_str() and are no longer static. New options
14523 to verify nonce values and to disable verification. OCSP response
14524 printout format cleaned up.
14528 * Add additional OCSP certificate checks. These are those specified
14529 in RFC2560. This consists of two separate checks: the CA of the
14530 certificate being checked must either be the OCSP signer certificate
14531 or the issuer of the OCSP signer certificate. In the latter case the
14532 OCSP signer certificate must contain the OCSP signing extended key
14533 usage. This check is performed by attempting to match the OCSP
14534 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14535 in the OCSP_CERTID structures of the response.
14539 * Initial OCSP certificate verification added to OCSP_basic_verify()
14540 and related routines. This uses the standard OpenSSL certificate
14541 verify routines to perform initial checks (just CA validity) and
14542 to obtain the certificate chain. Then additional checks will be
14543 performed on the chain. Currently the root CA is checked to see
14544 if it is explicitly trusted for OCSP signing. This is used to set
14545 a root CA as a global signing root: that is any certificate that
14546 chains to that CA is an acceptable OCSP signing certificate.
14550 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14551 extensions from a separate configuration file.
14552 As when reading extensions from the main configuration file,
14553 the '-extensions ...' option may be used for specifying the
14556 *Massimiliano Pala <madwolf@comune.modena.it>*
14558 * New OCSP utility. Allows OCSP requests to be generated or
14559 read. The request can be sent to a responder and the output
14560 parsed, outputted or printed in text form. Not complete yet:
14561 still needs to check the OCSP response validity.
14565 * New subcommands for 'openssl ca':
14566 `openssl ca -status <serial>` prints the status of the cert with
14567 the given serial number (according to the index file).
14568 `openssl ca -updatedb` updates the expiry status of certificates
14571 *Massimiliano Pala <madwolf@comune.modena.it>*
14573 * New '-newreq-nodes' command option to CA.pl. This is like
14574 '-newreq', but calls 'openssl req' with the '-nodes' option
14575 so that the resulting key is not encrypted.
14577 *Damien Miller <djm@mindrot.org>*
14579 * New configuration for the GNU Hurd.
14581 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14583 * Initial code to implement OCSP basic response verify. This
14584 is currently incomplete. Currently just finds the signer's
14585 certificate and verifies the signature on the response.
14589 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14590 value of OPENSSLDIR. This is available via the new '-d' option
14591 to 'openssl version', and is also included in 'openssl version -a'.
14595 * Allowing defining memory allocation callbacks that will be given
14596 file name and line number information in additional arguments
14597 (a `const char*` and an int). The basic functionality remains, as
14598 well as the original possibility to just replace malloc(),
14599 realloc() and free() by functions that do not know about these
14600 additional arguments. To register and find out the current
14601 settings for extended allocation functions, the following
14602 functions are provided:
14604 CRYPTO_set_mem_ex_functions
14605 CRYPTO_set_locked_mem_ex_functions
14606 CRYPTO_get_mem_ex_functions
14607 CRYPTO_get_locked_mem_ex_functions
14609 These work the same way as CRYPTO_set_mem_functions and friends.
14610 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14611 extended allocation function is enabled.
14612 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14613 a conventional allocation function is enabled.
14615 *Richard Levitte, Bodo Moeller*
14617 * Finish off removing the remaining LHASH function pointer casts.
14618 There should no longer be any prototype-casting required when using
14619 the LHASH abstraction, and any casts that remain are "bugs". See
14620 the callback types and macros at the head of lhash.h for details
14621 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14625 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14626 If /dev/[u]random devices are not available or do not return enough
14627 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14629 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14630 /etc/entropy will be queried once each in this sequence, querying stops
14631 when enough entropy was collected without querying more sockets.
14635 * Change the Unix RAND_poll() variant to be able to poll several
14636 random devices, as specified by DEVRANDOM, until a sufficient amount
14637 of data has been collected. We spend at most 10 ms on each file
14638 (select timeout) and read in non-blocking mode. DEVRANDOM now
14639 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14640 (previously it was just the string "/dev/urandom"), so on typical
14641 platforms the 10 ms delay will never occur.
14642 Also separate out the Unix variant to its own file, rand_unix.c.
14643 For VMS, there's a currently-empty rand_vms.c.
14647 * Move OCSP client related routines to ocsp_cl.c. These
14648 provide utility functions which an application needing
14649 to issue a request to an OCSP responder and analyse the
14650 response will typically need: as opposed to those which an
14651 OCSP responder itself would need which will be added later.
14653 OCSP_request_sign() signs an OCSP request with an API similar
14654 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14655 response. OCSP_response_get1_basic() extracts basic response
14656 from response. OCSP_resp_find_status(): finds and extracts status
14657 information from an OCSP_CERTID structure (which will be created
14658 when the request structure is built). These are built from lower
14659 level functions which work on OCSP_SINGLERESP structures but
14660 won't normally be used unless the application wishes to examine
14661 extensions in the OCSP response for example.
14663 Replace nonce routines with a pair of functions.
14664 OCSP_request_add1_nonce() adds a nonce value and optionally
14665 generates a random value. OCSP_check_nonce() checks the
14666 validity of the nonce in an OCSP response.
14670 * Change function OCSP_request_add() to OCSP_request_add0_id().
14671 This doesn't copy the supplied OCSP_CERTID and avoids the
14672 need to free up the newly created id. Change return type
14673 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14674 This can then be used to add extensions to the request.
14675 Deleted OCSP_request_new(), since most of its functionality
14676 is now in OCSP_REQUEST_new() (and the case insensitive name
14677 clash) apart from the ability to set the request name which
14678 will be added elsewhere.
14682 * Update OCSP API. Remove obsolete extensions argument from
14683 various functions. Extensions are now handled using the new
14684 OCSP extension code. New simple OCSP HTTP function which
14685 can be used to send requests and parse the response.
14689 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14690 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14691 uses the special reorder version of SET OF to sort the attributes
14692 and reorder them to match the encoded order. This resolves a long
14693 standing problem: a verify on a PKCS7 structure just after signing
14694 it used to fail because the attribute order did not match the
14695 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14696 it uses the received order. This is necessary to tolerate some broken
14697 software that does not order SET OF. This is handled by encoding
14698 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14699 to produce the required SET OF.
14703 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14704 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14705 files to get correct declarations of the ASN.1 item variables.
14709 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14710 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14711 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14712 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14713 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14714 ASN1_ITEM and no wrapper functions.
14718 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14719 replace the old function pointer based I/O routines. Change most of
14720 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14724 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14725 lines, recognize more "algorithms" that can be deselected, and make
14726 it complain about algorithm deselection that isn't recognised.
14730 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14731 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14732 to use new functions. Add NO_ASN1_OLD which can be set to remove
14733 some old style ASN1 functions: this can be used to determine if old
14734 code will still work when these eventually go away.
14738 * New extension functions for OCSP structures, these follow the
14739 same conventions as certificates and CRLs.
14743 * New function X509V3_add1_i2d(). This automatically encodes and
14744 adds an extension. Its behaviour can be customised with various
14745 flags to append, replace or delete. Various wrappers added for
14746 certificates and CRLs.
14750 * Fix to avoid calling the underlying ASN1 print routine when
14751 an extension cannot be parsed. Correct a typo in the
14752 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14756 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14757 entries for variables.
14761 * Add functionality to `apps/openssl.c` for detecting locking
14762 problems: As the program is single-threaded, all we have
14763 to do is register a locking callback using an array for
14764 storing which locks are currently held by the program.
14768 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14769 SSL_get_ex_data_X509_STORE_idx(), which is used in
14770 ssl_verify_cert_chain() and thus can be called at any time
14771 during TLS/SSL handshakes so that thread-safety is essential.
14772 Unfortunately, the ex_data design is not at all suited
14773 for multi-threaded use, so it probably should be abolished.
14777 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14779 *Broadcom, tweaked and integrated by Geoff Thorpe*
14781 * Move common extension printing code to new function
14782 X509V3_print_extensions(). Reorganise OCSP print routines and
14783 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14787 * New function X509_signature_print() to remove duplication in some
14792 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14793 set (this was treated exactly the same as SET OF previously). This
14794 is used to reorder the STACK representing the structure to match the
14795 encoding. This will be used to get round a problem where a PKCS7
14796 structure which was signed could not be verified because the STACK
14797 order did not reflect the encoded order.
14801 * Reimplement the OCSP ASN1 module using the new code.
14805 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14806 for its ASN1 operations. The old style function pointers still exist
14807 for now but they will eventually go away.
14811 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14812 completely replaces the old ASN1 functionality with a table driven
14813 encoder and decoder which interprets an ASN1_ITEM structure describing
14814 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14815 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14816 has also been converted to the new form.
14820 * Change BN_mod_exp_recp so that negative moduli are tolerated
14821 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14822 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14823 for negative moduli.
14827 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14828 of not touching the result's sign bit.
14832 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14837 * Changed the LHASH code to use prototypes for callbacks, and created
14838 macros to declare and implement thin (optionally static) functions
14839 that provide type-safety and avoid function pointer casting for the
14840 type-specific callbacks.
14844 * Added Kerberos Cipher Suites to be used with TLS, as written in
14846 *Veers Staats <staatsvr@asc.hpc.mil>,
14847 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14849 * Reformat the FAQ so the different questions and answers can be divided
14850 in sections depending on the subject.
14854 * Have the zlib compression code load ZLIB.DLL dynamically under
14859 * New function BN_mod_sqrt for computing square roots modulo a prime
14860 (using the probabilistic Tonelli-Shanks algorithm unless
14861 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14862 be handled deterministically).
14864 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14866 * Make BN_mod_inverse faster by explicitly handling small quotients
14867 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14868 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14872 * New function BN_kronecker.
14876 * Fix BN_gcd so that it works on negative inputs; the result is
14877 positive unless both parameters are zero.
14878 Previously something reasonably close to an infinite loop was
14879 possible because numbers could be growing instead of shrinking
14880 in the implementation of Euclid's algorithm.
14884 * Fix BN_is_word() and BN_is_one() macros to take into account the
14885 sign of the number in question.
14887 Fix BN_is_word(a,w) to work correctly for w == 0.
14889 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14890 because its test if the absolute value of 'a' equals 'w'.
14891 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14892 it exists mostly for use in the implementations of BN_is_zero(),
14893 BN_is_one(), and BN_is_word().
14897 * New function BN_swap.
14901 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14902 the exponentiation functions are more likely to produce reasonable
14903 results on negative inputs.
14907 * Change BN_mod_mul so that the result is always non-negative.
14908 Previously, it could be negative if one of the factors was negative;
14909 I don't think anyone really wanted that behaviour.
14913 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14914 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14915 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14916 and add new functions:
14925 BN_mod_lshift1_quick
14927 BN_mod_lshift_quick
14929 These functions always generate non-negative results.
14931 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14932 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14934 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14935 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14936 be reduced modulo `m`.
14938 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14941 The following entry accidentally appeared in the CHANGES file
14942 distributed with OpenSSL 0.9.7. The modifications described in
14943 it do *not* apply to OpenSSL 0.9.7.
14945 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14946 was actually never needed) and in BN_mul(). The removal in BN_mul()
14947 required a small change in bn_mul_part_recursive() and the addition
14948 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14949 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14950 bn_sub_words() and bn_add_words() except they take arrays with
14956 * In 'openssl passwd', verify passwords read from the terminal
14957 unless the '-salt' option is used (which usually means that
14958 verification would just waste user's time since the resulting
14959 hash is going to be compared with some given password hash)
14960 or the new '-noverify' option is used.
14962 This is an incompatible change, but it does not affect
14963 non-interactive use of 'openssl passwd' (passwords on the command
14964 line, '-stdin' option, '-in ...' option) and thus should not
14965 cause any problems.
14969 * Remove all references to RSAref, since there's no more need for it.
14973 * Make DSO load along a path given through an environment variable
14974 (SHLIB_PATH) with shl_load().
14978 * Constify the ENGINE code as a result of BIGNUM constification.
14979 Also constify the RSA code and most things related to it. In a
14980 few places, most notable in the depth of the ASN.1 code, ugly
14981 casts back to non-const were required (to be solved at a later
14986 * Make it so the openssl application has all engines loaded by default.
14990 * Constify the BIGNUM routines a little more.
14994 * Add the following functions:
14996 ENGINE_load_cswift()
14998 ENGINE_load_atalla()
14999 ENGINE_load_nuron()
15000 ENGINE_load_builtin_engines()
15002 That way, an application can itself choose if external engines that
15003 are built-in in OpenSSL shall ever be used or not. The benefit is
15004 that applications won't have to be linked with libdl or other dso
15005 libraries unless it's really needed.
15007 Changed 'openssl engine' to load all engines on demand.
15008 Changed the engine header files to avoid the duplication of some
15009 declarations (they differed!).
15013 * 'openssl engine' can now list capabilities.
15017 * Better error reporting in 'openssl engine'.
15021 * Never call load_dh_param(NULL) in s_server.
15025 * Add engine application. It can currently list engines by name and
15026 identity, and test if they are actually available.
15030 * Improve RPM specification file by forcing symbolic linking and making
15031 sure the installed documentation is also owned by root.root.
15033 *Damien Miller <djm@mindrot.org>*
15035 * Give the OpenSSL applications more possibilities to make use of
15036 keys (public as well as private) handled by engines.
15040 * Add OCSP code that comes from CertCo.
15044 * Add VMS support for the Rijndael code.
15048 * Added untested support for Nuron crypto accelerator.
15052 * Add support for external cryptographic devices. This code was
15053 previously distributed separately as the "engine" branch.
15055 *Geoff Thorpe, Richard Levitte*
15057 * Rework the filename-translation in the DSO code. It is now possible to
15058 have far greater control over how a "name" is turned into a filename
15059 depending on the operating environment and any oddities about the
15060 different shared library filenames on each system.
15064 * Support threads on FreeBSD-elf in Configure.
15068 * Fix for SHA1 assembly problem with MASM: it produces
15069 warnings about corrupt line number information when assembling
15070 with debugging information. This is caused by the overlapping
15073 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15076 NCONF_get_number() has no error checking at all. As a replacement,
15077 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15078 promoted strongly. The old NCONF_get_number is kept around for
15079 binary backward compatibility.
15080 Make it possible for methods to load from something other than a BIO,
15081 by providing a function pointer that is given a name instead of a BIO.
15082 For example, this could be used to load configuration data from an
15087 * Fix for non blocking accept BIOs. Added new I/O special reason
15088 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15089 with non blocking I/O was not possible because no retry code was
15090 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15095 * Added the beginnings of Rijndael support.
15099 * Fix for bug in DirectoryString mask setting. Add support for
15100 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15101 to allow certificate printing to more controllable, additional
15102 'certopt' option to 'x509' to allow new printing options to be
15107 * Clean old EAY MD5 hack from e_os.h.
15111 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15113 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15114 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15116 *Joe Orton, Steve Henson*
15118 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15120 * Fix additional bug revealed by the NISCC test suite:
15122 Stop bug triggering large recursion when presented with
15123 certain ASN.1 tags ([CVE-2003-0851])
15127 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15129 * Fix various bugs revealed by running the NISCC test suite:
15131 Stop out of bounds reads in the ASN1 code when presented with
15132 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15134 If verify callback ignores invalid public key errors don't try to check
15135 certificate signature with the NULL public key.
15139 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15140 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15145 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15146 extra data after the compression methods not only for TLS 1.0
15147 but also for SSL 3.0 (as required by the specification).
15149 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15151 * Change X509_certificate_type() to mark the key as exported/exportable
15152 when it's 512 *bits* long, not 512 bytes.
15156 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15158 * Countermeasure against the Klima-Pokorny-Rosa extension of
15159 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15160 a protocol version number mismatch like a decryption error
15161 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15165 * Turn on RSA blinding by default in the default implementation
15166 to avoid a timing attack. Applications that don't want it can call
15167 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15168 They would be ill-advised to do so in most cases.
15170 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15172 * Change RSA blinding code so that it works when the PRNG is not
15173 seeded (in this case, the secret RSA exponent is abused as
15174 an unpredictable seed -- if it is not unpredictable, there
15175 is no point in blinding anyway). Make RSA blinding thread-safe
15176 by remembering the creator's thread ID in rsa->blinding and
15177 having all other threads use local one-time blinding factors
15178 (this requires more computation than sharing rsa->blinding, but
15179 avoids excessive locking; and if an RSA object is not shared
15180 between threads, blinding will still be very fast).
15184 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15186 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15187 via timing by performing a MAC computation even if incorrect
15188 block cipher padding has been found. This is a countermeasure
15189 against active attacks where the attacker has to distinguish
15190 between bad padding and a MAC verification error. ([CVE-2003-0078])
15192 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15193 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15194 Martin Vuagnoux (EPFL, Ilion)*
15196 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15198 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15199 memory from its contents. This is done with a counter that will
15200 place alternating values in each byte. This can be used to solve
15201 two issues: 1) the removal of calls to memset() by highly optimizing
15202 compilers, and 2) cleansing with other values than 0, since those can
15203 be read through on certain media, for example a swap space on disk.
15207 * Bugfix: client side session caching did not work with external caching,
15208 because the session->cipher setting was not restored when reloading
15209 from the external cache. This problem was masked, when
15210 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15211 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15215 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15216 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15218 *Zeev Lieber <zeev-l@yahoo.com>*
15220 * Undo an undocumented change introduced in 0.9.6e which caused
15221 repeated calls to OpenSSL_add_all_ciphers() and
15222 OpenSSL_add_all_digests() to be ignored, even after calling
15227 * Change the default configuration reader to deal with last line not
15228 being properly terminated.
15232 * Change X509_NAME_cmp() so it applies the special rules on handling
15233 DN values that are of type PrintableString, as well as RDNs of type
15234 emailAddress where the value has the type ia5String.
15236 *stefank@valicert.com via Richard Levitte*
15238 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15239 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15240 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15241 the bitwise-OR of the two for use by the majority of applications
15242 wanting this behaviour, and update the docs. The documented
15243 behaviour and actual behaviour were inconsistent and had been
15244 changing anyway, so this is more a bug-fix than a behavioural
15247 *Geoff Thorpe, diagnosed by Nadav Har'El*
15249 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15250 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15254 * Fix initialization code race conditions in
15255 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15256 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15257 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15258 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15259 ssl2_get_cipher_by_char(),
15260 ssl3_get_cipher_by_char().
15262 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15264 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15265 the cached sessions are flushed, as the remove_cb() might use ex_data
15266 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15267 (see [openssl.org #212]).
15269 *Geoff Thorpe, Lutz Jaenicke*
15271 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15272 length, instead of the encoding length to d2i_ASN1_OBJECT.
15276 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15278 * [In 0.9.6g-engine release:]
15279 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15281 *Lynn Gazis <lgazis@rainbow.com>*
15283 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15285 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15286 and get fix the header length calculation.
15287 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15288 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15290 * Use proper error handling instead of 'assertions' in buffer
15291 overflow checks added in 0.9.6e. This prevents DoS (the
15292 assertions could call abort()).
15294 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15296 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15298 * Add various sanity checks to asn1_get_length() to reject
15299 the ASN1 length bytes if they exceed sizeof(long), will appear
15300 negative or the content length exceeds the length of the
15303 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15305 * Fix cipher selection routines: ciphers without encryption had no flags
15306 for the cipher strength set and where therefore not handled correctly
15307 by the selection routines (PR #130).
15311 * Fix EVP_dsa_sha macro.
15316 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15317 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15318 that was added in OpenSSL 0.9.6d.
15320 As the countermeasure turned out to be incompatible with some
15321 broken SSL implementations, the new option is part of SSL_OP_ALL.
15322 SSL_OP_ALL is usually employed when compatibility with weird SSL
15323 implementations is desired (e.g. '-bugs' option to 's_client' and
15324 's_server'), so the new option is automatically set in many
15329 * Changes in security patch:
15331 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15332 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15333 Air Force Materiel Command, USAF, under agreement number
15336 * Add various sanity checks to asn1_get_length() to reject
15337 the ASN1 length bytes if they exceed sizeof(long), will appear
15338 negative or the content length exceeds the length of the
15339 supplied buffer. ([CVE-2002-0659])
15341 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15343 * Assertions for various potential buffer overflows, not known to
15344 happen in practice.
15346 *Ben Laurie (CHATS)*
15348 * Various temporary buffers to hold ASCII versions of integers were
15349 too small for 64 bit platforms. ([CVE-2002-0655])
15350 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15352 * Remote buffer overflow in SSL3 protocol - an attacker could
15353 supply an oversized session ID to a client. ([CVE-2002-0656])
15355 *Ben Laurie (CHATS)*
15357 * Remote buffer overflow in SSL2 protocol - an attacker could
15358 supply an oversized client master key. ([CVE-2002-0656])
15360 *Ben Laurie (CHATS)*
15362 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15364 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15365 encoded as NULL) with id-dsa-with-sha1.
15367 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15369 * Check various `X509_...()` return values in `apps/req.c`.
15371 *Nils Larsch <nla@trustcenter.de>*
15373 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15374 an end-of-file condition would erroneously be flagged, when the CRLF
15375 was just at the end of a processed block. The bug was discovered when
15376 processing data through a buffering memory BIO handing the data to a
15377 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15378 <ptsekov@syntrex.com> and Nedelcho Stanev.
15382 * Implement a countermeasure against a vulnerability recently found
15383 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15384 before application data chunks to avoid the use of known IVs
15385 with data potentially chosen by the attacker.
15389 * Fix length checks in ssl3_get_client_hello().
15393 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15394 to prevent ssl3_read_internal() from incorrectly assuming that
15395 ssl3_read_bytes() found application data while handshake
15396 processing was enabled when in fact s->s3->in_read_app_data was
15397 merely automatically cleared during the initial handshake.
15399 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15401 * Fix object definitions for Private and Enterprise: they were not
15402 recognized in their shortname (=lowercase) representation. Extend
15403 obj_dat.pl to issue an error when using undefined keywords instead
15404 of silently ignoring the problem (Svenning Sorensen
15405 <sss@sss.dnsalias.net>).
15409 * Fix DH_generate_parameters() so that it works for 'non-standard'
15410 generators, i.e. generators other than 2 and 5. (Previously, the
15411 code did not properly initialise the 'add' and 'rem' values to
15412 BN_generate_prime().)
15414 In the new general case, we do not insist that 'generator' is
15415 actually a primitive root: This requirement is rather pointless;
15416 a generator of the order-q subgroup is just as good, if not
15421 * Map new X509 verification errors to alerts. Discovered and submitted by
15422 Tom Wu <tom@arcot.com>.
15426 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15427 returning non-zero before the data has been completely received
15428 when using non-blocking I/O.
15430 *Bodo Moeller; problem pointed out by John Hughes*
15432 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15434 *Ben Laurie, Lutz Jaenicke*
15436 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15437 Yoram Zahavi <YoramZ@gilian.com>).
15441 * Add information about CygWin 1.3 and on, and preserve proper
15442 configuration for the versions before that.
15444 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15446 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15447 check whether we deal with a copy of a session and do not delete from
15448 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15449 <izhar@checkpoint.com>.
15453 * Do not store session data into the internal session cache, if it
15454 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15455 flag is set). Proposed by Aslam <aslam@funk.com>.
15459 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15464 * [In 0.9.6d-engine release:]
15465 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15467 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15469 * Add the configuration target linux-s390x.
15471 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15473 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15474 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15475 variable as an indication that a ClientHello message has been
15476 received. As the flag value will be lost between multiple
15477 invocations of ssl3_accept when using non-blocking I/O, the
15478 function may not be aware that a handshake has actually taken
15479 place, thus preventing a new session from being added to the
15482 To avoid this problem, we now set s->new_session to 2 instead of
15483 using a local variable.
15485 *Lutz Jaenicke, Bodo Moeller*
15487 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15488 if the SSL_R_LENGTH_MISMATCH error is detected.
15490 *Geoff Thorpe, Bodo Moeller*
15492 * New 'shared_ldflag' column in Configure platform table.
15496 * Fix EVP_CIPHER_mode macro.
15498 *"Dan S. Camper" <dan@bti.net>*
15500 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15501 type, we must throw them away by setting rr->length to 0.
15503 *D P Chang <dpc@qualys.com>*
15505 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15507 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15508 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15509 worked incorrectly for those cases where range = `10..._2` and
15510 `3*range` is two bits longer than range.)
15514 * Only add signing time to PKCS7 structures if it is not already
15519 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15520 OBJ_ld_ce should be OBJ_id_ce.
15521 Also some ip-pda OIDs in crypto/objects/objects.txt were
15522 incorrect (cf. RFC 3039).
15524 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15526 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15527 returns early because it has nothing to do.
15529 *Andy Schneider <andy.schneider@bjss.co.uk>*
15531 * [In 0.9.6c-engine release:]
15532 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15534 *Andy Schneider <andy.schneider@bjss.co.uk>*
15536 * [In 0.9.6c-engine release:]
15537 Add support for Cryptographic Appliance's keyserver technology.
15538 (Use engine 'keyclient')
15540 *Cryptographic Appliances and Geoff Thorpe*
15542 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15543 is called via tools/c89.sh because arguments have to be
15544 rearranged (all '-L' options must appear before the first object
15547 *Richard Shapiro <rshapiro@abinitio.com>*
15549 * [In 0.9.6c-engine release:]
15550 Add support for Broadcom crypto accelerator cards, backported
15553 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15555 * [In 0.9.6c-engine release:]
15556 Add support for SureWare crypto accelerator cards from
15557 Baltimore Technologies. (Use engine 'sureware')
15559 *Baltimore Technologies and Mark Cox*
15561 * [In 0.9.6c-engine release:]
15562 Add support for crypto accelerator cards from Accelerated
15563 Encryption Processing, www.aep.ie. (Use engine 'aep')
15565 *AEP Inc. and Mark Cox*
15567 * Add a configuration entry for gcc on UnixWare.
15569 *Gary Benson <gbenson@redhat.com>*
15571 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15572 messages are stored in a single piece (fixed-length part and
15573 variable-length part combined) and fix various bugs found on the way.
15577 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15578 instead. BIO_gethostbyname() does not know what timeouts are
15579 appropriate, so entries would stay in cache even when they have
15581 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15583 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15584 faced with a pathologically small ClientHello fragment that does
15585 not contain client_version: Instead of aborting with an error,
15586 simply choose the highest available protocol version (i.e.,
15587 TLS 1.0 unless it is disabled). In practice, ClientHello
15588 messages are never sent like this, but this change gives us
15589 strictly correct behaviour at least for TLS.
15593 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15594 never resets s->method to s->ctx->method when called from within
15595 one of the SSL handshake functions.
15597 *Bodo Moeller; problem pointed out by Niko Baric*
15599 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15600 (sent using the client's version number) if client_version is
15601 smaller than the protocol version in use. Also change
15602 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15603 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15604 the client will at least see that alert.
15608 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15613 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15614 client receives HelloRequest while in a handshake.
15616 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15618 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15619 should end in 'break', not 'goto end' which circumvents various
15620 cleanups done in state SSL_ST_OK. But session related stuff
15621 must be disabled for SSL_ST_OK in the case that we just sent a
15624 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15625 before just sending a HelloRequest.
15627 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15629 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15630 reveal whether illegal block cipher padding was found or a MAC
15631 verification error occurred. (Neither SSLerr() codes nor alerts
15632 are directly visible to potential attackers, but the information
15633 may leak via logfiles.)
15635 Similar changes are not required for the SSL 2.0 implementation
15636 because the number of padding bytes is sent in clear for SSL 2.0,
15637 and the extra bytes are just ignored. However ssl/s2_pkt.c
15638 failed to verify that the purported number of padding bytes is in
15643 * Add OpenUNIX-8 support including shared libraries
15644 (Boyd Lynn Gerber <gerberb@zenez.com>).
15648 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15649 'wristwatch attack' using huge encoding parameters (cf.
15650 James H. Manger's CRYPTO 2001 paper). Note that the
15651 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15652 encoding parameters and hence was not vulnerable.
15656 * BN_sqr() bug fix.
15658 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15660 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15661 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15662 followed by modular reduction.
15664 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15666 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15667 equivalent based on BN_pseudo_rand() instead of BN_rand().
15671 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15672 This function was broken, as the check for a new client hello message
15673 to handle SGC did not allow these large messages.
15674 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15678 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15682 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15683 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15687 * Rework the configuration and shared library support for Tru64 Unix.
15688 The configuration part makes use of modern compiler features and
15689 still retains old compiler behavior for those that run older versions
15690 of the OS. The shared library support part includes a variant that
15691 uses the RPATH feature, and is available through the special
15692 configuration target "alpha-cc-rpath", which will never be selected
15695 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15697 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15698 with the same message size as in ssl3_get_certificate_request().
15699 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15700 messages might inadvertently be reject as too long.
15702 *Petr Lampa <lampa@fee.vutbr.cz>*
15704 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15708 * Modified SSL library such that the verify_callback that has been set
15709 specifically for an SSL object with SSL_set_verify() is actually being
15710 used. Before the change, a verify_callback set with this function was
15711 ignored and the verify_callback() set in the SSL_CTX at the time of
15712 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15713 to allow the necessary settings.
15717 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15718 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15719 done automatically (in contradiction to the requirements of the C
15720 standard). This made problems when used from OpenSSH.
15724 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15725 dh->length and always used
15727 BN_rand_range(priv_key, dh->p).
15729 BN_rand_range() is not necessary for Diffie-Hellman, and this
15730 specific range makes Diffie-Hellman unnecessarily inefficient if
15731 dh->length (recommended exponent length) is much smaller than the
15732 length of dh->p. We could use BN_rand_range() if the order of
15733 the subgroup was stored in the DH structure, but we only have
15738 BN_rand(priv_key, l, ...)
15740 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15747 RSA_eay_public_encrypt
15748 RSA_eay_private_decrypt
15749 RSA_eay_private_encrypt (signing)
15750 RSA_eay_public_decrypt (signature verification)
15752 (default implementations for RSA_public_encrypt,
15753 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15754 always reject numbers >= n.
15758 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15759 to synchronize access to 'locking_thread'. This is necessary on
15760 systems where access to 'locking_thread' (an 'unsigned long'
15761 variable) is not atomic.
15765 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15766 *before* setting the 'crypto_lock_rand' flag. The previous code had
15767 a race condition if 0 is a valid thread ID.
15769 *Travis Vitek <vitek@roguewave.com>*
15771 * Add support for shared libraries under Irix.
15773 *Albert Chin-A-Young <china@thewrittenword.com>*
15775 * Add configuration option to build on Linux on both big-endian and
15776 little-endian MIPS.
15778 *Ralf Baechle <ralf@uni-koblenz.de>*
15780 * Add the possibility to create shared libraries on HP-UX.
15784 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15786 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15787 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15788 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15789 PRNG state recovery was possible based on the output of
15790 one PRNG request appropriately sized to gain knowledge on
15791 'md' followed by enough consecutive 1-byte PRNG requests
15792 to traverse all of 'state'.
15794 1. When updating 'md_local' (the current thread's copy of 'md')
15795 during PRNG output generation, hash all of the previous
15796 'md_local' value, not just the half used for PRNG output.
15798 2. Make the number of bytes from 'state' included into the hash
15799 independent from the number of PRNG bytes requested.
15801 The first measure alone would be sufficient to avoid
15802 Markku-Juhani's attack. (Actually it had never occurred
15803 to me that the half of 'md_local' used for chaining was the
15804 half from which PRNG output bytes were taken -- I had always
15805 assumed that the secret half would be used.) The second
15806 measure makes sure that additional data from 'state' is never
15807 mixed into 'md_local' in small portions; this heuristically
15808 further strengthens the PRNG.
15812 * Fix crypto/bn/asm/mips3.s.
15816 * When only the key is given to "enc", the IV is undefined. Print out
15817 an error message in this case.
15821 * Handle special case when X509_NAME is empty in X509 printing routines.
15825 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15826 positive and less than q.
15830 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15831 used: it isn't thread safe and the add_lock_callback should handle
15834 *Paul Rose <Paul.Rose@bridge.com>*
15836 * Verify that incoming data obeys the block size in
15837 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15843 *Ulf Möller, Bodo Möller*
15845 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15846 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15847 when fixing the server behaviour for backwards-compatible 'client
15848 hello' messages. (Note that the attack is impractical against
15849 SSL 3.0 and TLS 1.0 anyway because length and version checking
15850 means that the probability of guessing a valid ciphertext is
15851 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15854 Before 0.9.5, the countermeasure (hide the error by generating a
15855 random 'decryption result') did not work properly because
15856 ERR_clear_error() was missing, meaning that SSL_get_error() would
15857 detect the supposedly ignored error.
15859 Both problems are now fixed.
15863 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15864 (previously it was 1024).
15868 * Fix for compatibility mode trust settings: ignore trust settings
15869 unless some valid trust or reject settings are present.
15873 * Fix for blowfish EVP: its a variable length cipher.
15877 * Fix various bugs related to DSA S/MIME verification. Handle missing
15878 parameters in DSA public key structures and return an error in the
15879 DSA routines if parameters are absent.
15883 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15884 in the current directory if neither $RANDFILE nor $HOME was set.
15885 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15886 caused some confusion to Windows users who haven't defined $HOME.
15887 Thus RAND_file_name() is changed again: e_os.h can define a
15888 DEFAULT_HOME, which will be used if $HOME is not set.
15889 For Windows, we use "C:"; on other platforms, we still require
15890 environment variables.
15892 * Move 'if (!initialized) RAND_poll()' into regions protected by
15893 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15894 having multiple threads call RAND_poll() concurrently.
15898 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15899 combination of a flag and a thread ID variable.
15900 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15901 flag), *other* threads can enter ssleay_add_bytes without obeying
15902 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15903 that they do not hold after the first thread unsets add_do_not_lock).
15907 * Change bctest again: '-x' expressions are not available in all
15908 versions of 'test'.
15912 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15914 * Fix a couple of memory leaks in PKCS7_dataDecode()
15916 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15918 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15919 the default extension for executables, if any. Also, make the perl
15920 scripts that use symlink() to test if it really exists and use "cp"
15921 if it doesn't. All this made OpenSSL compilable and installable in
15926 * Fix for asn1_GetSequence() for indefinite length constructed data.
15927 If SEQUENCE is length is indefinite just set c->slen to the total
15928 amount of data available.
15930 *Steve Henson, reported by shige@FreeBSD.org*
15932 *This change does not apply to 0.9.7.*
15934 * Change bctest to avoid here-documents inside command substitution
15935 (workaround for FreeBSD /bin/sh bug).
15936 For compatibility with Ultrix, avoid shell functions (introduced
15937 in the bctest version that searches along $PATH).
15941 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15942 with des_encrypt() defined on some operating systems, like Solaris
15947 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15948 On the Importance of Eliminating Errors in Cryptographic
15949 Computations, J. Cryptology 14 (2001) 2, 101-119,
15950 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15954 * MIPS assembler BIGNUM division bug fix.
15958 * Disabled incorrect Alpha assembler code.
15962 * Fix PKCS#7 decode routines so they correctly update the length
15963 after reading an EOC for the EXPLICIT tag.
15967 *This change does not apply to 0.9.7.*
15969 * Fix bug in PKCS#12 key generation routines. This was triggered
15970 if a 3DES key was generated with a 0 initial byte. Include
15971 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15972 (but broken) behaviour.
15976 * Enhance bctest to search for a working bc along $PATH and print
15979 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15981 * Fix memory leaks in err.c: free err_data string if necessary;
15982 don't write to the wrong index in ERR_set_error_data.
15986 * Implement ssl23_peek (analogous to ssl23_read), which previously
15991 * Replace rdtsc with `_emit` statements for VC++ version 5.
15993 *Jeremy Cooper <jeremy@baymoo.org>*
15995 * Make it possible to reuse SSLv2 sessions.
15999 * In copy_email() check for >= 0 as a return value for
16000 X509_NAME_get_index_by_NID() since 0 is a valid index.
16002 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16004 * Avoid coredump with unsupported or invalid public keys by checking if
16005 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16006 PKCS7_verify() fails with non detached data.
16010 * Don't use getenv in library functions when run as setuid/setgid.
16011 New function OPENSSL_issetugid().
16015 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16016 due to incorrect handling of multi-threading:
16018 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16020 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16022 3. Count how many times MemCheck_off() has been called so that
16023 nested use can be treated correctly. This also avoids
16024 inband-signalling in the previous code (which relied on the
16025 assumption that thread ID 0 is impossible).
16029 * Add "-rand" option also to s_client and s_server.
16033 * Fix CPU detection on Irix 6.x.
16034 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16035 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16037 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16042 *This change does not apply to 0.9.7.*
16044 * Use the cached encoding of an X509_NAME structure rather than
16045 copying it. This is apparently the reason for the libsafe "errors"
16046 but the code is actually correct.
16050 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16051 Bleichenbacher's DSA attack.
16052 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16053 to be set and top=0 forces the highest bit to be set; top=-1 is new
16054 and leaves the highest bit random.
16056 *Ulf Moeller, Bodo Moeller*
16058 * In the `NCONF_...`-based implementations for `CONF_...` queries
16059 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16060 a temporary CONF structure with the data component set to NULL
16061 (which gives segmentation faults in lh_retrieve).
16062 Instead, use NULL for the CONF pointer in CONF_get_string and
16063 CONF_get_number (which may use environment variables) and directly
16064 return NULL from CONF_get_section.
16068 * Fix potential buffer overrun for EBCDIC.
16072 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16073 keyUsage if basicConstraints absent for a CA.
16077 * Make SMIME_write_PKCS7() write mail header values with a format that
16078 is more generally accepted (no spaces before the semicolon), since
16079 some programs can't parse those values properly otherwise. Also make
16080 sure BIO's that break lines after each write do not create invalid
16085 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16086 macros previously used would not encode an empty SEQUENCE OF
16087 and break the signature.
16091 *This change does not apply to 0.9.7.*
16093 * Zero the premaster secret after deriving the master secret in
16098 * Add some EVP_add_digest_alias registrations (as found in
16099 OpenSSL_add_all_digests()) to SSL_library_init()
16100 aka OpenSSL_add_ssl_algorithms(). This provides improved
16101 compatibility with peers using X.509 certificates
16102 with unconventional AlgorithmIdentifier OIDs.
16106 * Fix for Irix with NO_ASM.
16108 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16110 * ./config script fixes.
16112 *Ulf Moeller, Richard Levitte*
16114 * Fix 'openssl passwd -1'.
16118 * Change PKCS12_key_gen_asc() so it can cope with non null
16119 terminated strings whose length is passed in the passlen
16120 parameter, for example from PEM callbacks. This was done
16121 by adding an extra length parameter to asc2uni().
16123 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16125 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16126 call failed, free the DSA structure.
16130 * Fix to uni2asc() to cope with zero length Unicode strings.
16131 These are present in some PKCS#12 files.
16135 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16136 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16137 when writing a 32767 byte record.
16139 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16141 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16142 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16144 (RSA objects have a reference count access to which is protected
16145 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16146 so they are meant to be shared between threads.)
16147 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16148 "Reddie, Steven" <Steven.Reddie@ca.com>*
16150 * Fix a deadlock in CRYPTO_mem_leaks().
16154 * Use better test patterns in bntest.
16158 * rand_win.c fix for Borland C.
16162 * BN_rshift bugfix for n == 0.
16166 * Add a 'bctest' script that checks for some known 'bc' bugs
16167 so that 'make test' does not abort just because 'bc' is broken.
16171 * Store verify_result within SSL_SESSION also for client side to
16172 avoid potential security hole. (Re-used sessions on the client side
16173 always resulted in verify_result==X509_V_OK, not using the original
16174 result of the server certificate verification.)
16178 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16179 SSL3_RT_APPLICATION_DATA, return 0.
16180 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16185 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16186 releases, have been re-implemented by renaming the previous
16187 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16188 and ssl3_read_internal, respectively, and adding 'peek' parameters
16189 to them. The new ssl[23]_{read,peek} functions are calls to
16190 ssl[23]_read_internal with the 'peek' flag set appropriately.
16191 A 'peek' parameter has also been added to ssl3_read_bytes, which
16192 does the actual work for ssl3_read_internal.
16196 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16197 the method-specific "init()" handler. Also clean up ex_data after
16198 calling the method-specific "finish()" handler. Previously, this was
16199 happening the other way round.
16203 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16204 The previous value, 12, was not always sufficient for BN_mod_exp().
16208 * Make sure that shared libraries get the internal name engine with
16209 the full version number and not just 0. This should mark the
16210 shared libraries as not backward compatible. Of course, this should
16211 be changed again when we can guarantee backward binary compatibility.
16215 * Fix typo in get_cert_by_subject() in by_dir.c
16217 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16219 * Rework the system to generate shared libraries:
16221 - Make note of the expected extension for the shared libraries and
16222 if there is a need for symbolic links from for example libcrypto.so.0
16223 to libcrypto.so.0.9.7. There is extended info in Configure for
16226 - Make as few rebuilds of the shared libraries as possible.
16228 - Still avoid linking the OpenSSL programs with the shared libraries.
16230 - When installing, install the shared libraries separately from the
16235 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16237 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16238 and not in SSL_clear because the latter is also used by the
16239 accept/connect functions; previously, the settings made by
16240 SSL_set_read_ahead would be lost during the handshake.
16242 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16244 * Correct util/mkdef.pl to be selective about disabled algorithms.
16245 Previously, it would create entries for disabled algorithms no
16250 * Added several new manual pages for SSL_* function.
16254 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16256 * In ssl23_get_client_hello, generate an error message when faced
16257 with an initial SSL 3.0/TLS record that is too small to contain the
16258 first two bytes of the ClientHello message, i.e. client_version.
16259 (Note that this is a pathologic case that probably has never happened
16260 in real life.) The previous approach was to use the version number
16261 from the record header as a substitute; but our protocol choice
16262 should not depend on that one because it is not authenticated
16263 by the Finished messages.
16267 * More robust randomness gathering functions for Windows.
16269 *Jeffrey Altman <jaltman@columbia.edu>*
16271 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16272 not set then we don't setup the error code for issuer check errors
16273 to avoid possibly overwriting other errors which the callback does
16274 handle. If an application does set the flag then we assume it knows
16275 what it is doing and can handle the new informational codes
16280 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16281 a general "ANY" type, as such it should be able to decode anything
16282 including tagged types. However it didn't check the class so it would
16283 wrongly interpret tagged types in the same way as their universal
16284 counterpart and unknown types were just rejected. Changed so that the
16285 tagged and unknown types are handled in the same way as a SEQUENCE:
16286 that is the encoding is stored intact. There is also a new type
16287 "V_ASN1_OTHER" which is used when the class is not universal, in this
16288 case we have no idea what the actual type is so we just lump them all
16293 * On VMS, stdout may very well lead to a file that is written to
16294 in a record-oriented fashion. That means that every write() will
16295 write a separate record, which will be read separately by the
16296 programs trying to read from it. This can be very confusing.
16298 The solution is to put a BIO filter in the way that will buffer
16299 text until a linefeed is reached, and then write everything a
16300 line at a time, so every record written will be an actual line,
16301 not chunks of lines and not (usually doesn't happen, but I've
16302 seen it once) several lines in one record. BIO_f_linebuffer() is
16305 Currently, it's a VMS-only method, because that's where it has
16306 been tested well enough.
16310 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16311 it can return incorrect results.
16312 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16313 but it was in 0.9.6-beta[12].)
16317 * Disable the check for content being present when verifying detached
16318 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16319 include zero length content when signing messages.
16323 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16324 BIO_ctrl (for BIO pairs).
16328 * Add DSO method for VMS.
16332 * Bug fix: Montgomery multiplication could produce results with the
16337 * Add RPM specification openssl.spec and modify it to build three
16338 packages. The default package contains applications, application
16339 documentation and run-time libraries. The devel package contains
16340 include files, static libraries and function documentation. The
16341 doc package contains the contents of the doc directory. The original
16342 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16346 * Add a large number of documentation files for many SSL routines.
16348 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16350 * Add a configuration entry for Sony News 4.
16352 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16354 * Don't set the two most significant bits to one when generating a
16355 random number < q in the DSA library.
16359 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16360 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16361 the underlying transport is blocking) if a handshake took place.
16362 (The default behaviour is needed by applications such as s_client
16363 and s_server that use select() to determine when to use SSL_read;
16364 but for applications that know in advance when to expect data, it
16365 just makes things more complicated.)
16369 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16374 * Add a few more EBCDIC conditionals that make `req` and `x509`
16375 work better on such systems.
16377 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16379 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16380 Update PKCS12_parse() so it copies the friendlyName and the
16381 keyid to the certificates aux info.
16385 * Fix bug in PKCS7_verify() which caused an infinite loop
16386 if there was more than one signature.
16388 *Sven Uszpelkat <su@celocom.de>*
16390 * Major change in util/mkdef.pl to include extra information
16391 about each symbol, as well as presenting variables as well
16392 as functions. This change means that there's n more need
16393 to rebuild the .num files when some algorithms are excluded.
16397 * Allow the verify time to be set by an application,
16398 rather than always using the current time.
16402 * Phase 2 verify code reorganisation. The certificate
16403 verify code now looks up an issuer certificate by a
16404 number of criteria: subject name, authority key id
16405 and key usage. It also verifies self signed certificates
16406 by the same criteria. The main comparison function is
16407 X509_check_issued() which performs these checks.
16409 Lot of changes were necessary in order to support this
16410 without completely rewriting the lookup code.
16412 Authority and subject key identifier are now cached.
16414 The LHASH 'certs' is X509_STORE has now been replaced
16415 by a STACK_OF(X509_OBJECT). This is mainly because an
16416 LHASH can't store or retrieve multiple objects with
16417 the same hash value.
16419 As a result various functions (which were all internal
16420 use only) have changed to handle the new X509_STORE
16421 structure. This will break anything that messed round
16422 with X509_STORE internally.
16424 The functions X509_STORE_add_cert() now checks for an
16425 exact match, rather than just subject name.
16427 The X509_STORE API doesn't directly support the retrieval
16428 of multiple certificates matching a given criteria, however
16429 this can be worked round by performing a lookup first
16430 (which will fill the cache with candidate certificates)
16431 and then examining the cache for matches. This is probably
16432 the best we can do without throwing out X509_LOOKUP
16433 entirely (maybe later...).
16435 The X509_VERIFY_CTX structure has been enhanced considerably.
16437 All certificate lookup operations now go via a get_issuer()
16438 callback. Although this currently uses an X509_STORE it
16439 can be replaced by custom lookups. This is a simple way
16440 to bypass the X509_STORE hackery necessary to make this
16441 work and makes it possible to use more efficient techniques
16442 in future. A very simple version which uses a simple
16443 STACK for its trusted certificate store is also provided
16444 using X509_STORE_CTX_trusted_stack().
16446 The verify_cb() and verify() callbacks now have equivalents
16447 in the X509_STORE_CTX structure.
16449 X509_STORE_CTX also has a 'flags' field which can be used
16450 to customise the verify behaviour.
16454 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16455 excludes S/MIME capabilities.
16459 * When a certificate request is read in keep a copy of the
16460 original encoding of the signed data and use it when outputting
16461 again. Signatures then use the original encoding rather than
16462 a decoded, encoded version which may cause problems if the
16463 request is improperly encoded.
16467 * For consistency with other BIO_puts implementations, call
16468 buffer_write(b, ...) directly in buffer_puts instead of calling
16471 In BIO_puts, increment b->num_write as in BIO_write.
16473 *Peter.Sylvester@EdelWeb.fr*
16475 * Fix BN_mul_word for the case where the word is 0. (We have to use
16476 BN_zero, we may not return a BIGNUM with an array consisting of
16477 words set to zero.)
16481 * Avoid calling abort() from within the library when problems are
16482 detected, except if preprocessor symbols have been defined
16483 (such as REF_CHECK, BN_DEBUG etc.).
16487 * New openssl application 'rsautl'. This utility can be
16488 used for low-level RSA operations. DER public key
16489 BIO/fp routines also added.
16493 * New Configure entry and patches for compiling on QNX 4.
16495 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16497 * A demo state-machine implementation was sponsored by
16498 Nuron (<http://www.nuron.com/>) and is now available in
16499 demos/state_machine.
16503 * New options added to the 'dgst' utility for signature
16504 generation and verification.
16508 * Unrecognized PKCS#7 content types are now handled via a
16509 catch all ASN1_TYPE structure. This allows unsupported
16510 types to be stored as a "blob" and an application can
16511 encode and decode it manually.
16515 * Fix various signed/unsigned issues to make a_strex.c
16516 compile under VC++.
16518 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16520 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16521 length if passed a buffer. ASN1_INTEGER_to_BN failed
16522 if passed a NULL BN and its argument was negative.
16524 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16526 * Modification to PKCS#7 encoding routines to output definite
16527 length encoding. Since currently the whole structures are in
16528 memory there's not real point in using indefinite length
16529 constructed encoding. However if OpenSSL is compiled with
16530 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16534 * Added BIO_vprintf() and BIO_vsnprintf().
16538 * Added more prefixes to parse for in the strings written
16539 through a logging bio, to cover all the levels that are available
16540 through syslog. The prefixes are now:
16542 PANIC, EMERG, EMR => LOG_EMERG
16543 ALERT, ALR => LOG_ALERT
16544 CRIT, CRI => LOG_CRIT
16545 ERROR, ERR => LOG_ERR
16546 WARNING, WARN, WAR => LOG_WARNING
16547 NOTICE, NOTE, NOT => LOG_NOTICE
16548 INFO, INF => LOG_INFO
16549 DEBUG, DBG => LOG_DEBUG
16551 and as before, if none of those prefixes are present at the
16552 beginning of the string, LOG_ERR is chosen.
16554 On Win32, the `LOG_*` levels are mapped according to this:
16556 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16557 LOG_WARNING => EVENTLOG_WARNING_TYPE
16558 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16562 * Made it possible to reconfigure with just the configuration
16563 argument "reconf" or "reconfigure". The command line arguments
16564 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16565 and are retrieved from there when reconfiguring.
16571 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16573 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16577 * The obj_dat.pl script was messing up the sorting of object
16578 names. The reason was that it compared the quoted version
16579 of strings as a result "OCSP" > "OCSP Signing" because
16580 " > SPACE. Changed script to store unquoted versions of
16581 names and add quotes on output. It was also omitting some
16582 names from the lookup table if they were given a default
16583 value (that is if SN is missing it is given the same
16584 value as LN and vice versa), these are now added on the
16585 grounds that if an object has a name we should be able to
16586 look it up. Finally added warning output when duplicate
16587 short or long names are found.
16591 * Changes needed for Tandem NSK.
16593 *Scott Uroff <scott@xypro.com>*
16595 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16596 RSA_padding_check_SSLv23(), special padding was never detected
16597 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16598 version rollback attacks was not effective.
16600 In s23_clnt.c, don't use special rollback-attack detection padding
16601 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16602 client; similarly, in s23_srvr.c, don't do the rollback check if
16603 SSL 2.0 is the only protocol enabled in the server.
16607 * Make it possible to get hexdumps of unprintable data with 'openssl
16608 asn1parse'. By implication, the functions ASN1_parse_dump() and
16609 BIO_dump_indent() are added.
16613 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16614 these print out strings and name structures based on various
16615 flags including RFC2253 support and proper handling of
16616 multibyte characters. Added options to the 'x509' utility
16617 to allow the various flags to be set.
16621 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16622 Also change the functions X509_cmp_current_time() and
16623 X509_gmtime_adj() work with an ASN1_TIME structure,
16624 this will enable certificates using GeneralizedTime in validity
16625 dates to be checked.
16629 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16630 negative public key encodings) on by default,
16631 NO_NEG_PUBKEY_BUG can be set to disable it.
16635 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16636 content octets. An i2c_ASN1_OBJECT is unnecessary because
16637 the encoding can be trivially obtained from the structure.
16641 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16642 not read locks (`CRYPTO_r_[un]lock`).
16646 * A first attempt at creating official support for shared
16647 libraries through configuration. I've kept it so the
16648 default is static libraries only, and the OpenSSL programs
16649 are always statically linked for now, but there are
16650 preparations for dynamic linking in place.
16651 This has been tested on Linux and Tru64.
16655 * Randomness polling function for Win9x, as described in:
16656 Peter Gutmann, Software Generation of Practically Strong
16661 * Fix so PRNG is seeded in req if using an already existing
16666 * New options to smime application. -inform and -outform
16667 allow alternative formats for the S/MIME message including
16668 PEM and DER. The -content option allows the content to be
16669 specified separately. This should allow things like Netscape
16670 form signing output easier to verify.
16674 * Fix the ASN1 encoding of tags using the 'long form'.
16678 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16679 STRING types. These convert content octets to and from the
16680 underlying type. The actual tag and length octets are
16681 already assumed to have been read in and checked. These
16682 are needed because all other string types have virtually
16683 identical handling apart from the tag. By having versions
16684 of the ASN1 functions that just operate on content octets
16685 IMPLICIT tagging can be handled properly. It also allows
16686 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16687 and ASN1_INTEGER are identical apart from the tag.
16691 * Change the handling of OID objects as follows:
16693 - New object identifiers are inserted in objects.txt, following
16694 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16695 - objects.pl is used to process obj_mac.num and create a new
16697 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16700 This is currently kind of a hack, and the perl code in objects.pl
16701 isn't very elegant, but it works as I intended. The simplest way
16702 to check that it worked correctly is to look in obj_dat.h and
16703 check the array nid_objs and make sure the objects haven't moved
16704 around (this is important!). Additions are OK, as well as
16705 consistent name changes.
16709 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16713 * Addition of the command line parameter '-rand file' to 'openssl req'.
16714 The given file adds to whatever has already been seeded into the
16715 random pool through the RANDFILE configuration file option or
16716 environment variable, or the default random state file.
16720 * mkstack.pl now sorts each macro group into lexical order.
16721 Previously the output order depended on the order the files
16722 appeared in the directory, resulting in needless rewriting
16727 * Patches to make OpenSSL compile under Win32 again. Mostly
16728 work arounds for the VC++ problem that it treats func() as
16729 func(void). Also stripped out the parts of mkdef.pl that
16730 added extra typesafe functions: these no longer exist.
16734 * Reorganisation of the stack code. The macros are now all
16735 collected in safestack.h . Each macro is defined in terms of
16736 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16737 DEBUG_SAFESTACK is now handled in terms of function casts,
16738 this has the advantage of retaining type safety without the
16739 use of additional functions. If DEBUG_SAFESTACK is not defined
16740 then the non typesafe macros are used instead. Also modified the
16741 mkstack.pl script to handle the new form. Needs testing to see
16742 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16743 the default if no major problems. Similar behaviour for ASN1_SET_OF
16744 and PKCS12_STACK_OF.
16748 * When some versions of IIS use the 'NET' form of private key the
16749 key derivation algorithm is different. Normally MD5(password) is
16750 used as a 128 bit RC4 key. In the modified case
16751 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16752 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16753 as the old Netscape_RSA functions except they have an additional
16754 'sgckey' parameter which uses the modified algorithm. Also added
16755 an -sgckey command line option to the rsa utility. Thanks to
16756 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16757 algorithm to openssl-dev.
16761 * The evp_local.h macros were using 'c.##kname' which resulted in
16762 invalid expansion on some systems (SCO 5.0.5 for example).
16763 Corrected to 'c.kname'.
16765 *Phillip Porch <root@theporch.com>*
16767 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16768 a STACK of email addresses from a certificate or request, these look
16769 in the subject name and the subject alternative name extensions and
16770 omit any duplicate addresses.
16774 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16775 This makes DSA verification about 2 % faster.
16779 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16780 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16781 plus overhead for 1024 bit moduli).
16782 This makes exponentiations about 0.5 % faster for 1024 bit
16783 exponents (as measured by "openssl speed rsa2048").
16787 * Rename memory handling macros to avoid conflicts with other
16789 Malloc => OPENSSL_malloc
16790 Malloc_locked => OPENSSL_malloc_locked
16791 Realloc => OPENSSL_realloc
16792 Free => OPENSSL_free
16796 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16797 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16801 * CygWin32 support.
16803 *John Jarvie <jjarvie@newsguy.com>*
16805 * The type-safe stack code has been rejigged. It is now only compiled
16806 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16807 by default all type-specific stack functions are "#define"d back to
16808 standard stack functions. This results in more streamlined output
16809 but retains the type-safety checking possibilities of the original
16814 * The STACK code has been cleaned up, and certain type declarations
16815 that didn't make a lot of sense have been brought in line. This has
16816 also involved a cleanup of sorts in safestack.h to more correctly
16817 map type-safe stack functions onto their plain stack counterparts.
16818 This work has also resulted in a variety of "const"ifications of
16819 lots of the code, especially `_cmp` operations which should normally
16820 be prototyped with "const" parameters anyway.
16824 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16825 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16826 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16827 where all of 'md' is used each time the PRNG is used, but 'state'
16828 is used only indexed by a cyclic counter. As entropy may not be
16829 well distributed from the beginning, 'md' is important as a
16830 chaining variable. However, the output function chains only half
16831 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16832 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16833 in all of 'state' being rewritten, with the new values depending
16834 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16838 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16839 the handshake is continued after ssl_verify_cert_chain();
16840 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16841 can lead to 'unexplainable' connection aborts later.
16843 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16845 * Major EVP API cipher revision.
16846 Add hooks for extra EVP features. This allows various cipher
16847 parameters to be set in the EVP interface. Support added for variable
16848 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16849 setting of RC2 and RC5 parameters.
16851 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16854 Remove lots of duplicated code from the EVP library. For example *every*
16855 cipher init() function handles the 'iv' in the same way according to the
16856 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16857 for CFB and OFB modes they zero ctx->num.
16859 New functionality allows removal of S/MIME code RC2 hack.
16861 Most of the routines have the same form and so can be declared in terms
16864 By shifting this to the top level EVP_CipherInit() it can be removed from
16865 all individual ciphers. If the cipher wants to handle IVs or keys
16866 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16869 Change lots of functions like EVP_EncryptUpdate() to now return a
16870 value: although software versions of the algorithms cannot fail
16871 any installed hardware versions can.
16875 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16876 this option is set, tolerate broken clients that send the negotiated
16877 protocol version number instead of the requested protocol version
16882 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16883 i.e. non-zero for export ciphersuites, zero otherwise.
16884 Previous versions had this flag inverted, inconsistent with
16885 rsa_tmp_cb (..._TMP_RSA_CB).
16887 *Bodo Moeller; problem reported by Amit Chopra*
16889 * Add missing DSA library text string. Work around for some IIS
16890 key files with invalid SEQUENCE encoding.
16894 * Add a document (doc/standards.txt) that list all kinds of standards
16895 and so on that are implemented in OpenSSL.
16899 * Enhance c_rehash script. Old version would mishandle certificates
16900 with the same subject name hash and wouldn't handle CRLs at all.
16901 Added -fingerprint option to crl utility, to support new c_rehash
16906 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16910 * Fix for SSL server purpose checking. Server checking was
16911 rejecting certificates which had extended key usage present
16912 but no ssl client purpose.
16914 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16916 * Make PKCS#12 code work with no password. The PKCS#12 spec
16917 is a little unclear about how a blank password is handled.
16918 Since the password in encoded as a BMPString with terminating
16919 double NULL a zero length password would end up as just the
16920 double NULL. However no password at all is different and is
16921 handled differently in the PKCS#12 key generation code. NS
16922 treats a blank password as zero length. MSIE treats it as no
16923 password on export: but it will try both on import. We now do
16924 the same: PKCS12_parse() tries zero length and no password if
16925 the password is set to "" or NULL (NULL is now a valid password:
16926 it wasn't before) as does the pkcs12 application.
16930 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16931 perror when PEM_read_bio_X509_REQ fails, the error message must
16932 be obtained from the error queue.
16936 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16937 it in ERR_remove_state if appropriate, and change ERR_get_state
16938 accordingly to avoid race conditions (this is necessary because
16939 thread_hash is no longer constant once set).
16943 * Bugfix for linux-elf makefile.one.
16947 * RSA_get_default_method() will now cause a default
16948 RSA_METHOD to be chosen if one doesn't exist already.
16949 Previously this was only set during a call to RSA_new()
16950 or RSA_new_method(NULL) meaning it was possible for
16951 RSA_get_default_method() to return NULL.
16955 * Added native name translation to the existing DSO code
16956 that will convert (if the flag to do so is set) filenames
16957 that are sufficiently small and have no path information
16958 into a canonical native form. Eg. "blah" converted to
16959 "libblah.so" or "blah.dll" etc.
16963 * New function ERR_error_string_n(e, buf, len) which is like
16964 ERR_error_string(e, buf), but writes at most 'len' bytes
16965 including the 0 terminator. For ERR_error_string_n, 'buf'
16968 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16970 * CONF library reworked to become more general. A new CONF
16971 configuration file reader "class" is implemented as well as a
16972 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16973 old `CONF_*` functions are still there, but are reimplemented to
16974 work in terms of the new functions. Also, a set of functions
16975 to handle the internal storage of the configuration data is
16976 provided to make it easier to write new configuration file
16977 reader "classes" (I can definitely see something reading a
16978 configuration file in XML format, for example), called `_CONF_*`,
16979 or "the configuration storage API"...
16981 The new configuration file reading functions are:
16983 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16984 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16986 NCONF_default, NCONF_WIN32
16988 NCONF_dump_fp, NCONF_dump_bio
16990 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16991 NCONF_new creates a new CONF object. This works in the same way
16992 as other interfaces in OpenSSL, like the BIO interface.
16993 `NCONF_dump_*` dump the internal storage of the configuration file,
16994 which is useful for debugging. All other functions take the same
16995 arguments as the old `CONF_*` functions with the exception of the
16996 first that must be a `CONF *` instead of a `LHASH *`.
16998 To make it easier to use the new classes with the old `CONF_*` functions,
16999 the function CONF_set_default_method is provided.
17003 * Add '-tls1' option to 'openssl ciphers', which was already
17004 mentioned in the documentation but had not been implemented.
17005 (This option is not yet really useful because even the additional
17006 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17010 * Initial DSO code added into libcrypto for letting OpenSSL (and
17011 OpenSSL-based applications) load shared libraries and bind to
17012 them in a portable way.
17014 *Geoff Thorpe, with contributions from Richard Levitte*
17016 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17018 * Make sure _lrotl and _lrotr are only used with MSVC.
17020 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17021 (the default implementation of RAND_status).
17023 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17024 to '-clrext' (= clear extensions), as intended and documented.
17025 *Bodo Moeller; inconsistency pointed out by Michael Attili
17026 <attili@amaxo.com>*
17028 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17029 was larger than the MD block size.
17031 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17033 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17034 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17035 using the passed key: if the passed key was a private key the result
17036 of X509_print(), for example, would be to print out all the private key
17041 * des_quad_cksum() byte order bug fix.
17042 *Ulf Möller, using the problem description in krb4-0.9.7, where
17043 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17045 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17048 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17050 * For easily testing in shell scripts whether some command
17051 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17052 returns with exit code 0 iff no command of the given name is available.
17053 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17054 the output goes to stdout and nothing is printed to stderr.
17055 Additional arguments are always ignored.
17057 Since for each cipher there is a command of the same name,
17058 the 'no-cipher' compilation switches can be tested this way.
17060 ('openssl no-XXX' is not able to detect pseudo-commands such
17061 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17065 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17069 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17070 is set; it will be thrown away anyway because each handshake creates
17072 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17073 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17074 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17075 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17079 * New s_client option -ign_eof: EOF at stdin is ignored, and
17080 'Q' and 'R' lose their special meanings (quit/renegotiate).
17081 This is part of what -quiet does; unlike -quiet, -ign_eof
17082 does not suppress any output.
17086 * Add compatibility options to the purpose and trust code. The
17087 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17088 accepts a certificate or CA, this was the previous behaviour,
17089 with all the associated security issues.
17091 X509_TRUST_COMPAT is the old trust behaviour: only and
17092 automatically trust self signed roots in certificate store. A
17093 new trust setting X509_TRUST_DEFAULT is used to specify that
17094 a purpose has no associated trust setting and it should instead
17095 use the value in the default purpose.
17099 * Fix the PKCS#8 DSA private key code so it decodes keys again
17100 and fix a memory leak.
17104 * In util/mkerr.pl (which implements 'make errors'), preserve
17105 reason strings from the previous version of the .c file, as
17106 the default to have only downcase letters (and digits) in
17107 automatically generated reasons codes is not always appropriate.
17111 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17112 using strerror. Previously, ERR_reason_error_string() returned
17113 library names as reason strings for SYSerr; but SYSerr is a special
17114 case where small numbers are errno values, not library numbers.
17118 * Add '-dsaparam' option to 'openssl dhparam' application. This
17119 converts DSA parameters into DH parameters. (When creating parameters,
17120 DSA_generate_parameters is used.)
17124 * Include 'length' (recommended exponent length) in C code generated
17125 by 'openssl dhparam -C'.
17129 * The second argument to set_label in perlasm was already being used
17130 so couldn't be used as a "file scope" flag. Moved to third argument
17135 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17136 instead of RAND_bytes for encryption IVs and salts.
17140 * Include RAND_status() into RAND_METHOD instead of implementing
17141 it only for md_rand.c Otherwise replacing the PRNG by calling
17142 RAND_set_rand_method would be impossible.
17146 * Don't let DSA_generate_key() enter an infinite loop if the random
17147 number generation fails.
17151 * New 'rand' application for creating pseudo-random output.
17155 * Added configuration support for Linux/IA64
17157 *Rolf Haberrecker <rolf@suse.de>*
17159 * Assembler module support for Mingw32.
17163 * Shared library support for HPUX (in shlib/).
17165 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17167 * Shared library support for Solaris gcc.
17169 *Lutz Behnke <behnke@trustcenter.de>*
17171 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17173 * PKCS7_encrypt() was adding text MIME headers twice because they
17174 were added manually and by SMIME_crlf_copy().
17178 * In bntest.c don't call BN_rand with zero bits argument.
17180 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17182 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17183 case was implemented. This caused BN_div_recp() to fail occasionally.
17187 * Add an optional second argument to the set_label() in the perl
17188 assembly language builder. If this argument exists and is set
17189 to 1 it signals that the assembler should use a symbol whose
17190 scope is the entire file, not just the current function. This
17191 is needed with MASM which uses the format label:: for this scope.
17193 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17195 * Change the ASN1 types so they are typedefs by default. Before
17196 almost all types were #define'd to ASN1_STRING which was causing
17197 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17202 * Change names of new functions to the new get1/get0 naming
17203 convention: After 'get1', the caller owns a reference count
17204 and has to call `..._free`; 'get0' returns a pointer to some
17205 data structure without incrementing reference counters.
17206 (Some of the existing 'get' functions increment a reference
17207 counter, some don't.)
17208 Similarly, 'set1' and 'add1' functions increase reference
17209 counters or duplicate objects.
17213 * Allow for the possibility of temp RSA key generation failure:
17214 the code used to assume it always worked and crashed on failure.
17218 * Fix potential buffer overrun problem in BIO_printf().
17219 *Ulf Möller, using public domain code by Patrick Powell; problem
17220 pointed out by David Sacerdote <das33@cornell.edu>*
17222 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17223 RAND_egd() and RAND_status(). In the command line application,
17224 the EGD socket can be specified like a seed file using RANDFILE
17229 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17230 Some CAs (e.g. Verisign) distribute certificates in this form.
17234 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17235 list to exclude them. This means that no special compilation option
17236 is needed to use anonymous DH: it just needs to be included in the
17241 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17242 EVP_MD_type. The old functionality is available in a new macro called
17243 EVP_MD_md(). Change code that uses it and update docs.
17247 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17248 where the `void *` argument is replaced by a function pointer argument.
17249 Previously `void *` was abused to point to functions, which works on
17250 many platforms, but is not correct. As these functions are usually
17251 called by macros defined in OpenSSL header files, most source code
17252 should work without changes.
17256 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17257 sections with information on -D... compiler switches used for
17258 compiling the library so that applications can see them. To enable
17259 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17260 must be defined. E.g.,
17261 #define OPENSSL_ALGORITHM_DEFINES
17262 #include <openssl/opensslconf.h>
17263 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17265 *Richard Levitte, Ulf and Bodo Möller*
17267 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17272 * Change the 'other' type in certificate aux info to a STACK_OF
17273 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17274 the required ASN1 format: arbitrary types determined by an OID.
17278 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17279 argument to 'req'. This is not because the function is newer or
17280 better than others it just uses the work 'NEW' in the certificate
17281 request header lines. Some software needs this.
17285 * Reorganise password command line arguments: now passwords can be
17286 obtained from various sources. Delete the PEM_cb function and make
17287 it the default behaviour: i.e. if the callback is NULL and the
17288 usrdata argument is not NULL interpret it as a null terminated pass
17289 phrase. If usrdata and the callback are NULL then the pass phrase
17290 is prompted for as usual.
17294 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17295 the support is automatically enabled. The resulting binaries will
17296 autodetect the card and use it if present.
17298 *Ben Laurie and Compaq Inc.*
17300 * Work around for Netscape hang bug. This sends certificate request
17301 and server done in one record. Since this is perfectly legal in the
17302 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17303 the bugs/SSLv3 entry for more info.
17307 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17311 * Add -rand argument to smime and pkcs12 applications and read/write
17316 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17320 * Add command line password options to the remaining applications.
17324 * Bug fix for BN_div_recp() for numerators with an even number of
17329 * More tests in bntest.c, and changed test_bn output.
17333 * ./config recognizes MacOS X now.
17337 * Bug fix for BN_div() when the first words of num and divisor are
17338 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17342 * Add support for various broken PKCS#8 formats, and command line
17343 options to produce them.
17347 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17348 get temporary BIGNUMs from a BN_CTX.
17352 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17357 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17358 include a #define from the old name to the new. The original intent
17359 was that statically linked binaries could for example just call
17360 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17361 link with digests. This never worked because SSLeay_add_all_digests()
17362 and SSLeay_add_all_ciphers() were in the same source file so calling
17363 one would link with the other. They are now in separate source files.
17367 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17371 * Use a less unusual form of the Miller-Rabin primality test (it used
17372 a binary algorithm for exponentiation integrated into the Miller-Rabin
17373 loop, our standard modexp algorithms are faster).
17377 * Support for the EBCDIC character set completed.
17379 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17381 * Source code cleanups: use const where appropriate, eliminate casts,
17382 use `void *` instead of `char *` in lhash.
17386 * Bugfix: ssl3_send_server_key_exchange was not restartable
17387 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17388 this the server could overwrite ephemeral keys that the client
17393 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17394 using 50 iterations of the Rabin-Miller test.
17396 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17397 iterations of the Rabin-Miller test as required by the appendix
17398 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17399 As BN_is_prime_fasttest includes trial division, DSA parameter
17400 generation becomes much faster.
17402 This implies a change for the callback functions in DSA_is_prime
17403 and DSA_generate_parameters: The callback function is called once
17404 for each positive witness in the Rabin-Miller test, not just
17405 occasionally in the inner loop; and the parameters to the
17406 callback function now provide an iteration count for the outer
17407 loop rather than for the current invocation of the inner loop.
17408 DSA_generate_parameters additionally can call the callback
17409 function with an 'iteration count' of -1, meaning that a
17410 candidate has passed the trial division test (when q is generated
17411 from an application-provided seed, trial division is skipped).
17415 * New function BN_is_prime_fasttest that optionally does trial
17416 division before starting the Rabin-Miller test and has
17417 an additional BN_CTX * argument (whereas BN_is_prime always
17418 has to allocate at least one BN_CTX).
17419 'callback(1, -1, cb_arg)' is called when a number has passed the
17420 trial division stage.
17424 * Fix for bug in CRL encoding. The validity dates weren't being handled
17429 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17433 * New function BN_pseudo_rand().
17437 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17438 bignum version of BN_from_montgomery() with the working code from
17439 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17444 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17445 made it impossible to use the same SSL_SESSION data structure in
17446 SSL2 clients in multiple threads.
17450 * The return value of RAND_load_file() no longer counts bytes obtained
17451 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17452 to seed the PRNG (previously an explicit byte count was required).
17454 *Ulf Möller, Bodo Möller*
17456 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17457 used `char *` instead of `void *` and had casts all over the place.
17461 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17465 * Retain source code compatibility for BN_prime_checks macro:
17466 BN_is_prime(..., BN_prime_checks, ...) now uses
17467 BN_prime_checks_for_size to determine the appropriate number of
17468 Rabin-Miller iterations.
17472 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17473 DH_CHECK_P_NOT_SAFE_PRIME.
17474 (Check if this is true? OpenPGP calls them "strong".)
17478 * Merge the functionality of "dh" and "gendh" programs into a new program
17479 "dhparam". The old programs are retained for now but will handle DH keys
17480 (instead of parameters) in future.
17484 * Make the ciphers, s_server and s_client programs check the return values
17485 when a new cipher list is set.
17489 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17490 ciphers. Before when the 56bit ciphers were enabled the sorting was
17493 The syntax for the cipher sorting has been extended to support sorting by
17494 cipher-strength (using the strength_bits hard coded in the tables).
17495 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17497 Fix a bug in the cipher-command parser: when supplying a cipher command
17498 string with an "undefined" symbol (neither command nor alphanumeric
17499 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17500 an error is flagged.
17502 Due to the strength-sorting extension, the code of the
17503 ssl_create_cipher_list() function was completely rearranged. I hope that
17504 the readability was also increased :-)
17506 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17508 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17509 for the first serial number and places 2 in the serial number file. This
17510 avoids problems when the root CA is created with serial number zero and
17511 the first user certificate has the same issuer name and serial number
17516 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17517 the new code. Add documentation for this stuff.
17521 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17522 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17523 structures and behave in an analogous way to the X509v3 functions:
17524 they shouldn't be called directly but wrapper functions should be used
17527 So we also now have some wrapper functions that call the X509at functions
17528 when passed certificate requests. (TO DO: similar things can be done with
17529 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17530 things. Some of these need some d2i or i2d and print functionality
17531 because they handle more complex structures.)
17535 * Add missing #ifndefs that caused missing symbols when building libssl
17536 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17537 NO_RSA in `ssl/s2*.c`.
17539 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17541 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17542 has a return value which indicates the quality of the random data
17543 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17544 error queue. New function RAND_pseudo_bytes() generates output that is
17545 guaranteed to be unique but not unpredictable. RAND_add is like
17546 RAND_seed, but takes an extra argument for an entropy estimate
17547 (RAND_seed always assumes full entropy).
17551 * Do more iterations of Rabin-Miller probable prime test (specifically,
17552 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17553 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17554 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17555 false-positive rate of at most 2^-80 for random input.
17559 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17563 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17564 in the 0.9.5 release), this returns the chain
17565 from an X509_CTX structure with a dup of the stack and all
17566 the X509 reference counts upped: so the stack will exist
17567 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17570 Also make SSL_SESSION_print() print out the verify return
17575 * Add manpage for the pkcs12 command. Also change the default
17576 behaviour so MAC iteration counts are used unless the new
17577 -nomaciter option is used. This improves file security and
17578 only older versions of MSIE (4.0 for example) need it.
17582 * Honor the no-xxx Configure options when creating .DEF files.
17586 * Add PKCS#10 attributes to field table: challengePassword,
17587 unstructuredName and unstructuredAddress. These are taken from
17588 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17589 international characters are used.
17591 More changes to X509_ATTRIBUTE code: allow the setting of types
17592 based on strings. Remove the 'loc' parameter when adding
17593 attributes because these will be a SET OF encoding which is sorted
17598 * Initial changes to the 'req' utility to allow request generation
17599 automation. This will allow an application to just generate a template
17600 file containing all the field values and have req construct the
17603 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17604 used all over the place including certificate requests and PKCS#7
17605 structures. They are currently handled manually where necessary with
17606 some primitive wrappers for PKCS#7. The new functions behave in a
17607 manner analogous to the X509 extension functions: they allow
17608 attributes to be looked up by NID and added.
17610 Later something similar to the X509V3 code would be desirable to
17611 automatically handle the encoding, decoding and printing of the
17612 more complex types. The string types like challengePassword can
17613 be handled by the string table functions.
17615 Also modified the multi byte string table handling. Now there is
17616 a 'global mask' which masks out certain types. The table itself
17617 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17618 is useful when for example there is only one permissible type
17619 (as in countryName) and using the mask might result in no valid
17624 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17625 SSL_get_peer_finished to allow applications to obtain the latest
17626 Finished messages sent to the peer or expected from the peer,
17627 respectively. (SSL_get_peer_finished is usually the Finished message
17628 actually received from the peer, otherwise the protocol will be aborted.)
17630 As the Finished message are message digests of the complete handshake
17631 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17632 be used for external authentication procedures when the authentication
17633 provided by SSL/TLS is not desired or is not enough.
17637 * Enhanced support for Alpha Linux is added. Now ./config checks if
17638 the host supports BWX extension and if Compaq C is present on the
17639 $PATH. Just exploiting of the BWX extension results in 20-30%
17640 performance kick for some algorithms, e.g. DES and RC4 to mention
17641 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17646 * Add support for MS "fast SGC". This is arguably a violation of the
17647 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17648 weak crypto and after checking the certificate is SGC a second one
17649 with strong crypto. MS SGC stops the first handshake after receiving
17650 the server certificate message and sends a second client hello. Since
17651 a server will typically do all the time consuming operations before
17652 expecting any further messages from the client (server key exchange
17653 is the most expensive) there is little difference between the two.
17655 To get OpenSSL to support MS SGC we have to permit a second client
17656 hello message after we have sent server done. In addition we have to
17657 reset the MAC if we do get this second client hello.
17661 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17662 if a DER encoded private key is RSA or DSA traditional format. Changed
17663 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17664 format DER encoded private key. Newer code should use PKCS#8 format which
17665 has the key type encoded in the ASN1 structure. Added DER private key
17666 support to pkcs8 application.
17670 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17671 ciphersuites has been selected (as required by the SSL 3/TLS 1
17672 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17673 is set, we interpret this as a request to violate the specification
17674 (the worst that can happen is a handshake failure, and 'correct'
17675 behaviour would result in a handshake failure anyway).
17679 * In SSL_CTX_add_session, take into account that there might be multiple
17680 SSL_SESSION structures with the same session ID (e.g. when two threads
17681 concurrently obtain them from an external cache).
17682 The internal cache can handle only one SSL_SESSION with a given ID,
17683 so if there's a conflict, we now throw out the old one to achieve
17688 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17689 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17690 some routines that use cipher OIDs: some ciphers do not have OIDs
17691 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17696 * Simplify the trust setting structure and code. Now we just have
17697 two sequences of OIDs for trusted and rejected settings. These will
17698 typically have values the same as the extended key usage extension
17699 and any application specific purposes.
17701 The trust checking code now has a default behaviour: it will just
17702 check for an object with the same NID as the passed id. Functions can
17703 be provided to override either the default behaviour or the behaviour
17704 for a given id. SSL client, server and email already have functions
17705 in place for compatibility: they check the NID and also return "trusted"
17706 if the certificate is self signed.
17710 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17711 traditional format into an EVP_PKEY structure.
17715 * Add a password callback function PEM_cb() which either prompts for
17716 a password if usr_data is NULL or otherwise assumes it is a null
17717 terminated password. Allow passwords to be passed on command line
17718 environment or config files in a few more utilities.
17722 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17723 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17724 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17725 Update documentation.
17729 * Support for ASN1 "NULL" type. This could be handled before by using
17730 ASN1_TYPE but there wasn't any function that would try to read a NULL
17731 and produce an error if it couldn't. For compatibility we also have
17732 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17733 don't allocate anything because they don't need to.
17737 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17740 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17742 * Rebuild of the memory allocation routines used by OpenSSL code and
17743 possibly others as well. The purpose is to make an interface that
17744 provide hooks so anyone can build a separate set of allocation and
17745 deallocation routines to be used by OpenSSL, for example memory
17746 pool implementations, or something else, which was previously hard
17747 since Malloc(), Realloc() and Free() were defined as macros having
17748 the values malloc, realloc and free, respectively (except for Win32
17749 compilations). The same is provided for memory debugging code.
17750 OpenSSL already comes with functionality to find memory leaks, but
17751 this gives people a chance to debug other memory problems.
17753 With these changes, a new set of functions and macros have appeared:
17755 CRYPTO_set_mem_debug_functions() [F]
17756 CRYPTO_get_mem_debug_functions() [F]
17757 CRYPTO_dbg_set_options() [F]
17758 CRYPTO_dbg_get_options() [F]
17759 CRYPTO_malloc_debug_init() [M]
17761 The memory debug functions are NULL by default, unless the library
17762 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17763 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17764 gives the standard debugging functions that come with OpenSSL) or
17765 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17766 provided by the library user) must be used. When the standard
17767 debugging functions are used, CRYPTO_dbg_set_options can be used to
17768 request additional information:
17769 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17770 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17772 Also, things like CRYPTO_set_mem_functions will always give the
17773 expected result (the new set of functions is used for allocation
17774 and deallocation) at all times, regardless of platform and compiler
17777 To finish it up, some functions that were never use in any other
17778 way than through macros have a new API and new semantic:
17780 CRYPTO_dbg_malloc()
17781 CRYPTO_dbg_realloc()
17784 All macros of value have retained their old syntax.
17786 *Richard Levitte and Bodo Moeller*
17788 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17789 ordering of SMIMECapabilities wasn't in "strength order" and there
17790 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17795 * Some ASN1 types with illegal zero length encoding (INTEGER,
17796 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17798 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17800 * Merge in my S/MIME library for OpenSSL. This provides a simple
17801 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17802 functionality to handle multipart/signed properly) and a utility
17803 called 'smime' to call all this stuff. This is based on code I
17804 originally wrote for Celo who have kindly allowed it to be
17805 included in OpenSSL.
17809 * Add variants des_set_key_checked and des_set_key_unchecked of
17810 des_set_key (aka des_key_sched). Global variable des_check_key
17811 decides which of these is called by des_set_key; this way
17812 des_check_key behaves as it always did, but applications and
17813 the library itself, which was buggy for des_check_key == 1,
17814 have a cleaner way to pick the version they need.
17818 * New function PKCS12_newpass() which changes the password of a
17823 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17824 dynamic mix. In both cases the ids can be used as an index into the
17825 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17826 functions so they accept a list of the field values and the
17827 application doesn't need to directly manipulate the X509_TRUST
17832 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17837 * Modify the way the V3 extension code looks up extensions. This now
17838 works in a similar way to the object code: we have some "standard"
17839 extensions in a static table which is searched with OBJ_bsearch()
17840 and the application can add dynamic ones if needed. The file
17841 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17842 updated whenever a new extension is added to the core code and kept
17843 in ext_nid order. There is a simple program 'tabtest.c' which checks
17844 this. New extensions are not added too often so this file can readily
17845 be maintained manually.
17847 There are two big advantages in doing things this way. The extensions
17848 can be looked up immediately and no longer need to be "added" using
17849 X509V3_add_standard_extensions(): this function now does nothing.
17850 Side note: I get *lots* of email saying the extension code doesn't
17851 work because people forget to call this function.
17852 Also no dynamic allocation is done unless new extensions are added:
17853 so if we don't add custom extensions there is no need to call
17854 X509V3_EXT_cleanup().
17858 * Modify enc utility's salting as follows: make salting the default. Add a
17859 magic header, so unsalted files fail gracefully instead of just decrypting
17860 to garbage. This is because not salting is a big security hole, so people
17861 should be discouraged from doing it.
17865 * Fixes and enhancements to the 'x509' utility. It allowed a message
17866 digest to be passed on the command line but it only used this
17867 parameter when signing a certificate. Modified so all relevant
17868 operations are affected by the digest parameter including the
17869 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17870 DSA key was used because it didn't fix the digest.
17874 * Initial certificate chain verify code. Currently tests the untrusted
17875 certificates for consistency with the verify purpose (which is set
17876 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17878 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17879 this is because it will reject chains with invalid extensions whereas
17880 every previous version of OpenSSL and SSLeay made no checks at all.
17882 Trust code: checks the root CA for the relevant trust settings. Trust
17883 settings have an initial value consistent with the verify purpose: e.g.
17884 if the verify purpose is for SSL client use it expects the CA to be
17885 trusted for SSL client use. However the default value can be changed to
17886 permit custom trust settings: one example of this would be to only trust
17887 certificates from a specific "secure" set of CAs.
17889 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17890 which should be used for version portability: especially since the
17891 verify structure is likely to change more often now.
17893 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17894 to set them. If not set then assume SSL clients will verify SSL servers
17897 Two new options to the verify program: -untrusted allows a set of
17898 untrusted certificates to be passed in and -purpose which sets the
17899 intended purpose of the certificate. If a purpose is set then the
17900 new chain verify code is used to check extension consistency.
17904 * Support for the authority information access extension.
17908 * Modify RSA and DSA PEM read routines to transparently handle
17909 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17910 public keys in a format compatible with certificate
17911 SubjectPublicKeyInfo structures. Unfortunately there were already
17912 functions called *_PublicKey_* which used various odd formats so
17913 these are retained for compatibility: however the DSA variants were
17914 never in a public release so they have been deleted. Changed dsa/rsa
17915 utilities to handle the new format: note no releases ever handled public
17916 keys so we should be OK.
17918 The primary motivation for this change is to avoid the same fiasco
17919 that dogs private keys: there are several incompatible private key
17920 formats some of which are standard and some OpenSSL specific and
17921 require various evil hacks to allow partial transparent handling and
17922 even then it doesn't work with DER formats. Given the option anything
17923 other than PKCS#8 should be dumped: but the other formats have to
17924 stay in the name of compatibility.
17926 With public keys and the benefit of hindsight one standard format
17927 is used which works with EVP_PKEY, RSA or DSA structures: though
17928 it clearly returns an error if you try to read the wrong kind of key.
17930 Added a -pubkey option to the 'x509' utility to output the public key.
17931 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17932 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17933 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17934 that do the same as the `EVP_PKEY_assign_*()` except they up the
17935 reference count of the added key (they don't "swallow" the
17940 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17941 CRLs would fail if the file contained no certificates or no CRLs:
17942 added a new function to read in both types and return the number
17943 read: this means that if none are read it will be an error. The
17944 DER versions of the certificate and CRL reader would always fail
17945 because it isn't possible to mix certificates and CRLs in DER format
17946 without choking one or the other routine. Changed this to just read
17947 a certificate: this is the best we can do. Also modified the code
17948 in `apps/verify.c` to take notice of return codes: it was previously
17949 attempting to read in certificates from NULL pointers and ignoring
17950 any errors: this is one reason why the cert and CRL reader seemed
17951 to work. It doesn't check return codes from the default certificate
17952 routines: these may well fail if the certificates aren't installed.
17956 * Code to support otherName option in GeneralName.
17960 * First update to verify code. Change the verify utility
17961 so it warns if it is passed a self signed certificate:
17962 for consistency with the normal behaviour. X509_verify
17963 has been modified to it will now verify a self signed
17964 certificate if *exactly* the same certificate appears
17965 in the store: it was previously impossible to trust a
17966 single self signed certificate. This means that:
17967 openssl verify ss.pem
17968 now gives a warning about a self signed certificate but
17969 openssl verify -CAfile ss.pem ss.pem
17974 * For servers, store verify_result in SSL_SESSION data structure
17975 (and add it to external session representation).
17976 This is needed when client certificate verifications fails,
17977 but an application-provided verification callback (set by
17978 SSL_CTX_set_cert_verify_callback) allows accepting the session
17979 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17980 but returns 1): When the session is reused, we have to set
17981 ssl->verify_result to the appropriate error code to avoid
17984 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17986 * Fix a bug in the new PKCS#7 code: it didn't consider the
17987 case in PKCS7_dataInit() where the signed PKCS7 structure
17988 didn't contain any existing data because it was being created.
17990 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17992 * Add a salt to the key derivation routines in enc.c. This
17993 forms the first 8 bytes of the encrypted file. Also add a
17994 -S option to allow a salt to be input on the command line.
17998 * New function X509_cmp(). Oddly enough there wasn't a function
17999 to compare two certificates. We do this by working out the SHA1
18000 hash and comparing that. X509_cmp() will be needed by the trust
18005 * SSL_get1_session() is like SSL_get_session(), but increments
18006 the reference count in the SSL_SESSION returned.
18008 *Geoff Thorpe <geoff@eu.c2.net>*
18010 * Fix for 'req': it was adding a null to request attributes.
18011 Also change the X509_LOOKUP and X509_INFO code to handle
18012 certificate auxiliary information.
18016 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18021 * Add the possibility to add extra information to the memory leak
18022 detecting output, to form tracebacks, showing from where each
18023 allocation was originated: CRYPTO_push_info("constant string") adds
18024 the string plus current file name and line number to a per-thread
18025 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18026 is like calling CYRPTO_pop_info() until the stack is empty.
18027 Also updated memory leak detection code to be multi-thread-safe.
18031 * Add options -text and -noout to pkcs7 utility and delete the
18032 encryption options which never did anything. Update docs.
18036 * Add options to some of the utilities to allow the pass phrase
18037 to be included on either the command line (not recommended on
18038 OSes like Unix) or read from the environment. Update the
18039 manpages and fix a few bugs.
18043 * Add a few manpages for some of the openssl commands.
18047 * Fix the -revoke option in ca. It was freeing up memory twice,
18048 leaking and not finding already revoked certificates.
18052 * Extensive changes to support certificate auxiliary information.
18053 This involves the use of X509_CERT_AUX structure and X509_AUX
18054 functions. An X509_AUX function such as PEM_read_X509_AUX()
18055 can still read in a certificate file in the usual way but it
18056 will also read in any additional "auxiliary information". By
18057 doing things this way a fair degree of compatibility can be
18058 retained: existing certificates can have this information added
18059 using the new 'x509' options.
18061 Current auxiliary information includes an "alias" and some trust
18062 settings. The trust settings will ultimately be used in enhanced
18063 certificate chain verification routines: currently a certificate
18064 can only be trusted if it is self signed and then it is trusted
18069 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18070 The problem was that one of the replacement routines had not been working
18071 since SSLeay releases. For now the offending routine has been replaced
18072 with non-optimised assembler. Even so, this now gives around 95%
18073 performance improvement for 1024 bit RSA signs.
18077 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18078 handling. Most clients have the effective key size in bits equal to
18079 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18080 A few however don't do this and instead use the size of the decrypted key
18081 to determine the RC2 key length and the AlgorithmIdentifier to determine
18082 the effective key length. In this case the effective key length can still
18083 be 40 bits but the key length can be 168 bits for example. This is fixed
18084 by manually forcing an RC2 key into the EVP_PKEY structure because the
18085 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18086 the key length and effective key length are equal.
18090 * Add a bunch of functions that should simplify the creation of
18091 X509_NAME structures. Now you should be able to do:
18092 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18093 and have it automatically work out the correct field type and fill in
18094 the structures. The more adventurous can try:
18095 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18096 and it will (hopefully) work out the correct multibyte encoding.
18100 * Change the 'req' utility to use the new field handling and multibyte
18101 copy routines. Before the DN field creation was handled in an ad hoc
18102 way in req, ca, and x509 which was rather broken and didn't support
18103 BMPStrings or UTF8Strings. Since some software doesn't implement
18104 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18105 using the dirstring_type option. See the new comment in the default
18106 openssl.cnf for more info.
18110 * Make crypto/rand/md_rand.c more robust:
18111 - Assure unique random numbers after fork().
18112 - Make sure that concurrent threads access the global counter and
18113 md serializably so that we never lose entropy in them
18114 or use exactly the same state in multiple threads.
18115 Access to the large state is not always serializable because
18116 the additional locking could be a performance killer, and
18117 md should be large enough anyway.
18121 * New file `apps/app_rand.c` with commonly needed functionality
18122 for handling the random seed file.
18124 Use the random seed file in some applications that previously did not:
18126 dsaparam -genkey (which also ignored its '-rand' option),
18129 x509 (when signing).
18130 Except on systems with /dev/urandom, it is crucial to have a random
18131 seed file at least for key creation, DSA signing, and for DH exchanges;
18132 for RSA signatures we could do without one.
18134 gendh and gendsa (unlike genrsa) used to read only the first byte
18135 of each file listed in the '-rand' option. The function as previously
18136 found in genrsa is now in app_rand.c and is used by all programs
18137 that support '-rand'.
18141 * In RAND_write_file, use mode 0600 for creating files;
18142 don't just chmod when it may be too late.
18146 * Report an error from X509_STORE_load_locations
18147 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18151 * New function ASN1_mbstring_copy() this copies a string in either
18152 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18153 into an ASN1_STRING type. A mask of permissible types is passed
18154 and it chooses the "minimal" type to use or an error if not type
18159 * Add function equivalents to the various macros in asn1.h. The old
18160 macros are retained with an `M_` prefix. Code inside the library can
18161 use the `M_` macros. External code (including the openssl utility)
18162 should *NOT* in order to be "shared library friendly".
18166 * Add various functions that can check a certificate's extensions
18167 to see if it usable for various purposes such as SSL client,
18168 server or S/MIME and CAs of these types. This is currently
18169 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18170 verification. Also added a -purpose flag to x509 utility to
18171 print out all the purposes.
18175 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18180 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18181 for, obtain and decode and extension and obtain its critical flag.
18182 This allows all the necessary extension code to be handled in a
18183 single function call.
18187 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18188 platforms. See crypto/rc4/rc4_enc.c for further details.
18192 * New -noout option to asn1parse. This causes no output to be produced
18193 its main use is when combined with -strparse and -out to extract data
18194 from a file (which may not be in ASN.1 format).
18198 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18199 when producing the local key id.
18201 *Richard Levitte <levitte@stacken.kth.se>*
18203 * New option -dhparam in s_server. This allows a DH parameter file to be
18204 stated explicitly. If it is not stated then it tries the first server
18205 certificate file. The previous behaviour hard coded the filename
18210 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18211 a public key to be input or output. For example:
18212 openssl rsa -in key.pem -pubout -out pubkey.pem
18213 Also added necessary DSA public key functions to handle this.
18217 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18218 in the message. This was handled by allowing
18219 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18221 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18223 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18224 to the end of the strings whereas this didn't. This would cause problems
18225 if strings read with d2i_ASN1_bytes() were later modified.
18227 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18229 * Fix for base64 decode bug. When a base64 bio reads only one line of
18230 data and it contains EOF it will end up returning an error. This is
18231 caused by input 46 bytes long. The cause is due to the way base64
18232 BIOs find the start of base64 encoded data. They do this by trying a
18233 trial decode on each line until they find one that works. When they
18234 do a flag is set and it starts again knowing it can pass all the
18235 data directly through the decoder. Unfortunately it doesn't reset
18236 the context it uses. This means that if EOF is reached an attempt
18237 is made to pass two EOFs through the context and this causes the
18238 resulting error. This can also cause other problems as well. As is
18239 usual with these problems it takes *ages* to find and the fix is
18240 trivial: move one line.
18242 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18244 * Ugly workaround to get s_client and s_server working under Windows. The
18245 old code wouldn't work because it needed to select() on sockets and the
18246 tty (for keypresses and to see if data could be written). Win32 only
18247 supports select() on sockets so we select() with a 1s timeout on the
18248 sockets and then see if any characters are waiting to be read, if none
18249 are present then we retry, we also assume we can always write data to
18250 the tty. This isn't nice because the code then blocks until we've
18251 received a complete line of data and it is effectively polling the
18252 keyboard at 1s intervals: however it's quite a bit better than not
18253 working at all :-) A dedicated Windows application might handle this
18254 with an event loop for example.
18258 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18259 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18260 will be called when RSA_sign() and RSA_verify() are used. This is useful
18261 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18262 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18263 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18264 This necessitated the support of an extra signature type NID_md5_sha1
18265 for SSL signatures and modifications to the SSL library to use it instead
18266 of calling RSA_public_decrypt() and RSA_private_encrypt().
18270 * Add new -verify -CAfile and -CApath options to the crl program, these
18271 will lookup a CRL issuers certificate and verify the signature in a
18272 similar way to the verify program. Tidy up the crl program so it
18273 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18274 less strict. It will now permit CRL extensions even if it is not
18275 a V2 CRL: this will allow it to tolerate some broken CRLs.
18279 * Initialize all non-automatic variables each time one of the openssl
18280 sub-programs is started (this is necessary as they may be started
18281 multiple times from the "OpenSSL>" prompt).
18283 *Lennart Bang, Bodo Moeller*
18285 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18286 removing all other RSA functionality (this is what NO_RSA does). This
18287 is so (for example) those in the US can disable those operations covered
18288 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18293 * Non-copying interface to BIO pairs.
18294 (still largely untested)
18298 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18299 ASCII string. This was handled independently in various places before.
18303 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18304 UTF8 strings a character at a time.
18308 * Use client_version from client hello to select the protocol
18309 (s23_srvr.c) and for RSA client key exchange verification
18310 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18314 * Add various utility functions to handle SPKACs, these were previously
18315 handled by poking round in the structure internals. Added new function
18316 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18317 print, verify and generate SPKACs. Based on an original idea from
18318 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18322 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18326 * Allow the config file extension section to be overwritten on the
18327 command line. Based on an original idea from Massimiliano Pala
18328 <madwolf@comune.modena.it>. The new option is called -extensions
18329 and can be applied to ca, req and x509. Also -reqexts to override
18330 the request extensions in req and -crlexts to override the crl extensions
18335 * Add new feature to the SPKAC handling in ca. Now you can include
18336 the same field multiple times by preceding it by "XXXX." for example:
18339 this is the same syntax as used in the req config file.
18343 * Allow certificate extensions to be added to certificate requests. These
18344 are specified in a 'req_extensions' option of the req section of the
18345 config file. They can be printed out with the -text option to req but
18346 are otherwise ignored at present.
18350 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18351 data read consists of only the final block it would not decrypted because
18352 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18353 A misplaced 'break' also meant the decrypted final block might not be
18354 copied until the next read.
18358 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18359 a few extra parameters to the DH structure: these will be useful if
18360 for example we want the value of 'q' or implement X9.42 DH.
18364 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18365 provides hooks that allow the default DSA functions or functions on a
18366 "per key" basis to be replaced. This allows hardware acceleration and
18367 hardware key storage to be handled without major modification to the
18368 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18369 associated functions.
18373 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18374 as "read only": it can't be written to and the buffer it points to will
18375 not be freed. Reading from a read only BIO is much more efficient than
18376 a normal memory BIO. This was added because there are several times when
18377 an area of memory needs to be read from a BIO. The previous method was
18378 to create a memory BIO and write the data to it, this results in two
18379 copies of the data and an O(n^2) reading algorithm. There is a new
18380 function BIO_new_mem_buf() which creates a read only memory BIO from
18381 an area of memory. Also modified the PKCS#7 routines to use read only
18386 * Bugfix: ssl23_get_client_hello did not work properly when called in
18387 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18388 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18389 but a retry condition occurred while trying to read the rest.
18393 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18394 NID_pkcs7_encrypted by default: this was wrong since this should almost
18395 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18396 the encrypted data type: this is a more sensible place to put it and it
18397 allows the PKCS#12 code to be tidied up that duplicated this
18402 * Changed obj_dat.pl script so it takes its input and output files on
18403 the command line. This should avoid shell escape redirection problems
18408 * Initial support for certificate extension requests, these are included
18409 in things like Xenroll certificate requests. Included functions to allow
18410 extensions to be obtained and added.
18414 * -crlf option to s_client and s_server for sending newlines as
18415 CRLF (as required by many protocols).
18419 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18421 * Install libRSAglue.a when OpenSSL is built with RSAref.
18423 *Ralf S. Engelschall*
18425 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18427 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18429 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18434 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18435 DH parameters/keys (q is lost during that conversion, but the resulting
18436 DH parameters contain its length).
18438 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18439 much faster than DH_generate_parameters (which creates parameters
18440 where `p = 2*q + 1`), and also the smaller q makes DH computations
18441 much more efficient (160-bit exponentiation instead of 1024-bit
18442 exponentiation); so this provides a convenient way to support DHE
18443 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18444 utter importance to use
18445 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18447 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18448 when such DH parameters are used, because otherwise small subgroup
18449 attacks may become possible!
18453 * Avoid memory leak in i2d_DHparams.
18457 * Allow the -k option to be used more than once in the enc program:
18458 this allows the same encrypted message to be read by multiple recipients.
18462 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18463 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18464 it will always use the numerical form of the OID, even if it has a short
18469 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18470 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18471 otherwise bn_mod_exp was called. In the case of hardware keys for example
18472 no private key components need be present and it might store extra data
18473 in the RSA structure, which cannot be accessed from bn_mod_exp.
18474 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18475 private key operations.
18479 * Added support for SPARC Linux.
18483 * pem_password_cb function type incompatibly changed from
18484 typedef int pem_password_cb(char *buf, int size, int rwflag);
18486 ....(char *buf, int size, int rwflag, void *userdata);
18487 so that applications can pass data to their callbacks:
18488 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18489 additional void * argument, which is just handed through whenever
18490 the password callback is called.
18492 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18494 New function SSL_CTX_set_default_passwd_cb_userdata.
18496 Compatibility note: As many C implementations push function arguments
18497 onto the stack in reverse order, the new library version is likely to
18498 interoperate with programs that have been compiled with the old
18499 pem_password_cb definition (PEM_whatever takes some data that
18500 happens to be on the stack as its last argument, and the callback
18501 just ignores this garbage); but there is no guarantee whatsoever that
18504 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18505 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18506 problems not only on Windows, but also on some Unix platforms.
18507 To avoid problematic command lines, these definitions are now in an
18508 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18509 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18513 * MIPS III/IV assembler module is reimplemented.
18517 * More DES library cleanups: remove references to srand/rand and
18518 delete an unused file.
18522 * Add support for the free Netwide assembler (NASM) under Win32,
18523 since not many people have MASM (ml) and it can be hard to obtain.
18524 This is currently experimental but it seems to work OK and pass all
18525 the tests. Check out INSTALL.W32 for info.
18529 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18530 without temporary keys kept an extra copy of the server key,
18531 and connections with temporary keys did not free everything in case
18536 * New function RSA_check_key and new openssl rsa option -check
18537 for verifying the consistency of RSA keys.
18539 *Ulf Moeller, Bodo Moeller*
18541 * Various changes to make Win32 compile work:
18542 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18543 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18544 comparison" warnings.
18545 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18549 * Add a debugging option to PKCS#5 v2 key generation function: when
18550 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18551 derived keys are printed to stderr.
18555 * Copy the flags in ASN1_STRING_dup().
18557 *Roman E. Pavlov <pre@mo.msk.ru>*
18559 * The x509 application mishandled signing requests containing DSA
18560 keys when the signing key was also DSA and the parameters didn't match.
18562 It was supposed to omit the parameters when they matched the signing key:
18563 the verifying software was then supposed to automatically use the CA's
18564 parameters if they were absent from the end user certificate.
18566 Omitting parameters is no longer recommended. The test was also
18567 the wrong way round! This was probably due to unusual behaviour in
18568 EVP_cmp_parameters() which returns 1 if the parameters match.
18569 This meant that parameters were omitted when they *didn't* match and
18570 the certificate was useless. Certificates signed with 'ca' didn't have
18573 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18575 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18576 The interface is as follows:
18577 Applications can use
18578 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18579 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18580 "off" is now the default.
18581 The library internally uses
18582 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18583 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18584 to disable memory-checking temporarily.
18586 Some inconsistent states that previously were possible (and were
18587 even the default) are now avoided.
18589 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18590 with each memory chunk allocated; this is occasionally more helpful
18591 than just having a counter.
18593 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18595 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18600 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18601 which largely parallels "options", but is for changing API behaviour,
18602 whereas "options" are about protocol behaviour.
18603 Initial "mode" flags are:
18605 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18606 a single record has been written.
18607 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18608 retries use the same buffer location.
18609 (But all of the contents must be
18614 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18617 * Fix problems with no-hmac etc.
18619 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18621 * New functions RSA_get_default_method(), RSA_set_method() and
18622 RSA_get_method(). These allows replacement of RSA_METHODs without having
18623 to mess around with the internals of an RSA structure.
18627 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18628 Also really enable memory leak checks in openssl.c and in some
18631 *Chad C. Mulligan, Bodo Moeller*
18633 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18634 up the length of negative integers. This has now been simplified to just
18635 store the length when it is first determined and use it later, rather
18636 than trying to keep track of where data is copied and updating it to
18638 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18640 * Add a new function PKCS7_signatureVerify. This allows the verification
18641 of a PKCS#7 signature but with the signing certificate passed to the
18642 function itself. This contrasts with PKCS7_dataVerify which assumes the
18643 certificate is present in the PKCS#7 structure. This isn't always the
18644 case: certificates can be omitted from a PKCS#7 structure and be
18645 distributed by "out of band" means (such as a certificate database).
18649 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18650 function prototypes in pem.h, also change util/mkdef.pl to add the
18651 necessary function names.
18655 * mk1mf.pl (used by Windows builds) did not properly read the
18656 options set by Configure in the top level Makefile, and Configure
18657 was not even able to write more than one option correctly.
18658 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18662 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18663 file to be loaded from a BIO or FILE pointer. The BIO version will
18664 for example allow memory BIOs to contain config info.
18668 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18669 Whoever hopes to achieve shared-library compatibility across versions
18670 must use this, not the compile-time macro.
18671 (Exercise 0.9.4: Which is the minimum library version required by
18673 Note: All this applies only to multi-threaded programs, others don't
18678 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18679 through a BIO pair triggered the default case, i.e.
18680 SSLerr(...,SSL_R_UNKNOWN_STATE).
18684 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18685 can use the SSL library even if none of the specific BIOs is
18690 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18691 for the encoded length.
18693 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18695 * Add initial documentation of the X509V3 functions.
18699 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18700 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18701 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18702 secure PKCS#8 private key format with a high iteration count.
18706 * Fix determination of Perl interpreter: A perl or perl5
18707 *directory* in $PATH was also accepted as the interpreter.
18709 *Ralf S. Engelschall*
18711 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18712 wrong with it but it was very old and did things like calling
18713 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18714 unusual formatting.
18718 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18719 to use the new extension code.
18723 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18724 with macros. This should make it easier to change their form, add extra
18725 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18730 * Add to configuration table a new entry that can specify an alternative
18731 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18732 according to Mark Crispin <MRC@Panda.COM>.
18736 * DES CBC did not update the IV. Weird.
18740 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18741 Changing the behaviour of the former might break existing programs --
18742 where IV updating is needed, des_ncbc_encrypt can be used.
18745 * When bntest is run from "make test" it drives bc to check its
18746 calculations, as well as internally checking them. If an internal check
18747 fails, it needs to cause bc to give a non-zero result or make test carries
18748 on without noticing the failure. Fixed.
18752 * DES library cleanups.
18756 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18757 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18758 ciphers. NOTE: although the key derivation function has been verified
18759 against some published test vectors it has not been extensively tested
18760 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18765 * Instead of "mkdir -p", which is not fully portable, use new
18766 Perl script "util/mkdir-p.pl".
18770 * Rewrite the way password based encryption (PBE) is handled. It used to
18771 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18772 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18773 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18774 the 'parameter' field of the AlgorithmIdentifier is passed to the
18775 underlying key generation function so it must do its own ASN1 parsing.
18776 This has also changed the EVP_PBE_CipherInit() function which now has a
18777 'parameter' argument instead of literal salt and iteration count values
18778 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18782 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18783 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18784 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18785 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18786 value was just used as a "magic string" and not used directly its
18787 value doesn't matter.
18791 * Introduce some semblance of const correctness to BN. Shame C doesn't
18796 * "linux-sparc64" configuration (ultrapenguin).
18798 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18799 "linux-sparc" configuration.
18801 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18803 * config now generates no-xxx options for missing ciphers.
18807 * Support the EBCDIC character set (work in progress).
18808 File ebcdic.c not yet included because it has a different license.
18810 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18812 * Support BS2000/OSD-POSIX.
18814 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18816 * Make callbacks for key generation use `void *` instead of `char *`.
18820 * Make S/MIME samples compile (not yet tested).
18824 * Additional typesafe stacks.
18828 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18832 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18834 * New configuration variant "sco5-gcc".
18836 * Updated some demos.
18838 *Sean O Riordain, Wade Scholine*
18840 * Add missing BIO_free at exit of pkcs12 application.
18844 * Fix memory leak in conf.c.
18848 * Updates for Win32 to assembler version of MD5.
18852 * Set #! path to perl in `apps/der_chop` to where we found it
18853 instead of using a fixed path.
18857 * SHA library changes for irix64-mips4-cc.
18861 * Improvements for VMS support.
18865 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18867 * Bignum library bug fix. IRIX 6 passes "make test" now!
18868 This also avoids the problems with SC4.2 and unpatched SC5.
18870 *Andy Polyakov <appro@fy.chalmers.se>*
18872 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18873 These are required because of the typesafe stack would otherwise break
18874 existing code. If old code used a structure member which used to be STACK
18875 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18876 sk_num or sk_value it would produce an error because the num, data members
18877 are not present in STACK_OF. Now it just produces a warning. sk_set
18878 replaces the old method of assigning a value to sk_value
18879 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18880 that does this will no longer work (and should use sk_set instead) but
18881 this could be regarded as a "questionable" behaviour anyway.
18885 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18886 correctly handle encrypted S/MIME data.
18890 * Change type of various DES function arguments from des_cblock
18891 (which means, in function argument declarations, pointer to char)
18892 to des_cblock * (meaning pointer to array with 8 char elements),
18893 which allows the compiler to do more typechecking; it was like
18894 that back in SSLeay, but with lots of ugly casts.
18896 Introduce new type const_des_cblock.
18900 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18901 problems: find RecipientInfo structure that matches recipient certificate
18902 and initialise the ASN1 structures properly based on passed cipher.
18906 * Belatedly make the BN tests actually check the results.
18910 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18911 to and from BNs: it was completely broken. New compilation option
18912 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18913 key elements as negative integers.
18917 * Reorganize and speed up MD5.
18919 *Andy Polyakov <appro@fy.chalmers.se>*
18923 *Richard Levitte <richard@levitte.org>*
18925 * New option -out to asn1parse to allow the parsed structure to be
18926 output to a file. This is most useful when combined with the -strparse
18927 option to examine the output of things like OCTET STRINGS.
18931 * Make SSL library a little more fool-proof by not requiring any longer
18932 that `SSL_set_{accept,connect}_state` be called before
18933 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18934 in many applications because usually everything *appeared* to work as
18935 intended anyway -- now it really works as intended).
18939 * Move openssl.cnf out of lib/.
18943 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18944 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18945 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18947 *Ralf S. Engelschall*
18949 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18950 handle PKCS#7 enveloped data properly.
18952 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18954 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18955 copying pointers. The cert_st handling is changed by this in
18956 various ways (and thus what used to be known as ctx->default_cert
18957 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18958 any longer when s->cert does not give us what we need).
18959 ssl_cert_instantiate becomes obsolete by this change.
18960 As soon as we've got the new code right (possibly it already is?),
18961 we have solved a couple of bugs of the earlier code where s->cert
18962 was used as if it could not have been shared with other SSL structures.
18964 Note that using the SSL API in certain dirty ways now will result
18965 in different behaviour than observed with earlier library versions:
18966 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18967 does not influence s as it used to.
18969 In order to clean up things more thoroughly, inside SSL_SESSION
18970 we don't use CERT any longer, but a new structure SESS_CERT
18971 that holds per-session data (if available); currently, this is
18972 the peer's certificate chain and, for clients, the server's certificate
18973 and temporary key. CERT holds only those values that can have
18974 meaningful defaults in an SSL_CTX.
18978 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18979 from the internal representation. Various PKCS#7 fixes: remove some
18980 evil casts and set the enc_dig_alg field properly based on the signing
18985 * Allow PKCS#12 password to be set from the command line or the
18986 environment. Let 'ca' get its config file name from the environment
18987 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18992 * Allow certificate policies extension to use an IA5STRING for the
18993 organization field. This is contrary to the PKIX definition but
18994 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18999 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19000 without disallowing inline assembler and the like for non-pedantic builds.
19004 * Support Borland C++ builder.
19006 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19012 * SHA-1 cleanups and performance enhancements.
19014 *Andy Polyakov <appro@fy.chalmers.se>*
19016 * Sparc v8plus assembler for the bignum library.
19018 *Andy Polyakov <appro@fy.chalmers.se>*
19020 * Accept any -xxx and +xxx compiler options in Configure.
19024 * Update HPUX configuration.
19028 * Add missing `sk_<type>_unshift()` function to safestack.h
19030 *Ralf S. Engelschall*
19032 * New function SSL_CTX_use_certificate_chain_file that sets the
19033 "extra_cert"s in addition to the certificate. (This makes sense
19034 only for "PEM" format files, as chains as a whole are not
19039 * Support verify_depth from the SSL API.
19040 x509_vfy.c had what can be considered an off-by-one-error:
19041 Its depth (which was not part of the external interface)
19042 was actually counting the number of certificates in a chain;
19043 now it really counts the depth.
19047 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19048 instead of X509err, which often resulted in confusing error
19049 messages since the error codes are not globally unique
19050 (e.g. an alleged error in ssl3_accept when a certificate
19051 didn't match the private key).
19053 * New function SSL_CTX_set_session_id_context that allows to set a default
19054 value (so that you don't need SSL_set_session_id_context for each
19055 connection using the SSL_CTX).
19059 * OAEP decoding bug fix.
19063 * Support INSTALL_PREFIX for package builders, as proposed by
19068 * New Configure options "threads" and "no-threads". For systems
19069 where the proper compiler options are known (currently Solaris
19070 and Linux), "threads" is the default.
19074 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19078 * Install various scripts to $(OPENSSLDIR)/misc, not to
19079 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19080 such as /usr/local/bin.
19084 * "make linux-shared" to build shared libraries.
19086 *Niels Poppe <niels@netbox.org>*
19088 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19092 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19093 extension adding in x509 utility.
19097 * Remove NOPROTO sections and error code comments.
19101 * Partial rewrite of the DEF file generator to now parse the ANSI
19106 * New Configure options --prefix=DIR and --openssldir=DIR.
19110 * Complete rewrite of the error code script(s). It is all now handled
19111 by one script at the top level which handles error code gathering,
19112 header rewriting and C source file generation. It should be much better
19113 than the old method: it now uses a modified version of Ulf's parser to
19114 read the ANSI prototypes in all header files (thus the old K&R definitions
19115 aren't needed for error creation any more) and do a better job of
19116 translating function codes into names. The old 'ASN1 error code embedded
19117 in a comment' is no longer necessary and it doesn't use .err files which
19118 have now been deleted. Also the error code call doesn't have to appear all
19119 on one line (which resulted in some large lines...).
19123 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19127 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19128 0 (which usually indicates a closed connection), but continue reading.
19132 * Fix some race conditions.
19136 * Add support for CRL distribution points extension. Add Certificate
19137 Policies and CRL distribution points documentation.
19141 * Move the autogenerated header file parts to crypto/opensslconf.h.
19145 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19146 8 of keying material. Merlin has also confirmed interop with this fix
19147 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19149 *Merlin Hughes <merlin@baltimore.ie>*
19151 * Fix lots of warnings.
19153 *Richard Levitte <levitte@stacken.kth.se>*
19155 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19156 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19158 *Richard Levitte <levitte@stacken.kth.se>*
19160 * Fix problems with sizeof(long) == 8.
19162 *Andy Polyakov <appro@fy.chalmers.se>*
19164 * Change functions to ANSI C.
19168 * Fix typos in error codes.
19170 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19172 * Remove defunct assembler files from Configure.
19176 * SPARC v8 assembler BIGNUM implementation.
19178 *Andy Polyakov <appro@fy.chalmers.se>*
19180 * Support for Certificate Policies extension: both print and set.
19181 Various additions to support the r2i method this uses.
19185 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19186 return a const string when you are expecting an allocated buffer.
19190 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19191 types DirectoryString and DisplayText.
19195 * Add code to allow r2i extensions to access the configuration database,
19196 add an LHASH database driver and add several ctx helper functions.
19200 * Fix an evil bug in bn_expand2() which caused various BN functions to
19201 fail when they extended the size of a BIGNUM.
19205 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19206 support typesafe stack.
19210 * Fix typo in SSL_[gs]et_options().
19212 *Nils Frostberg <nils@medcom.se>*
19214 * Delete various functions and files that belonged to the (now obsolete)
19215 old X509V3 handling code.
19219 * New Configure option "rsaref".
19223 * Don't auto-generate pem.h.
19227 * Introduce type-safe ASN.1 SETs.
19231 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19233 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19235 * Introduce type-safe STACKs. This will almost certainly break lots of code
19236 that links with OpenSSL (well at least cause lots of warnings), but fear
19237 not: the conversion is trivial, and it eliminates loads of evil casts. A
19238 few STACKed things have been converted already. Feel free to convert more.
19239 In the fullness of time, I'll do away with the STACK type altogether.
19243 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19244 specified in `<certfile>` by updating the entry in the index.txt file.
19245 This way one no longer has to edit the index.txt file manually for
19246 revoking a certificate. The -revoke option does the gory details now.
19248 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19250 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19251 `-text` option at all and this way the `-noout -text` combination was
19252 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19254 *Ralf S. Engelschall*
19256 * Make sure a corresponding plain text error message exists for the
19257 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19258 verify callback function determined that a certificate was revoked.
19260 *Ralf S. Engelschall*
19262 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19263 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19264 all available ciphers including rc5, which was forgotten until now.
19265 In order to let the testing shell script know which algorithms
19266 are available, a new (up to now undocumented) command
19267 `openssl list-cipher-commands` is used.
19271 * Bugfix: s_client occasionally would sleep in select() when
19272 it should have checked SSL_pending() first.
19276 * New functions DSA_do_sign and DSA_do_verify to provide access to
19277 the raw DSA values prior to ASN.1 encoding.
19281 * Tweaks to Configure
19283 *Niels Poppe <niels@netbox.org>*
19285 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19290 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19294 * New config option to avoid instructions that are illegal on the 80386.
19295 The default code is faster, but requires at least a 486.
19299 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19300 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19301 same as SSL2_VERSION anyway.
19305 * New "-showcerts" option for s_client.
19309 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19310 application. Various cleanups and fixes.
19314 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19315 modify error routines to work internally. Add error codes and PBE init
19316 to library startup routines.
19320 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19321 packing functions to asn1 and evp. Changed function names and error
19322 codes along the way.
19326 * PKCS12 integration: and so it begins... First of several patches to
19327 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19328 objects to objects.h
19332 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19333 and display support for Thawte strong extranet extension.
19337 * Add LinuxPPC support.
19339 *Jeff Dubrule <igor@pobox.org>*
19341 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19342 bn_div_words in alpha.s.
19344 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19346 * Make sure the RSA OAEP test is skipped under -DRSAref because
19347 OAEP isn't supported when OpenSSL is built with RSAref.
19349 *Ulf Moeller <ulf@fitug.de>*
19351 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19352 so they no longer are missing under -DNOPROTO.
19354 *Soren S. Jorvang <soren@t.dk>*
19356 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19358 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19359 doesn't work when the session is reused. Coming soon!
19363 * Fix a security hole, that allows sessions to be reused in the wrong
19364 context thus bypassing client cert protection! All software that uses
19365 client certs and session caches in multiple contexts NEEDS PATCHING to
19366 allow session reuse! A fuller solution is in the works.
19368 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19370 * Some more source tree cleanups (removed obsolete files
19371 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19372 permission on "config" script to be executable) and a fix for the INSTALL
19375 *Ulf Moeller <ulf@fitug.de>*
19377 * Remove some legacy and erroneous uses of malloc, free instead of
19380 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19382 * Make rsa_oaep_test return non-zero on error.
19384 *Ulf Moeller <ulf@fitug.de>*
19386 * Add support for native Solaris shared libraries. Configure
19387 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19388 if someone would make that last step automatic.
19390 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19392 * ctx_size was not built with the right compiler during "make links". Fixed.
19396 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19397 except NULL ciphers". This means the default cipher list will no longer
19398 enable NULL ciphers. They need to be specifically enabled e.g. with
19399 the string "DEFAULT:eNULL".
19403 * Fix to RSA private encryption routines: if p < q then it would
19404 occasionally produce an invalid result. This will only happen with
19405 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19409 * Be less restrictive and allow also `perl util/perlpath.pl
19410 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19411 because this way one can also use an interpreter named `perl5` (which is
19412 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19413 installed as `perl`).
19415 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19417 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19419 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19421 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19422 advapi32.lib to Win32 build and change the pem test comparison
19423 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19424 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19425 and crypto/des/ede_cbcm_enc.c.
19429 * DES quad checksum was broken on big-endian architectures. Fixed.
19433 * Comment out two functions in bio.h that aren't implemented. Fix up the
19434 Win32 test batch file so it (might) work again. The Win32 test batch file
19435 is horrible: I feel ill....
19439 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19440 in e_os.h. Audit of header files to check ANSI and non ANSI
19441 sections: 10 functions were absent from non ANSI section and not exported
19442 from Windows DLLs. Fixed up libeay.num for new functions.
19446 * Make `openssl version` output lines consistent.
19448 *Ralf S. Engelschall*
19450 * Fix Win32 symbol export lists for BIO functions: Added
19451 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19452 to ms/libeay{16,32}.def.
19454 *Ralf S. Engelschall*
19456 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19457 fine under Unix and passes some trivial tests I've now added. But the
19458 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19459 added to make sure no one expects that this stuff really works in the
19460 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19461 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19464 *Ralf S. Engelschall*
19466 * Fix the generation of two part addresses in perl.
19468 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19470 * Add config entry for Linux on MIPS.
19472 *John Tobey <jtobey@channel1.com>*
19474 * Make links whenever Configure is run, unless we are on Windoze.
19478 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19479 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19484 * Add a useful kludge to allow package maintainers to specify compiler and
19485 other platforms details on the command line without having to patch the
19486 Configure script every time: One now can use
19487 `perl Configure <id>:<details>`,
19488 i.e. platform ids are allowed to have details appended
19489 to them (separated by colons). This is treated as there would be a static
19490 pre-configured entry in Configure's %table under key `<id>` with value
19491 `<details>` and `perl Configure <id>` is called. So, when you want to
19492 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19493 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19494 now, which overrides the FreeBSD-elf entry on-the-fly.
19496 *Ralf S. Engelschall*
19498 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19502 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19503 on the `perl Configure ...` command line. This way one can compile
19504 OpenSSL libraries with Position Independent Code (PIC) which is needed
19505 for linking it into DSOs.
19507 *Ralf S. Engelschall*
19509 * Remarkably, export ciphers were totally broken and no-one had noticed!
19514 * Cleaned up the LICENSE document: The official contact for any license
19515 questions now is the OpenSSL core team under openssl-core@openssl.org.
19516 And add a paragraph about the dual-license situation to make sure people
19517 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19518 to the OpenSSL toolkit.
19520 *Ralf S. Engelschall*
19522 * General source tree makefile cleanups: Made `making xxx in yyy...`
19523 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19524 Additionally cleaned up the `make links` target: Remove unnecessary
19525 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19526 to speed processing and no longer clutter the display with confusing
19527 stuff. Instead only the actually done links are displayed.
19529 *Ralf S. Engelschall*
19531 * Permit null encryption ciphersuites, used for authentication only. It used
19532 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19533 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19538 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19539 signed attributes when verifying signatures (this would break them),
19540 the detached data encoding was wrong and public keys obtained using
19541 X509_get_pubkey() weren't freed.
19545 * Add text documentation for the BUFFER functions. Also added a work around
19546 to a Win95 console bug. This was triggered by the password read stuff: the
19547 last character typed gets carried over to the next fread(). If you were
19548 generating a new cert request using 'req' for example then the last
19549 character of the passphrase would be CR which would then enter the first
19554 * Added the new 'Includes OpenSSL Cryptography Software' button as
19555 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19556 button and can be used by applications based on OpenSSL to show the
19557 relationship to the OpenSSL project.
19559 *Ralf S. Engelschall*
19561 * Remove confusing variables in function signatures in files
19562 ssl/ssl_lib.c and ssl/ssl.h.
19564 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19566 * Don't install bss_file.c under PREFIX/include/
19568 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19570 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19571 functions that return function pointers and has support for NT specific
19572 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19573 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19574 unsigned to signed types: this was killing the Win32 compile.
19578 * Add new certificate file to stack functions,
19579 SSL_add_dir_cert_subjects_to_stack() and
19580 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19581 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19582 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19583 This means that Apache-SSL and similar packages don't have to mess around
19584 to add as many CAs as they want to the preferred list.
19588 * Experiment with doxygen documentation. Currently only partially applied to
19590 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19591 openssl.doxy as the configuration file.
19595 * Get rid of remaining C++-style comments which strict C compilers hate.
19597 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19599 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19600 compiled in by default: it has problems with large keys.
19604 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19605 DH private keys and/or callback functions which directly correspond to
19606 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19607 is needed for applications which have to configure certificates on a
19608 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19610 For the RSA certificate situation is makes no difference, but
19611 for the DSA certificate situation this fixes the "no shared cipher"
19612 problem where the OpenSSL cipher selection procedure failed because the
19613 temporary keys were not overtaken from the context and the API provided
19614 no way to reconfigure them.
19615 The new functions now let applications reconfigure the stuff and they
19616 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19617 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19618 non-public-API function ssl_cert_instantiate() is used as a helper
19619 function and also to reduce code redundancy inside ssl_rsa.c.
19621 *Ralf S. Engelschall*
19623 * Move s_server -dcert and -dkey options out of the undocumented feature
19624 area because they are useful for the DSA situation and should be
19625 recognized by the users.
19627 *Ralf S. Engelschall*
19629 * Fix the cipher decision scheme for export ciphers: the export bits are
19630 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19631 SSL_EXP_MASK. So, the original variable has to be used instead of the
19632 already masked variable.
19634 *Richard Levitte <levitte@stacken.kth.se>*
19636 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19638 *Richard Levitte <levitte@stacken.kth.se>*
19640 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19641 from `int` to `unsigned int` because it is a length and initialized by
19642 EVP_DigestFinal() which expects an `unsigned int *`.
19644 *Richard Levitte <levitte@stacken.kth.se>*
19646 * Don't hard-code path to Perl interpreter on shebang line of Configure
19647 script. Instead use the usual Shell->Perl transition trick.
19649 *Ralf S. Engelschall*
19651 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19652 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19653 -noout -modulus` as it's already the case for `openssl rsa -noout
19654 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19655 currently the public key is printed (a decision which was already done by
19656 `openssl dsa -modulus` in the past) which serves a similar purpose.
19657 Additionally the NO_RSA no longer completely removes the whole -modulus
19658 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19661 *Ralf S. Engelschall*
19663 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19664 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19666 *Arne Ansper <arne@ats.cyber.ee>*
19668 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19669 to be added. Now both 'req' and 'ca' can use new objects defined in the
19674 * Add cool BIO that does syslog (or event log on NT).
19676 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19678 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19679 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19680 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19681 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19685 * Add preliminary config info for new extension code.
19689 * Make RSA_NO_PADDING really use no padding.
19691 *Ulf Moeller <ulf@fitug.de>*
19693 * Generate errors when private/public key check is done.
19697 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19698 for some CRL extensions and new objects added.
19702 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19703 key usage extension and fuller support for authority key id.
19707 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19708 padding method for RSA, which is recommended for new applications in PKCS
19709 #1 v2.0 (RFC 2437, October 1998).
19710 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19711 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19712 against Bleichbacher's attack on RSA.
19713 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19716 * Updates to the new SSL compression code
19718 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19720 * Fix so that the version number in the master secret, when passed
19721 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19722 (because the server will not accept higher), that the version number
19723 is 0x03,0x01, not 0x03,0x00
19725 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19727 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19728 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19729 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19733 * Support for RAW extensions where an arbitrary extension can be
19734 created by including its DER encoding. See `apps/openssl.cnf` for
19739 * Make sure latest Perl versions don't interpret some generated C array
19740 code as Perl array code in the crypto/err/err_genc.pl script.
19742 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19744 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19745 not many people have the assembler. Various Win32 compilation fixes and
19746 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19747 build instructions.
19751 * Modify configure script 'Configure' to automatically create crypto/date.h
19752 file under Win32 and also build pem.h from pem.org. New script
19753 util/mkfiles.pl to create the MINFO file on environments that can't do a
19754 'make files': perl util/mkfiles.pl >MINFO should work.
19758 * Major rework of DES function declarations, in the pursuit of correctness
19759 and purity. As a result, many evil casts evaporated, and some weirdness,
19760 too. You may find this causes warnings in your code. Zapping your evil
19761 casts will probably fix them. Mostly.
19765 * Fix for a typo in asn1.h. Bug fix to object creation script
19766 obj_dat.pl. It considered a zero in an object definition to mean
19767 "end of object": none of the objects in objects.h have any zeros
19768 so it wasn't spotted.
19770 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19772 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19773 Masking (CBCM). In the absence of test vectors, the best I have been able
19774 to do is check that the decrypt undoes the encrypt, so far. Send me test
19775 vectors if you have them.
19779 * Correct calculation of key length for export ciphers (too much space was
19780 allocated for null ciphers). This has not been tested!
19784 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19785 message is now correct (it understands "crypto" and "ssl" on its
19786 command line). There is also now an "update" option. This will update
19787 the util/ssleay.num and util/libeay.num files with any new functions.
19789 perl util/mkdef.pl crypto ssl update
19790 it will update them.
19794 * Overhauled the Perl interface:
19795 - ported BN stuff to OpenSSL's different BN library
19796 - made the perl/ source tree CVS-aware
19797 - renamed the package from SSLeay to OpenSSL (the files still contain
19798 their history because I've copied them in the repository)
19799 - removed obsolete files (the test scripts will be replaced
19800 by better Test::Harness variants in the future)
19802 *Ralf S. Engelschall*
19804 * First cut for a very conservative source tree cleanup:
19805 1. merge various obsolete readme texts into doc/ssleay.txt
19806 where we collect the old documents and readme texts.
19807 2. remove the first part of files where I'm already sure that we no
19808 longer need them because of three reasons: either they are just temporary
19809 files which were left by Eric or they are preserved original files where
19810 I've verified that the diff is also available in the CVS via "cvs diff
19811 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19812 the crypto/md/ stuff).
19814 *Ralf S. Engelschall*
19816 * More extension code. Incomplete support for subject and issuer alt
19817 name, issuer and authority key id. Change the i2v function parameters
19818 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19819 what that's for :-) Fix to ASN1 macro which messed up
19820 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19824 * Preliminary support for ENUMERATED type. This is largely copied from the
19829 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19831 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19833 * Make sure `make rehash` target really finds the `openssl` program.
19835 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19837 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19838 like to hear about it if this slows down other processors.
19842 * Add CygWin32 platform information to Configure script.
19844 *Alan Batie <batie@aahz.jf.intel.com>*
19846 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19848 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19850 * New program nseq to manipulate netscape certificate sequences
19854 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19859 * Fixes to BN code. Previously the default was to define BN_RECURSION
19860 but the BN code had some problems that would cause failures when
19861 doing certificate verification and some other functions.
19863 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19865 * Add ASN1 and PEM code to support netscape certificate sequences.
19869 * Add ASN1 and PEM code to support netscape certificate sequences.
19873 * Add several PKIX and private extended key usage OIDs.
19877 * Modify the 'ca' program to handle the new extension code. Modify
19878 openssl.cnf for new extension format, add comments.
19882 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19883 and add a sample to openssl.cnf so req -x509 now adds appropriate
19888 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19889 error code, add initial support to X509_print() and x509 application.
19893 * Takes a deep breath and start adding X509 V3 extension support code. Add
19894 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19895 stuff is currently isolated and isn't even compiled yet.
19899 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19900 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19901 Removed the versions check from X509 routines when loading extensions:
19902 this allows certain broken certificates that don't set the version
19903 properly to be processed.
19907 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19908 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19909 can still be regenerated with "make depend".
19913 * Spelling mistake in C version of CAST-128.
19915 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19917 * Changes to the error generation code. The perl script err-code.pl
19918 now reads in the old error codes and retains the old numbers, only
19919 adding new ones if necessary. It also only changes the .err files if new
19920 codes are added. The makefiles have been modified to only insert errors
19921 when needed (to avoid needlessly modifying header files). This is done
19922 by only inserting errors if the .err file is newer than the auto generated
19923 C file. To rebuild all the error codes from scratch (the old behaviour)
19924 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19925 or delete all the .err files.
19929 * CAST-128 was incorrectly implemented for short keys. The C version has
19930 been fixed, but is untested. The assembler versions are also fixed, but
19931 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19932 to regenerate it if needed.
19933 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19934 Hagino <itojun@kame.net>*
19936 * File was opened incorrectly in randfile.c.
19938 *Ulf Möller <ulf@fitug.de>*
19940 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19941 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19942 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19943 al: it's just almost always a UTCTime. Note this patch adds new error
19944 codes so do a "make errors" if there are problems.
19948 * Correct Linux 1 recognition in config.
19950 *Ulf Möller <ulf@fitug.de>*
19952 * Remove pointless MD5 hash when using DSA keys in ca.
19954 *Anonymous <nobody@replay.com>*
19956 * Generate an error if given an empty string as a cert directory. Also
19957 generate an error if handed NULL (previously returned 0 to indicate an
19958 error, but didn't set one).
19960 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19962 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19966 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19967 parameters. This was causing a warning which killed off the Win32 compile.
19971 * Remove C++ style comments from crypto/bn/bn_local.h.
19973 *Neil Costigan <neil.costigan@celocom.com>*
19975 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19976 based on a text string, looking up short and long names and finally
19977 "dot" format. The "dot" format stuff didn't work. Added new function
19978 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19979 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19980 OID is not part of the table.
19984 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19985 X509_LOOKUP_by_alias().
19989 * Sort openssl functions by name.
19993 * Get the `gendsa` command working and add it to the `list` command. Remove
19994 encryption from sample DSA keys (in case anyone is interested the password
19999 * Make *all* `*_free` functions accept a NULL pointer.
20001 *Frans Heymans <fheymans@isaserver.be>*
20003 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20006 *Anonymous <nobody@replay.com>*
20008 * s_server should send the CAfile as acceptable CAs, not its own cert.
20010 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20012 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20014 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20016 * Temp key "for export" tests were wrong in s3_srvr.c.
20018 *Anonymous <nobody@replay.com>*
20020 * Add prototype for temp key callback functions
20021 SSL_CTX_set_tmp_{rsa,dh}_callback().
20025 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20026 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20030 * X509_name_add_entry() freed the wrong thing after an error.
20032 *Arne Ansper <arne@ats.cyber.ee>*
20034 * rsa_eay.c would attempt to free a NULL context.
20036 *Arne Ansper <arne@ats.cyber.ee>*
20038 * BIO_s_socket() had a broken should_retry() on Windoze.
20040 *Arne Ansper <arne@ats.cyber.ee>*
20042 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20044 *Arne Ansper <arne@ats.cyber.ee>*
20046 * Make sure the already existing X509_STORE->depth variable is initialized
20047 in X509_STORE_new(), but document the fact that this variable is still
20048 unused in the certificate verification process.
20050 *Ralf S. Engelschall*
20052 * Fix the various library and `apps/` files to free up pkeys obtained from
20053 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20057 * Fix reference counting in X509_PUBKEY_get(). This makes
20058 demos/maurice/example2.c work, amongst others, probably.
20060 *Steve Henson and Ben Laurie*
20062 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20063 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20064 are no longer created. This way we have a single and consistent command
20065 line interface `openssl <command>`, similar to `cvs <command>`.
20067 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20069 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20070 BIT STRING wrapper always have zero unused bits.
20074 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20078 * Make the top-level INSTALL documentation easier to understand.
20082 * Makefiles updated to exit if an error occurs in a sub-directory
20083 make (including if user presses ^C) [Paul Sutton]
20085 * Make Montgomery context stuff explicit in RSA data structure.
20089 * Fix build order of pem and err to allow for generated pem.h.
20093 * Fix renumbering bug in X509_NAME_delete_entry().
20097 * Enhanced the err-ins.pl script so it makes the error library number
20098 global and can add a library name. This is needed for external ASN1 and
20099 other error libraries.
20103 * Fixed sk_insert which never worked properly.
20107 * Fix ASN1 macros so they can handle indefinite length constructed
20108 EXPLICIT tags. Some non standard certificates use these: they can now
20113 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20114 into a single doc/ssleay.txt bundle. This way the information is still
20115 preserved but no longer messes up this directory. Now it's new room for
20116 the new set of documentation files.
20118 *Ralf S. Engelschall*
20120 * SETs were incorrectly DER encoded. This was a major pain, because they
20121 shared code with SEQUENCEs, which aren't coded the same. This means that
20122 almost everything to do with SETs or SEQUENCEs has either changed name or
20123 number of arguments.
20125 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20127 * Fix test data to work with the above.
20131 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20132 was already fixed by Eric for 0.9.1 it seems.
20134 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20136 * Autodetect FreeBSD3.
20140 * Fix various bugs in Configure. This affects the following platforms:
20144 unixware-2.0-pentium
20149 * Eliminate generated files from CVS. Reorder tests to regenerate files
20150 before they are needed.
20154 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20158 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20160 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20161 changed SSLeay to OpenSSL in version strings.
20163 *Ralf S. Engelschall*
20165 * Some fixups to the top-level documents.
20169 * Fixed the nasty bug where rsaref.h was not found under compile-time
20170 because the symlink to include/ was missing.
20172 *Ralf S. Engelschall*
20174 * Incorporated the popular no-RSA/DSA-only patches
20175 which allow to compile an RSA-free SSLeay.
20177 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20179 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20180 when "ssleay" is still not found.
20182 *Ralf S. Engelschall*
20184 * Added more platforms to Configure: Cray T3E, HPUX 11,
20186 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20188 * Updated the README file.
20190 *Ralf S. Engelschall*
20192 * Added various .cvsignore files in the CVS repository subdirs
20193 to make a "cvs update" really silent.
20195 *Ralf S. Engelschall*
20197 * Recompiled the error-definition header files and added
20198 missing symbols to the Win32 linker tables.
20200 *Ralf S. Engelschall*
20202 * Cleaned up the top-level documents;
20203 o new files: CHANGES and LICENSE
20204 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20205 o merged COPYRIGHT into LICENSE
20206 o removed obsolete TODO file
20207 o renamed MICROSOFT to INSTALL.W32
20209 *Ralf S. Engelschall*
20211 * Removed dummy files from the 0.9.1b source tree:
20212 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20213 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20214 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20215 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20216 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20218 *Ralf S. Engelschall*
20220 * Added various platform portability fixes.
20224 * The Genesis of the OpenSSL rpject:
20225 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20226 Young and Tim J. Hudson created while they were working for C2Net until
20229 *The OpenSSL Project*
20231 ### Changes between 0.9.0b and 0.9.1b [not released]
20233 * Updated a few CA certificates under certs/
20237 * Changed some BIGNUM api stuff.
20241 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20242 DGUX x86, Linux Alpha, etc.
20246 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20247 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20252 * Add -strparse option to asn1pars program which parses nested
20255 *Dr Stephen Henson <shenson@bigfoot.com>*
20257 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20261 * DSA fix for "ca" program.
20265 * Added "-genkey" option to "dsaparam" program.
20269 * Added RIPE MD160 (rmd160) message digest.
20273 * Added -a (all) option to "ssleay version" command.
20277 * Added PLATFORM define which is the id given to Configure.
20281 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20285 * Extended the ASN.1 parser routines.
20289 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20293 * Added a BN_CTX to the BN library.
20297 * Fixed the weak key values in DES library
20301 * Changed API in EVP library for cipher aliases.
20305 * Added support for RC2/64bit cipher.
20309 * Converted the lhash library to the crypto/mem.c functions.
20313 * Added more recognized ASN.1 object ids.
20317 * Added more RSA padding checks for SSL/TLS.
20321 * Added BIO proxy/filter functionality.
20325 * Added extra_certs to SSL_CTX which can be used
20326 send extra CA certificates to the client in the CA cert chain sending
20327 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20331 * Now Fortezza is denied in the authentication phase because
20332 this is key exchange mechanism is not supported by SSLeay at all.
20336 * Additional PKCS1 checks.
20340 * Support the string "TLSv1" for all TLS v1 ciphers.
20344 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20345 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20349 * Fixed a few memory leaks.
20353 * Fixed various code and comment typos.
20357 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20358 bytes sent in the client random.
20360 *Edward Bishop <ebishop@spyglass.com>*
20364 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20365 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20366 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20367 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20368 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20369 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20370 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20371 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20372 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20373 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20374 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20375 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20376 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20377 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20378 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20379 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20380 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20381 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20382 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20383 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20384 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20385 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20386 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20387 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20388 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20389 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20390 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20391 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20392 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20393 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20394 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20395 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20396 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20397 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20398 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20399 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20400 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20401 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20402 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20403 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20404 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20405 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20406 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20407 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20408 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20409 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20410 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20411 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20412 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20413 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20414 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20415 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20416 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20417 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20418 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20419 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20420 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20421 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20422 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20423 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20424 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20425 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20426 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20427 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20428 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20429 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20430 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20431 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20432 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20433 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20434 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20435 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20436 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20437 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20438 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20439 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20440 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20441 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20442 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20443 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20444 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20445 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20446 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20447 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20448 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20449 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20450 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20451 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20452 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20453 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20454 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20455 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20456 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20457 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20458 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20459 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20460 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20461 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20462 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20463 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20464 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20465 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20466 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20467 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20468 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20469 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20470 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20471 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20472 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20473 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20474 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20475 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20476 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20477 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20478 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20479 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20480 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20481 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20482 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20483 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20484 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20485 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20486 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20487 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20488 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20489 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20490 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20491 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20492 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20493 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20494 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20495 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20496 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20497 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20498 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20499 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20500 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20501 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20502 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20503 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20504 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20505 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20506 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20507 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20508 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20509 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20510 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20511 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20512 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20513 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20514 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20515 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20516 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20517 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20518 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20519 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20520 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20521 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20522 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20523 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20524 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20525 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20526 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20527 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20528 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20529 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20530 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20531 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20532 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20533 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20534 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20535 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20536 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20537 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20538 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20539 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20540 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20541 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20542 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20543 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20544 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20545 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20546 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20547 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20548 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655