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
46 - support for delayed delivery of all types of response messages
50 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
51 be less hard coded in the build file templates, and to allow easier
52 addition of more exporters. With that, an exporter for CMake is also
57 * The BLAKE2s hash algorithm matches BLAKE2b's support
58 for configurable output length.
62 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
63 server to prefer session resumption using PSK-only key exchange over PSK
64 with DHE, if both are available.
66 *Markus Minichmayr, Tapkey GmbH*
71 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
73 * The BLAKE2b hash algorithm supports a configurable output length
74 by setting the "size" parameter.
76 *Čestmír Kalina and Tomáš Mráz*
78 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
82 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
83 and the corresponding provider-storemgmt API function
84 OSSL_FUNC_store_delete().
88 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
89 a passphrase callback when opening a store.
93 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
94 from 8 bytes to 16 bytes.
95 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
96 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
97 requires a salt length of 128 bits. This affects OpenSSL command line
98 applications such as "genrsa" and "pkcs8" and API's such as
99 PEM_write_bio_PrivateKey() that are reliant on the default value.
100 The additional commandline option 'saltlen' has been added to the
101 OpenSSL command line applications for "pkcs8" and "enc" to allow the
102 salt length to be set to a non default value.
106 * Changed the default value of the `ess_cert_id_alg` configuration
107 option which is used to calculate the TSA's public key certificate
108 identifier. The default algorithm is updated to be sha256 instead
111 *Małgorzata Olszówka*
113 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
114 table for point multiplication of the base point, which increases the size of
115 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
116 been added to disable the precomputed table.
120 * Added client side support for QUIC
122 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
124 * Added multiple tutorials on the OpenSSL library and in particular
125 on writing various clients (using TLS and QUIC protocols) with libssl.
129 * Added secp384r1 implementation using Solinas' reduction to improve
130 speed of the NIST P-384 elliptic curve. To enable the implementation
131 the build option `enable-ec_nistp_64_gcc_128` must be used.
135 * Improved RFC7468 compliance of the asn1parse command.
137 *Matthias St. Pierre*
139 * Added SHA256/192 algorithm support.
143 * Added support for securely getting root CA certificate update in
148 * Improved contention on global write locks by using more read locks where
153 * Improved performance of OSSL_PARAM lookups in performance critical
158 * Added the SSL_get0_group_name() function to provide access to the
159 name of the group used for the TLS key exchange.
163 * Provide a new configure option `no-http` that can be used to disable the
164 HTTP support. Provide new configure options `no-apps` and `no-docs` to
165 disable building the openssl command line application and the documentation.
169 * Provide a new configure option `no-ecx` that can be used to disable the
170 X25519, X448, and EdDSA support.
174 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
175 the EVP_KDF_CTX_set_params() function they are now concatenated not just
176 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
180 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
181 the provider context as a parameter.
185 * TLS round-trip time calculation was added by a Brigham Young University
186 Capstone team partnering with Sandia National Laboratories. A new function
187 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
192 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
193 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
194 option. Use of the "advanced" s_client command command via the "-adv" option
199 * Added an "advanced" command mode to s_client. Use this with the "-adv"
200 option. The old "basic" command mode recognises certain letters that must
201 always appear at the start of a line and cannot be escaped. The advanced
202 command mode enables commands to be entered anywhere and there is an
203 escaping mechanism. After starting s_client with "-adv" type "{help}"
204 to show a list of available commands.
208 * Add Raw Public Key (RFC7250) support. Authentication is supported
209 by matching keys against either local policy (TLSA records synthesised
210 from the expected keys) or DANE (TLSA records obtained by the
211 application from DNS). TLSA records will also match the same key in
212 the server certificate, should RPK use not happen to be negotiated.
216 * Added support for modular exponentiation and CRT offloading for the
221 * Added further assembler code for the RISC-V architecture.
225 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
226 from a given EC_GROUP.
230 * Improved support for non-default library contexts and property queries
231 when parsing PKCS#12 files.
235 * Implemented support for all five instances of EdDSA from RFC8032:
236 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
237 The streaming is not yet supported for the HashEdDSA variants
238 (Ed25519ph and Ed448ph).
242 * Added SM4 optimization for ARM processors using ASIMD and AES HW
247 * Implemented SM4-XTS support.
251 * Added platform-agnostic OSSL_sleep() function.
255 * Implemented deterministic ECDSA signatures (RFC6979) support.
259 * Implemented AES-GCM-SIV (RFC8452) support.
263 * Added support for pluggable (provider-based) TLS signature algorithms.
264 This enables TLS 1.3 authentication operations with algorithms embedded
265 in providers not included by default in OpenSSL. In combination with
266 the already available pluggable KEM and X.509 support, this enables
267 for example suitable providers to deliver post-quantum or quantum-safe
268 cryptography to OpenSSL users.
272 * Added support for pluggable (provider-based) CMS signature algorithms.
273 This enables CMS sign and verify operations with algorithms embedded
274 in providers not included by default in OpenSSL.
278 * Added support for Hybrid Public Key Encryption (HPKE) as defined
279 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
280 Message Layer Security (MLS) and other IETF specifications.
281 HPKE can also be used by other applications that require
282 encrypting "to" an ECDH public key. External APIs are defined in
283 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
287 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
292 * Add support for certificate compression (RFC8879), including
293 library support for Brotli and Zstandard compression.
297 * Add the ability to add custom attributes to PKCS12 files. Add a new API
298 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
299 for a user specified callback and optional argument.
300 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
301 added to the existing STACK_OF attrs.
305 * Major refactor of the libssl record layer.
309 * Add a mac salt length option for the pkcs12 command.
313 * Add more SRTP protection profiles from RFC8723 and RFC8269.
317 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
319 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
321 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
322 supported and enabled.
326 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
327 to the list of ciphersuites providing Perfect Forward Secrecy as
328 required by SECLEVEL >= 3.
330 *Dmitry Belyavskiy, Nicola Tuveri*
332 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
333 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
334 SSL_get0_iana_groups() function-like macro, retrieves the list of
335 supported groups sent by the peer.
336 The function SSL_client_hello_get_extension_order() populates
337 a caller-supplied array with the list of extension types present in the
338 ClientHello, in order of appearance.
342 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
343 to make it possible to use empty passphrase strings.
347 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
351 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
352 to change functions used for allocating the memory of asynchronous call stack.
356 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
360 * A failure exit code is returned when using the openssl x509 command to check
361 certificate attributes and the checks fail.
365 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
366 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
367 of 160 bits and above and less than 224 bits were previously accepted by
368 default but are now no longer allowed. By default TLS compression was
369 already disabled in previous OpenSSL versions. At security level 2 it cannot
374 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
379 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
380 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
381 will need to load the legacy crypto provider.
385 * CCM8 cipher suites in TLS have been downgraded to security level zero
386 because they use a short authentication tag which lowers their strength.
390 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
395 * Add X.509 certificate codeSigning purpose and related checks on key usage and
396 extended key usage of the leaf certificate according to the CA/Browser Forum.
400 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
401 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
402 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
403 X.509 version 3 if the certificate information includes X.509 extensions.
407 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
408 such as adding a trace facility for debugging certificate chain building.
412 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
413 in particular supporting requests for central key generation, generalized
414 polling, and various types of genm/genp exchanges defined in CMP Updates.
418 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
419 like correcting the TLS and proxy support and adding tracing for debugging.
423 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
427 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
428 a certificate to be added is already present. `CMS_sign_ex()` and
429 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
430 and no longer throw an error for them.
434 * Fixed and extended `util/check-format.pl` for checking adherence to the
435 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
436 The checks are meanwhile more complete and yield fewer false positives.
440 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
441 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
442 calls. They can be used as the transport BIOs for QUIC.
444 *Hugo Landau, Matt Caswell and Tomáš Mráz*
446 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
447 sending and receiving multiple messages in a single call. An implementation
448 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
452 * Support for loading root certificates from the Windows certificate store
453 has been added. The support is in the form of a store which recognises the
454 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
455 arguments. This store is built by default and can be disabled using the new
456 compile-time option `no-winstore`. This store is not currently used by
457 default and must be loaded explicitly using the above store URI. It is
458 expected to be loaded by default in the future.
462 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
463 kernel versions that support KTLS have a known bug in CCM processing. That
464 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
465 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
470 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
475 * Zerocopy KTLS sendfile() support on Linux.
479 * The OBJ_ calls are now thread safe using a global lock.
483 * New parameter `-digest` for openssl cms command allowing signing
484 pre-computed digests and new CMS API functions supporting that
489 * OPENSSL_malloc() and other allocation functions now raise errors on
490 allocation failures. The callers do not need to explicitly raise errors
491 unless they want to for tracing purposes.
495 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
496 decryption as a protection against Bleichenbacher-like attacks.
497 The RSA decryption API will now return a randomly generated deterministic
498 message instead of an error in case it detects an error when checking
499 padding during PKCS#1 v1.5 decryption. This is a general protection against
500 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
502 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
503 on the RSA decryption context.
507 * Added support for Brainpool curves in TLS-1.3.
509 *Bernd Edlinger and Matt Caswell*
511 * Added OpenBSD specific build targets.
515 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
516 a basic thread pool implementation for select platforms.
523 ### Changes between 3.1.4 and 3.1.5 [xx XXX xxxx]
525 * Fix excessive time spent in DH check / generation with large Q parameter
528 Applications that use the functions DH_generate_key() to generate an
529 X9.42 DH key may experience long delays. Likewise, applications that use
530 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
531 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
532 Where the key or parameters that are being checked have been obtained from
533 an untrusted source this may lead to a Denial of Service.
539 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
541 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
542 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
543 that alter the key or IV length ([CVE-2023-5363]).
547 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
549 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
551 The POLY1305 MAC (message authentication code) implementation in OpenSSL
552 does not save the contents of non-volatile XMM registers on Windows 64
553 platform when calculating the MAC of data larger than 64 bytes. Before
554 returning to the caller all the XMM registers are set to zero rather than
555 restoring their previous content. The vulnerable code is used only on newer
556 x86_64 processors supporting the AVX512-IFMA instructions.
558 The consequences of this kind of internal application state corruption can
559 be various - from no consequences, if the calling application does not
560 depend on the contents of non-volatile XMM registers at all, to the worst
561 consequences, where the attacker could get complete control of the
562 application process. However given the contents of the registers are just
563 zeroized so the attacker cannot put arbitrary values inside, the most likely
564 consequence, if any, would be an incorrect result of some application
565 dependent calculations or a crash leading to a denial of service.
571 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
573 * Fix excessive time spent checking DH q parameter value.
575 The function DH_check() performs various checks on DH parameters. After
576 fixing CVE-2023-3446 it was discovered that a large q parameter value can
577 also trigger an overly long computation during some of these checks.
578 A correct q value, if present, cannot be larger than the modulus p
579 parameter, thus it is unnecessary to perform these checks if q is larger
582 If DH_check() is called with such q parameter value,
583 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
584 intensive checks are skipped.
590 * Fix DH_check() excessive time with over sized modulus.
592 The function DH_check() performs various checks on DH parameters. One of
593 those checks confirms that the modulus ("p" parameter) is not too large.
594 Trying to use a very large modulus is slow and OpenSSL will not normally use
595 a modulus which is over 10,000 bits in length.
597 However the DH_check() function checks numerous aspects of the key or
598 parameters that have been supplied. Some of those checks use the supplied
599 modulus value even if it has already been found to be too large.
601 A new limit has been added to DH_check of 32,768 bits. Supplying a
602 key/parameters with a modulus over this size will simply cause DH_check() to
609 * Do not ignore empty associated data entries with AES-SIV.
611 The AES-SIV algorithm allows for authentication of multiple associated
612 data entries along with the encryption. To authenticate empty data the
613 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
614 with NULL pointer as the output buffer and 0 as the input buffer length.
615 The AES-SIV implementation in OpenSSL just returns success for such call
616 instead of performing the associated data authentication operation.
617 The empty data thus will not be authenticated. ([CVE-2023-2975])
619 Thanks to Juerg Wullschleger (Google) for discovering the issue.
621 The fix changes the authentication tag value and the ciphertext for
622 applications that use empty associated data entries with AES-SIV.
623 To decrypt data encrypted with previous versions of OpenSSL the application
624 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
629 * When building with the `enable-fips` option and using the resulting
630 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
631 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
632 not operate with truncated digests (FIPS 140-3 IG G.R).
636 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
638 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
639 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
641 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
642 numeric text form. For gigantic sub-identifiers, this would take a very
643 long time, the time complexity being O(n^2) where n is the size of that
644 sub-identifier. ([CVE-2023-2650])
646 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
647 IDENTIFIER to canonical numeric text form if the size of that OBJECT
648 IDENTIFIER is 586 bytes or less, and fail otherwise.
650 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
651 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
652 most 128 sub-identifiers, and that the maximum value that each sub-
653 identifier may have is 2^32-1 (4294967295 decimal).
655 For each byte of every sub-identifier, only the 7 lower bits are part of
656 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
657 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
662 * Multiple algorithm implementation fixes for ARM BE platforms.
666 * Added a -pedantic option to fipsinstall that adjusts the various
667 settings to ensure strict FIPS compliance rather than backwards
672 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
673 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
674 trigger a crash of an application using AES-XTS decryption if the memory
675 just after the buffer being decrypted is not mapped.
676 Thanks to Anton Romanov (Amazon) for discovering the issue.
681 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
682 The previous fix for this timing side channel turned out to cause
683 a severe 2-3x performance regression in the typical use case
684 compared to 3.0.7. The new fix uses existing constant time
685 code paths, and restores the previous performance level while
686 fully eliminating all existing timing side channels.
687 The fix was developed by Bernd Edlinger with testing support
692 * Add FIPS provider configuration option to disallow the use of
693 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
694 The option '-no_drbg_truncated_digests' can optionally be
695 supplied to 'openssl fipsinstall'.
699 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
700 that it does not enable policy checking. Thanks to David Benjamin for
701 discovering this issue.
706 * Fixed an issue where invalid certificate policies in leaf certificates are
707 silently ignored by OpenSSL and other certificate policy checks are skipped
708 for that certificate. A malicious CA could use this to deliberately assert
709 invalid certificate policies in order to circumvent policy checking on the
710 certificate altogether.
715 * Limited the number of nodes created in a policy tree to mitigate
716 against CVE-2023-0464. The default limit is set to 1000 nodes, which
717 should be sufficient for most installations. If required, the limit
718 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
719 time define to a desired maximum number of nodes or zero to allow
725 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
727 * Add FIPS provider configuration option to enforce the
728 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
729 The option '-ems-check' can optionally be supplied to
730 'openssl fipsinstall'.
734 * The FIPS provider includes a few non-approved algorithms for
735 backward compatibility purposes and the "fips=yes" property query
736 must be used for all algorithm fetches to ensure FIPS compliance.
738 The algorithms that are included but not approved are Triple DES ECB,
739 Triple DES CBC and EdDSA.
743 * Added support for KMAC in KBKDF.
747 * RNDR and RNDRRS support in provider functions to provide
748 random number generation for Arm CPUs (aarch64).
752 * s_client and s_server apps now explicitly say when the TLS version
753 does not include the renegotiation mechanism. This avoids confusion
754 between that scenario versus when the TLS version includes secure
755 renegotiation but the peer lacks support for it.
759 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
761 *Tomasz Kantecki, Andrey Matyukov*
763 * The various OBJ_* functions have been made thread safe.
767 * Parallel dual-prime 1536/2048-bit modular exponentiation for
768 AVX512_IFMA capable processors.
770 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
772 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
773 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
774 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
775 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
776 `OPENSSL_NO_DEPRECATED_3_1`.
778 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
779 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
780 definitions for these functions regardless of whether
781 `OPENSSL_NO_DEPRECATED_3_1` is defined.
783 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
784 functions regardless of whether they are using them. It is recommended that
785 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
789 * When generating safe-prime DH parameters set the recommended private key
790 length equivalent to minimum key lengths as in RFC 7919.
794 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
795 maximum size that is smaller or equal to the digest length to comply with
796 FIPS 186-4 section 5. This is implemented by a new option
797 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
798 `rsa_pss_saltlen` parameter, which is now the default. Signature
799 verification is not affected by this change and continues to work as before.
806 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
807 listed here are only a brief description.
808 The migration guide contains more detailed information related to new features,
809 breaking changes, and mappings for the large list of deprecated functions.
811 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
813 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
815 * Fixed NULL dereference during PKCS7 data verification.
817 A NULL pointer can be dereferenced when signatures are being
818 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
819 algorithm used for the signature is known to the OpenSSL library but
820 the implementation of the hash algorithm is not available the digest
821 initialization will fail. There is a missing check for the return
822 value from the initialization function which later leads to invalid
823 usage of the digest API most likely leading to a crash.
826 PKCS7 data is processed by the SMIME library calls and also by the
827 time stamp (TS) library calls. The TLS implementation in OpenSSL does
828 not call these functions however third party applications would be
829 affected if they call these functions to verify signatures on untrusted
834 * Fixed X.400 address type confusion in X.509 GeneralName.
836 There is a type confusion vulnerability relating to X.400 address processing
837 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
838 but the public structure definition for GENERAL_NAME incorrectly specified
839 the type of the x400Address field as ASN1_TYPE. This field is subsequently
840 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
843 When CRL checking is enabled (i.e. the application sets the
844 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
845 pass arbitrary pointers to a memcmp call, enabling them to read memory
846 contents or enact a denial of service.
851 * Fixed NULL dereference validating DSA public key.
853 An invalid pointer dereference on read can be triggered when an
854 application tries to check a malformed DSA public key by the
855 EVP_PKEY_public_check() function. This will most likely lead
856 to an application crash. This function can be called on public
857 keys supplied from untrusted sources which could allow an attacker
858 to cause a denial of service attack.
860 The TLS implementation in OpenSSL does not call this function
861 but applications might call the function if there are additional
862 security requirements imposed by standards such as FIPS 140-3.
865 *Shane Lontis, Tomáš Mráz*
867 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
869 An invalid pointer dereference on read can be triggered when an
870 application tries to load malformed PKCS7 data with the
871 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
873 The result of the dereference is an application crash which could
874 lead to a denial of service attack. The TLS implementation in OpenSSL
875 does not call this function however third party applications might
876 call these functions on untrusted data.
881 * Fixed Use-after-free following BIO_new_NDEF.
883 The public API function BIO_new_NDEF is a helper function used for
884 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
885 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
886 be called directly by end user applications.
888 The function receives a BIO from the caller, prepends a new BIO_f_asn1
889 filter BIO onto the front of it to form a BIO chain, and then returns
890 the new head of the BIO chain to the caller. Under certain conditions,
891 for example if a CMS recipient public key is invalid, the new filter BIO
892 is freed and the function returns a NULL result indicating a failure.
893 However, in this case, the BIO chain is not properly cleaned up and the
894 BIO passed by the caller still retains internal pointers to the previously
895 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
896 then a use-after-free will occur. This will most likely result in a crash.
899 *Viktor Dukhovni, Matt Caswell*
901 * Fixed Double free after calling PEM_read_bio_ex.
903 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
904 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
905 data. If the function succeeds then the "name_out", "header" and "data"
906 arguments are populated with pointers to buffers containing the relevant
907 decoded data. The caller is responsible for freeing those buffers. It is
908 possible to construct a PEM file that results in 0 bytes of payload data.
909 In this case PEM_read_bio_ex() will return a failure code but will populate
910 the header argument with a pointer to a buffer that has already been freed.
911 If the caller also frees this buffer then a double free will occur. This
912 will most likely lead to a crash.
914 The functions PEM_read_bio() and PEM_read() are simple wrappers around
915 PEM_read_bio_ex() and therefore these functions are also directly affected.
917 These functions are also called indirectly by a number of other OpenSSL
918 functions including PEM_X509_INFO_read_bio_ex() and
919 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
920 internal uses of these functions are not vulnerable because the caller does
921 not free the header argument if PEM_read_bio_ex() returns a failure code.
924 *Kurt Roeckx, Matt Caswell*
926 * Fixed Timing Oracle in RSA Decryption.
928 A timing based side channel exists in the OpenSSL RSA Decryption
929 implementation which could be sufficient to recover a plaintext across
930 a network in a Bleichenbacher style attack. To achieve a successful
931 decryption an attacker would have to be able to send a very large number
932 of trial messages for decryption. The vulnerability affects all RSA padding
933 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
936 *Dmitry Belyavsky, Hubert Kario*
938 * Fixed X.509 Name Constraints Read Buffer Overflow.
940 A read buffer overrun can be triggered in X.509 certificate verification,
941 specifically in name constraint checking. The read buffer overrun might
942 result in a crash which could lead to a denial of service attack.
943 In a TLS client, this can be triggered by connecting to a malicious
944 server. In a TLS server, this can be triggered if the server requests
945 client authentication and a malicious client connects.
950 * Fixed X.509 Policy Constraints Double Locking security issue.
952 If an X.509 certificate contains a malformed policy constraint and
953 policy processing is enabled, then a write lock will be taken twice
954 recursively. On some operating systems (most widely: Windows) this
955 results in a denial of service when the affected process hangs. Policy
956 processing being enabled on a publicly facing server is not considered
957 to be a common setup.
962 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
963 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
964 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
965 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
966 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
967 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
968 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
969 for legacy EC and SM2 keys is also changed similarly to honor the
970 equivalent conversion format flag as specified in the underlying
971 `EC_KEY` object being exported to a provider, when this function is
972 called through `EVP_PKEY_export()`.
976 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
978 * Fixed two buffer overflows in punycode decoding functions.
980 A buffer overrun can be triggered in X.509 certificate verification,
981 specifically in name constraint checking. Note that this occurs after
982 certificate chain signature verification and requires either a CA to
983 have signed the malicious certificate or for the application to continue
984 certificate verification despite failure to construct a path to a trusted
987 In a TLS client, this can be triggered by connecting to a malicious
988 server. In a TLS server, this can be triggered if the server requests
989 client authentication and a malicious client connects.
991 An attacker can craft a malicious email address to overflow
992 an arbitrary number of bytes containing the `.` character (decimal 46)
993 on the stack. This buffer overflow could result in a crash (causing a
997 An attacker can craft a malicious email address to overflow four
998 attacker-controlled bytes on the stack. This buffer overflow could
999 result in a crash (causing a denial of service) or potentially remote code
1000 execution depending on stack layout for any given platform/compiler.
1005 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1006 parameters in OpenSSL code.
1007 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1008 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1009 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1010 Using these invalid names may cause algorithms to use slower methods
1011 that ignore the CRT parameters.
1015 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1020 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1021 data to be signed before signing the certificate.
1025 * Added RIPEMD160 to the default provider.
1029 * Ensured that the key share group sent or accepted for the key exchange
1030 is allowed for the protocol version.
1034 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1036 * OpenSSL supports creating a custom cipher via the legacy
1037 EVP_CIPHER_meth_new() function and associated function calls. This function
1038 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1039 to use the new provider mechanism in order to implement custom ciphers.
1041 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1042 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1043 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1044 and decryption initialisation functions). Instead of using the custom cipher
1045 directly it incorrectly tries to fetch an equivalent cipher from the
1046 available providers. An equivalent cipher is found based on the NID passed to
1047 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1048 given cipher. However it is possible for an application to incorrectly pass
1049 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1050 is used in this way the OpenSSL encryption/decryption initialisation function
1051 will match the NULL cipher as being equivalent and will fetch this from the
1052 available providers. This will succeed if the default provider has been
1053 loaded (or if a third party provider has been loaded that offers this
1054 cipher). Using the NULL cipher means that the plaintext is emitted as the
1057 Applications are only affected by this issue if they call
1058 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1059 encryption/decryption initialisation function. Applications that only use
1060 SSL/TLS are not impacted by this issue.
1065 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1070 * Fixed the linux-mips64 Configure target which was missing the
1071 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1076 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1081 * Correctly handle a retransmitted ClientHello in DTLS
1085 * Fixed detection of ktls support in cross-compile environment on Linux
1089 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1094 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1095 report correct results in some cases
1099 * Fix UWP builds by defining VirtualLock
1103 * For known safe primes use the minimum key length according to RFC 7919.
1104 Longer private key sizes unnecessarily raise the cycles needed to compute the
1105 shared secret without any increase of the real security. This fixes a
1106 regression from 1.1.1 where these shorter keys were generated for the known
1111 * Added the loongarch64 target
1115 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1116 only passed to the FIPS provider and not to the default or legacy provider.
1120 * Fixed reported performance degradation on aarch64. Restored the
1121 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1122 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1123 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1124 The new algorithm is still used for 32 bit targets.
1128 * Added a missing header for memcmp that caused compilation failure on some
1133 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1135 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1136 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1137 This issue makes the RSA implementation with 2048 bit private keys
1138 incorrect on such machines and memory corruption will happen during
1139 the computation. As a consequence of the memory corruption an attacker
1140 may be able to trigger a remote code execution on the machine performing
1143 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1144 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1145 are affected by this issue.
1150 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1151 implementation would not encrypt the entirety of the data under some
1152 circumstances. This could reveal sixteen bytes of data that was
1153 preexisting in the memory that wasn't written. In the special case of
1154 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1156 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1157 they are both unaffected.
1160 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1162 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1164 * In addition to the c_rehash shell command injection identified in
1165 CVE-2022-1292, further bugs where the c_rehash script does not
1166 properly sanitise shell metacharacters to prevent command injection have been
1169 When the CVE-2022-1292 was fixed it was not discovered that there
1170 are other places in the script where the file names of certificates
1171 being hashed were possibly passed to a command executed through the shell.
1173 This script is distributed by some operating systems in a manner where
1174 it is automatically executed. On such operating systems, an attacker
1175 could execute arbitrary commands with the privileges of the script.
1177 Use of the c_rehash script is considered obsolete and should be replaced
1178 by the OpenSSL rehash command line tool.
1181 *Daniel Fiala, Tomáš Mráz*
1183 * Case insensitive string comparison no longer uses locales. It has instead
1184 been directly implemented.
1188 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1190 * Case insensitive string comparison is reimplemented via new locale-agnostic
1191 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1192 comparison. The previous implementation had problems when the Turkish locale
1197 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1198 metacharacters to prevent command injection. This script is distributed by
1199 some operating systems in a manner where it is automatically executed. On
1200 such operating systems, an attacker could execute arbitrary commands with the
1201 privileges of the script.
1203 Use of the c_rehash script is considered obsolete and should be replaced
1204 by the OpenSSL rehash command line tool.
1209 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1210 certificate on an OCSP response. The bug caused the function in the case
1211 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1212 response (meaning a successful verification) even in the case where the
1213 response signing certificate fails to verify.
1215 It is anticipated that most users of `OCSP_basic_verify` will not use the
1216 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1217 a negative value (indicating a fatal error) in the case of a certificate
1218 verification failure. The normal expected return value in this case would be
1221 This issue also impacts the command line OpenSSL "ocsp" application. When
1222 verifying an ocsp response with the "-no_cert_checks" option the command line
1223 application will report that the verification is successful even though it
1224 has in fact failed. In this case the incorrect successful response will also
1225 be accompanied by error messages showing the failure and contradicting the
1226 apparently successful result.
1231 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1232 AAD data as the MAC key. This made the MAC key trivially predictable.
1234 An attacker could exploit this issue by performing a man-in-the-middle attack
1235 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1236 that the modified data would still pass the MAC integrity check.
1238 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1239 endpoint will always be rejected by the recipient and the connection will
1240 fail at that point. Many application protocols require data to be sent from
1241 the client to the server first. Therefore, in such a case, only an OpenSSL
1242 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1244 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1245 sent in both directions. In this case both clients and servers could be
1246 affected, regardless of the application protocol.
1248 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1249 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1250 the handshake when using this ciphersuite.
1252 The confidentiality of data is not impacted by this issue, i.e. an attacker
1253 cannot decrypt data that has been encrypted using this ciphersuite - they can
1256 In order for this attack to work both endpoints must legitimately negotiate
1257 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1258 OpenSSL 3.0, and is not available within the default provider or the default
1259 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1260 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1261 following must have occurred:
1263 1) OpenSSL must have been compiled with the (non-default) compile time option
1264 enable-weak-ssl-ciphers
1266 2) OpenSSL must have had the legacy provider explicitly loaded (either
1267 through application code or via configuration)
1269 3) The ciphersuite must have been explicitly added to the ciphersuite list
1271 4) The libssl security level must have been set to 0 (default is 1)
1273 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1275 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1276 others that both endpoints have in common
1281 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1282 occupied by the removed hash table entries.
1284 This function is used when decoding certificates or keys. If a long lived
1285 process periodically decodes certificates or keys its memory usage will
1286 expand without bounds and the process might be terminated by the operating
1287 system causing a denial of service. Also traversing the empty hash table
1288 entries will take increasingly more time.
1290 Typically such long lived processes might be TLS clients or TLS servers
1291 configured to accept client certificate authentication.
1294 *Hugo Landau, Aliaksei Levin*
1296 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1297 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1298 statistics are no longer supported. For compatibility, these statistics are
1299 still listed in the output but are now always reported as zero.
1303 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1305 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1306 for non-prime moduli.
1308 Internally this function is used when parsing certificates that contain
1309 elliptic curve public keys in compressed form or explicit elliptic curve
1310 parameters with a base point encoded in compressed form.
1312 It is possible to trigger the infinite loop by crafting a certificate that
1313 has invalid explicit curve parameters.
1315 Since certificate parsing happens prior to verification of the certificate
1316 signature, any process that parses an externally supplied certificate may thus
1317 be subject to a denial of service attack. The infinite loop can also be
1318 reached when parsing crafted private keys as they can contain explicit
1319 elliptic curve parameters.
1321 Thus vulnerable situations include:
1323 - TLS clients consuming server certificates
1324 - TLS servers consuming client certificates
1325 - Hosting providers taking certificates or private keys from customers
1326 - Certificate authorities parsing certification requests from subscribers
1327 - Anything else which parses ASN.1 elliptic curve parameters
1329 Also any other applications that use the BN_mod_sqrt() where the attacker
1330 can control the parameter values are vulnerable to this DoS issue.
1335 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1336 to the list of ciphersuites providing Perfect Forward Secrecy as
1337 required by SECLEVEL >= 3.
1339 *Dmitry Belyavskiy, Nicola Tuveri*
1341 * Made the AES constant time code for no-asm configurations
1342 optional due to the resulting 95% performance degradation.
1343 The AES constant time code can be enabled, for no assembly
1344 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1348 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1353 * The negative return value handling of the certificate verification callback
1354 was reverted. The replacement is to set the verification retry state with
1355 the SSL_set_retry_verify() function.
1359 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1361 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1362 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1363 verify a certificate supplied by a server. That function may return a
1364 negative return value to indicate an internal error (for example out of
1365 memory). Such a negative return value is mishandled by OpenSSL and will cause
1366 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1367 success and a subsequent call to SSL_get_error() to return the value
1368 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1369 returned by OpenSSL if the application has previously called
1370 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1371 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1372 totally unexpected and applications may not behave correctly as a result. The
1373 exact behaviour will depend on the application but it could result in
1374 crashes, infinite loops or other similar incorrect responses.
1376 This issue is made more serious in combination with a separate bug in OpenSSL
1377 3.0 that will cause X509_verify_cert() to indicate an internal error when
1378 processing a certificate chain. This will occur where a certificate does not
1379 include the Subject Alternative Name extension but where a Certificate
1380 Authority has enforced name constraints. This issue can occur even with valid
1386 * Corrected a few file name and file reference bugs in the build,
1387 installation and setup scripts, which lead to installation verification
1388 failures. Slightly enhanced the installation verification script.
1392 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1397 * Fixed PVK encoder to properly query for the passphrase.
1401 * Multiple fixes in the OSSL_HTTP API functions.
1405 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1406 OSSL_PARAM_INTEGER data type and return error on negative numbers
1407 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1408 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1412 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1416 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1420 * Multiple threading fixes.
1424 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1428 * Allow fetching an operation from the provider that owns an unexportable key
1429 as a fallback if that is still allowed by the property query.
1433 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1435 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1440 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1441 S390X capability vector to zero. This simplifies testing of different code
1442 paths on S390X architecture.
1446 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1447 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1448 SP 800-38D". The communication will fail at this point.
1452 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1453 confidential in EC_GROUP data.
1457 * The byte order mark (BOM) character is ignored if encountered at the
1458 beginning of a PEM-formatted file.
1462 * Added CMS support for the Russian GOST algorithms.
1466 * Due to move of the implementation of cryptographic operations
1467 to the providers, validation of various operation parameters can
1468 be postponed until the actual operation is executed where previously
1469 it happened immediately when an operation parameter was set.
1471 For example when setting an unsupported curve with
1472 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1473 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1475 *OpenSSL team members and many third party contributors*
1477 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1478 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1479 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1480 instead to retrieve these algorithms from a provider.
1484 * On build targets where the multilib postfix is set in the build
1485 configuration the libdir directory was changing based on whether
1486 the lib directory with the multilib postfix exists on the system
1487 or not. This unpredictable behavior was removed and eventual
1488 multilib postfix is now always added to the default libdir. Use
1489 `--libdir=lib` to override the libdir if adding the postfix is
1494 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1495 no longer interoperable with OpenSSL 1.1.1.
1499 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1500 function codes, this function can only cause problems for calling
1505 * Add a configurable flag to output date formats as ISO 8601. Does not
1506 change the default date format.
1510 * Version of MSVC earlier than 1300 could get link warnings, which could
1511 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1512 Support for this flag has been removed.
1516 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1517 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1518 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1519 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1520 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1524 * The signatures of the functions to get and set options on SSL and
1525 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1526 Some source code changes may be required.
1530 * The public definitions of conf_method_st and conf_st have been
1531 deprecated. They will be made opaque in a future release.
1533 *Rich Salz and Tomáš Mráz*
1535 * Client-initiated renegotiation is disabled by default. To allow it, use
1536 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1537 flag, or the "ClientRenegotiation" config parameter as appropriate.
1541 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1542 or modify relative pathname inclusion.
1546 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1547 validated. Please consult the README-FIPS and
1548 README-PROVIDERS files, as well as the migration guide.
1550 *OpenSSL team members and many third party contributors*
1552 * For the key types DH and DHX the allowed settable parameters are now different.
1556 * The openssl commands that read keys, certificates, and CRLs now
1557 automatically detect the PEM or DER format of the input files.
1559 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1561 * Added enhanced PKCS#12 APIs which accept a library context.
1565 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1569 * Added support for Kernel TLS (KTLS).
1571 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1573 * Support for RFC 5746 secure renegotiation is now required by default for
1574 SSL or TLS connections to succeed.
1578 * The signature of the `copy` functional parameter of the
1579 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1580 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1581 the signature of the `pub_decode` functional parameter of the
1582 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1583 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1587 * The error return values from some control calls (ctrl) have changed.
1591 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1595 * Many functions in the EVP_ namespace that are getters of values from
1596 implementations or contexts were renamed to include get or get0 in their
1597 names. Old names are provided as macro aliases for compatibility and
1602 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1603 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1604 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1609 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1612 * The output from the command line applications may have minor
1617 * The output from numerous "printing" may have minor changes.
1621 * Windows thread synchronization uses read/write primitives (SRWLock) when
1622 supported by the OS, otherwise CriticalSection continues to be used.
1626 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1627 work on read only BIO source/sinks that do not support these functions.
1628 This allows piping or redirection of a file BIO using stdin to be buffered
1629 into memory. This is used internally in OSSL_DECODER_from_bio().
1633 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1634 this function would return one of the values OSSL_STORE_INFO_NAME,
1635 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1636 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1637 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1638 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1639 using this function should be amended to handle the changed return value.
1643 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1644 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1645 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1646 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1647 contains more than one certificate identifier: This means that all
1648 certificates referenced there MUST be part of the validation chain.
1652 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1653 RC5, DESX and DES have been moved to the legacy provider.
1657 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1658 RIPEMD-160 have been moved to the legacy provider.
1662 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1667 * The deprecated functions EVP_PKEY_get0_RSA(),
1668 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1669 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1670 well as the similarly named "get1" functions behave differently in
1675 * A number of functions handling low-level keys or engines were deprecated
1676 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1677 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1678 EVP_PKEY_get0_siphash().
1682 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1683 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1684 will need to load the legacy crypto provider. This includes these PBE
1685 algorithms which use this KDF:
1686 - NID_pbeWithMD2AndDES_CBC
1687 - NID_pbeWithMD5AndDES_CBC
1688 - NID_pbeWithSHA1AndRC2_CBC
1689 - NID_pbeWithMD2AndRC2_CBC
1690 - NID_pbeWithMD5AndRC2_CBC
1691 - NID_pbeWithSHA1AndDES_CBC
1695 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1696 BIO_debug_callback() functions.
1700 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1701 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1705 * The RAND_METHOD APIs have been deprecated.
1709 * The SRP APIs have been deprecated.
1713 * Add a compile time option to prevent the caching of provider fetched
1714 algorithms. This is enabled by including the no-cached-fetch option
1715 at configuration time.
1719 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1720 count of PKCS12_DEFAULT_ITER.
1722 *Tomáš Mráz and Sahana Prasad*
1724 * The openssl speed command does not use low-level API calls anymore.
1728 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1731 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1733 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1737 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1738 providers may supply their own group implementations (using either the "key
1739 exchange" or the "key encapsulation" methods) which will automatically be
1740 detected and used by libssl.
1742 *Matt Caswell, Nicola Tuveri*
1744 * The undocumented function X509_certificate_type() has been deprecated;
1748 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1752 * Removed RSA padding mode for SSLv23 (which was only used for
1753 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1754 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1759 * Deprecated the obsolete X9.31 RSA key generation related functions.
1761 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1762 is not allowed to return a value > 1, this is no more taken as failure.
1764 *Viktor Dukhovni and David von Oheimb*
1766 * Deprecated the obsolete X9.31 RSA key generation related functions
1767 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1768 BN_X931_generate_prime_ex().
1772 * The default key generation method for the regular 2-prime RSA keys was
1773 changed to the FIPS 186-4 B.3.6 method.
1777 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1781 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1785 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1786 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1788 *Rich Salz, Richard Levitte, and David von Oheimb*
1790 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1794 * Deprecated `OCSP_parse_url()`.
1798 * Validation of SM2 keys has been separated from the validation of regular EC
1803 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1804 switches: a validation failure triggers an early exit, returning a failure
1805 exit status to the parent process.
1809 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1810 to ignore unknown ciphers.
1814 * The `-cipher-commands` and `-digest-commands` options
1815 of the command line utility `list` have been deprecated.
1816 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1820 * Added convenience functions for generating asymmetric key pairs:
1821 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1822 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1826 * All of the low-level EC_KEY functions have been deprecated.
1828 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1830 * Deprecated all the libcrypto and libssl error string loading
1835 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1836 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1841 * The `-crypt` option to the `passwd` command line tool has been removed.
1845 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1850 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1854 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1855 EVP_PKEY_get1_tls_encodedpoint().
1859 * The security callback, which can be customised by application code, supports
1860 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1861 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1865 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1866 interface. Their functionality remains unchanged.
1870 * Added new option for 'openssl list', '-providers', which will display the
1871 list of loaded providers, their names, version and status. It optionally
1872 displays their gettable parameters.
1876 * Removed EVP_PKEY_set_alias_type().
1880 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1881 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1885 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1886 parameters are still checked for type safety at compile time via helper
1891 * Remove the RAND_DRBG API
1893 *Paul Dale and Matthias St. Pierre*
1895 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1896 as well as actual hostnames.
1900 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1901 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1902 conversely, silently ignore DTLS protocol version bounds when configuring
1903 TLS-based contexts. The commands can be repeated to set bounds of both
1904 types. The same applies with the corresponding "min_protocol" and
1905 "max_protocol" command-line switches, in case some application uses both TLS
1908 SSL_CTX instances that are created for a fixed protocol version (e.g.
1909 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1910 attempts to apply bounds to these protocol versions would result in an
1911 error. Now only the "version-flexible" SSL_CTX instances are subject to
1912 limits in configuration files in command-line options.
1916 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1921 * Reworked the recorded ERR codes to make better space for system errors.
1922 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1923 given code is a system error (true) or an OpenSSL error (false).
1927 * Reworked the test perl framework to better allow parallel testing.
1929 *Nicola Tuveri and David von Oheimb*
1931 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1932 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1936 * 'Configure' has been changed to figure out the configuration target if
1937 none is given on the command line. Consequently, the 'config' script is
1938 now only a mere wrapper. All documentation is changed to only mention
1941 *Rich Salz and Richard Levitte*
1943 * Added a library context `OSSL_LIB_CTX` that applications as well as
1944 other libraries can use to form a separate context within which
1945 libcrypto operations are performed.
1949 * Added various `_ex` functions to the OpenSSL API that support using
1950 a non-default `OSSL_LIB_CTX`.
1954 * Handshake now fails if Extended Master Secret extension is dropped
1959 * Dropped interactive mode from the `openssl` program.
1963 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1965 *David von Oheimb and Shane Lontis*
1967 * Deprecated `EC_METHOD_get_field_type()`.
1971 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1972 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1973 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1977 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1981 * Add CAdES-BES signature verification support, mostly derived
1982 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1984 *Filipe Raimundo da Silva*
1986 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1990 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1991 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1995 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1999 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2000 EC_KEY_precompute_mult().
2004 * Deprecated EC_POINTs_mul().
2008 * Removed FIPS_mode() and FIPS_mode_set().
2012 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2016 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2017 EC_POINT_get_Jprojective_coordinates_GFp().
2021 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2022 arrays to be more easily constructed via a series of utility functions.
2023 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2024 the various push functions and finally convert to a passable OSSL_PARAM
2025 array using OSSL_PARAM_BLD_to_param().
2029 * The security strength of SHA1 and MD5 based signatures in TLS has been
2034 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2035 contain a provider side internal key.
2039 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2043 * Project text documents not yet having a proper file name extension
2044 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2045 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2049 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2050 have been converted to Markdown with the goal to produce documents
2051 which not only look pretty when viewed online in the browser, but
2052 remain well readable inside a plain text editor.
2054 To achieve this goal, a 'minimalistic' Markdown style has been applied
2055 which avoids formatting elements that interfere too much with the
2056 reading flow in the text file. For example, it
2058 * avoids [ATX headings][] and uses [setext headings][] instead
2059 (which works for `<h1>` and `<h2>` headings only).
2060 * avoids [inline links][] and uses [reference links][] instead.
2061 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2063 [ATX headings]: https://github.github.com/gfm/#atx-headings
2064 [setext headings]: https://github.github.com/gfm/#setext-headings
2065 [inline links]: https://github.github.com/gfm/#inline-link
2066 [reference links]: https://github.github.com/gfm/#reference-link
2067 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2068 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2070 *Matthias St. Pierre*
2072 * The test suite is changed to preserve results of each test recipe.
2073 A new directory test-runs/ with subdirectories named like the
2074 test recipes are created in the build tree for this purpose.
2078 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2079 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2080 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2082 *David von Oheimb, Martin Peylo*
2084 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2085 It supports arbitrary request and response content types, GET redirection,
2086 TLS, connections via HTTP(S) proxies, connections and exchange via
2087 user-defined BIOs (allowing implicit connections), persistent connections,
2088 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2089 The legacy OCSP-focused (and only partly documented) API
2090 is retained for backward compatibility, while most of it is deprecated.
2094 * Added `util/check-format.pl`, a tool for checking adherence to the
2095 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2096 The checks performed are incomplete and yield some false positives.
2097 Still the tool should be useful for detecting most typical glitches.
2101 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2102 If domain name resolution yields multiple IP addresses all of them are tried
2103 after `connect()` failures.
2107 * All of the low-level RSA functions have been deprecated.
2111 * X509 certificates signed using SHA1 are no longer allowed at security
2116 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2117 modified to use PKEY APIs. These commands are now in maintenance mode
2118 and no new features will be added to them.
2122 * The command line utility rsautl has been deprecated.
2126 * The command line utilities genrsa and rsa have been modified to use PKEY
2127 APIs. They now write PKCS#8 keys by default. These commands are now in
2128 maintenance mode and no new features will be added to them.
2132 * All of the low-level DH functions have been deprecated.
2134 *Paul Dale and Matt Caswell*
2136 * All of the low-level DSA functions have been deprecated.
2140 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2141 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2145 * Deprecated low-level ECDH and ECDSA functions.
2149 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2153 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2154 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2155 a new formulation to include all the things it can be used for,
2156 as well as words of caution.
2160 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2164 * All of the low-level HMAC functions have been deprecated.
2166 *Paul Dale and David von Oheimb*
2168 * Over two thousand fixes were made to the documentation, including:
2169 - Common options (such as -rand/-writerand, TLS version control, etc)
2170 were refactored and point to newly-enhanced descriptions in openssl.pod.
2171 - Added style conformance for all options (with help from Richard Levitte),
2172 documented all reported missing options, added a CI build to check
2173 that all options are documented and that no unimplemented options
2175 - Documented some internals, such as all use of environment variables.
2176 - Addressed all internal broken L<> references.
2180 * All of the low-level CMAC functions have been deprecated.
2184 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2185 functions have been deprecated.
2187 *Paul Dale and David von Oheimb*
2189 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2190 set of functions. The documentation mentioned negative values for some
2191 errors, but this was never the case, so the mention of negative values
2194 Code that followed the documentation and thereby check with something
2195 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2199 * All of the low-level cipher functions have been deprecated.
2201 *Matt Caswell and Paul Dale*
2203 * Removed include/openssl/opensslconf.h.in and replaced it with
2204 include/openssl/configuration.h.in, which differs in not including
2205 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2206 was added to include both.
2208 This allows internal hacks where one might need to modify the set
2209 of configured macros, for example this if deprecated symbols are
2210 still supposed to be available internally:
2212 #include <openssl/configuration.h>
2214 #undef OPENSSL_NO_DEPRECATED
2215 #define OPENSSL_SUPPRESS_DEPRECATED
2217 #include <openssl/macros.h>
2219 This should not be used by applications that use the exported
2220 symbols, as that will lead to linking errors.
2224 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2225 used in exponentiation with 512-bit moduli. No EC algorithms are
2226 affected. Analysis suggests that attacks against 2-prime RSA1024,
2227 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2228 difficult to perform and are not believed likely. Attacks against DH512
2229 are considered just feasible. However, for an attack the target would
2230 have to reuse the DH512 private key, which is not recommended anyway.
2231 Also applications directly using the low-level API BN_mod_exp may be
2232 affected if they use BN_FLG_CONSTTIME.
2237 * Most memory-debug features have been deprecated, and the functionality
2238 replaced with no-ops.
2242 * Added documentation for the STACK API.
2246 * Introduced a new method type and API, OSSL_ENCODER, to represent
2247 generic encoders. These do the same sort of job that PEM writers
2248 and d2i functions do, but with support for methods supplied by
2249 providers, and the possibility for providers to support other
2254 * Introduced a new method type and API, OSSL_DECODER, to represent
2255 generic decoders. These do the same sort of job that PEM readers
2256 and i2d functions do, but with support for methods supplied by
2257 providers, and the possibility for providers to support other
2262 * Added a .pragma directive to the syntax of configuration files, to
2263 allow varying behavior in a supported and predictable manner.
2264 Currently added pragma:
2268 This allows dollar signs to be a keyword character unless it's
2269 followed by a opening brace or parenthesis. This is useful for
2270 platforms where dollar signs are commonly used in names, such as
2271 volume names and system directory names on VMS.
2275 * Added functionality to create an EVP_PKEY from user data.
2279 * Change the interpretation of the '--api' configuration option to
2280 mean that this is a desired API compatibility level with no
2281 further meaning. The previous interpretation, that this would
2282 also mean to remove all deprecated symbols up to and including
2283 the given version, no requires that 'no-deprecated' is also used
2284 in the configuration.
2286 When building applications, the desired API compatibility level
2287 can be set with the OPENSSL_API_COMPAT macro like before. For
2288 API compatibility version below 3.0, the old style numerical
2289 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2290 For version 3.0 and on, the value is expected to be the decimal
2291 value calculated from the major and minor version like this:
2293 MAJOR * 10000 + MINOR * 100
2297 -DOPENSSL_API_COMPAT=30000 For 3.0
2298 -DOPENSSL_API_COMPAT=30200 For 3.2
2300 To hide declarations that are deprecated up to and including the
2301 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2302 given when building the application as well.
2306 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2307 access to certificate and CRL stores via URIs and OSSL_STORE
2310 This adds the following functions:
2312 - X509_LOOKUP_store()
2313 - X509_STORE_load_file()
2314 - X509_STORE_load_path()
2315 - X509_STORE_load_store()
2316 - SSL_add_store_cert_subjects_to_stack()
2317 - SSL_CTX_set_default_verify_store()
2318 - SSL_CTX_load_verify_file()
2319 - SSL_CTX_load_verify_dir()
2320 - SSL_CTX_load_verify_store()
2324 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2325 The presence of this system service is determined at run-time.
2329 * Added functionality to create an EVP_PKEY context based on data
2330 for methods from providers. This takes an algorithm name and a
2331 property query string and simply stores them, with the intent
2332 that any operation that uses this context will use those strings
2333 to fetch the needed methods implicitly, thereby making the port
2334 of application written for pre-3.0 OpenSSL easier.
2338 * The undocumented function NCONF_WIN32() has been deprecated; for
2339 conversion details see the HISTORY section of doc/man5/config.pod
2343 * Introduced the new functions EVP_DigestSignInit_ex() and
2344 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2345 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2346 pages for further details.
2350 * Over two thousand fixes were made to the documentation, including:
2351 adding missing command flags, better style conformance, documentation
2354 *Rich Salz, Richard Levitte*
2356 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2357 X25519, X448, Ed25519 and Ed448.
2361 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2366 * Deprecated the public definition of `ERR_STATE` as well as the function
2367 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2372 * Added ERR functionality to give callers access to the stored function
2373 names that have replaced the older function code based functions.
2375 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2376 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2377 ERR_peek_error_all() and ERR_peek_last_error_all().
2379 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2380 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2381 ERR_func_error_string().
2385 * Extended testing to be verbose for failing tests only. The make variables
2386 VERBOSE_FAILURE or VF can be used to enable this:
2388 $ make VF=1 test # Unix
2389 $ mms /macro=(VF=1) test ! OpenVMS
2390 $ nmake VF=1 test # Windows
2394 * Added the `-copy_extensions` option to the `x509` command for use with
2395 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2396 all extensions in the request are copied to the certificate or vice versa.
2398 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2400 * Added the `-copy_extensions` option to the `req` command for use with
2401 `-x509`. When given with the `copy` or `copyall` argument,
2402 all extensions in the certification request are copied to the certificate.
2406 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2407 they generate are by default RFC 5280 compliant in the following sense:
2408 There is a subjectKeyIdentifier extension with a hash value of the public key
2409 and for not self-signed certs there is an authorityKeyIdentifier extension
2410 with a keyIdentifier field or issuer information identifying the signing key.
2411 This is done unless some configuration overrides the new default behavior,
2412 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2416 * Added several checks to `X509_verify_cert()` according to requirements in
2417 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2418 (which may be done by using the CLI option `-x509_strict`):
2419 * The basicConstraints of CA certificates must be marked critical.
2420 * CA certificates must explicitly include the keyUsage extension.
2421 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2422 * The issuer name of any certificate must not be empty.
2423 * The subject name of CA certs, certs with keyUsage crlSign,
2424 and certs without subjectAlternativeName must not be empty.
2425 * If a subjectAlternativeName extension is given it must not be empty.
2426 * The signatureAlgorithm field and the cert signature must be consistent.
2427 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2428 must not be marked critical.
2429 * The authorityKeyIdentifier must be given for X.509v3 certs
2430 unless they are self-signed.
2431 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2435 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2436 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2440 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2441 used even when parsing explicit parameters, when loading a encoded key
2442 or calling `EC_GROUP_new_from_ecpkparameters()`/
2443 `EC_GROUP_new_from_ecparameters()`.
2444 This prevents bypass of security hardening and performance gains,
2445 especially for curves with specialized EC_METHODs.
2446 By default, if a key encoded with explicit parameters is loaded and later
2447 encoded, the output is still encoded with explicit parameters, even if
2448 internally a "named" EC_GROUP is used for computation.
2452 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2453 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2454 NULL. After this change, only the cofactor parameter can be NULL. It also
2455 does some minimal sanity checks on the passed order.
2460 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2461 An attack is simple, if the first CMS_recipientInfo is valid but the
2462 second CMS_recipientInfo is chosen ciphertext. If the second
2463 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2464 encryption key will be replaced by garbage, and the message cannot be
2465 decoded, but if the RSA decryption fails, the correct encryption key is
2466 used and the recipient will not notice the attack.
2467 As a work around for this potential attack the length of the decrypted
2468 key must be equal to the cipher default key length, in case the
2469 certificate is not given and all recipientInfo are tried out.
2470 The old behaviour can be re-enabled in the CMS code by setting the
2471 CMS_DEBUG_DECRYPT flag.
2475 * Early start up entropy quality from the DEVRANDOM seed source has been
2476 improved for older Linux systems. The RAND subsystem will wait for
2477 /dev/random to be producing output before seeding from /dev/urandom.
2478 The seeded state is stored for future library initialisations using
2479 a system global shared memory segment. The shared memory identifier
2480 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2481 the desired value. The default identifier is 114.
2485 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2486 when primes for RSA keys are computed.
2487 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2488 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2489 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2490 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2491 This avoids possible fingerprinting of newly generated RSA modules.
2495 * Correct the extended master secret constant on EBCDIC systems. Without this
2496 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2497 negotiate EMS will fail. Unfortunately this also means that TLS connections
2498 between EBCDIC systems with this fix, and EBCDIC systems without this
2499 fix will fail if they negotiate EMS.
2503 * Changed the library initialisation so that the config file is now loaded
2504 by default. This was already the case for libssl. It now occurs for both
2505 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2506 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2510 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2511 where the former acts as a replacement for `ERR_put_error()`, and the
2512 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2513 `ERR_raise_data()` adds more flexibility by taking a format string and
2514 an arbitrary number of arguments following it, to be processed with
2519 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2520 to check if a named provider is loaded and available. When called, it
2521 will also activate all fallback providers if such are still present.
2525 * Enforce a minimum DH modulus size of 512 bits.
2529 * Changed DH parameters to generate the order q subgroup instead of 2q.
2530 Previously generated DH parameters are still accepted by DH_check
2531 but DH_generate_key works around that by clearing bit 0 of the
2532 private key for those. This avoids leaking bit 0 of the private key.
2536 * Significantly reduce secure memory usage by the randomness pools.
2540 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2545 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2546 algorithms. An implementation of a key exchange algorithm can be obtained
2547 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2548 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2549 the older EVP_PKEY_derive_init() function. See the man pages for the new
2550 functions for further details.
2554 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2558 * Removed the function names from error messages and deprecated the
2563 * Removed NextStep support and the macro OPENSSL_UNISTD
2567 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2568 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2569 Also removed "export var as function" capability; we do not export
2570 variables, only functions.
2574 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2575 an error and 1 indicating success. In previous versions of OpenSSL this
2576 was a void type. If a key was set longer than the maximum possible this
2581 * Support SM2 signing and verification schemes with X509 certificate.
2585 * Use SHA256 as the default digest for TS query in the `ts` app.
2589 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2593 * Default cipher lists/suites are now available via a function, the
2594 #defines are deprecated.
2598 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2599 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2600 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2604 * Join the directories crypto/x509 and crypto/x509v3
2608 * Added command 'openssl kdf' that uses the EVP_KDF API.
2612 * Added command 'openssl mac' that uses the EVP_MAC API.
2616 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2617 as default directories. Also added the command 'openssl info'
2618 for scripting purposes.
2622 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2627 * Add prediction resistance to the DRBG reseeding process.
2631 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2632 mandated by IEEE Std 1619-2018.
2636 * Added newline escaping functionality to a filename when using openssl dgst.
2637 This output format is to replicate the output format found in the `*sum`
2638 checksum programs. This aims to preserve backward compatibility.
2640 *Matt Eaton, Richard Levitte, and Paul Dale*
2642 * Removed the heartbeat message in DTLS feature, as it has very
2643 little usage and doesn't seem to fulfill a valuable purpose.
2644 The configuration option is now deprecated.
2648 * Changed the output of 'openssl {digestname} < file' to display the
2649 digest name in its output.
2653 * Added a new generic trace API which provides support for enabling
2654 instrumentation through trace output.
2656 *Richard Levitte & Matthias St. Pierre*
2658 * Added build tests for C++. These are generated files that only do one
2659 thing, to include one public OpenSSL head file each. This tests that
2660 the public header files can be usefully included in a C++ application.
2662 This test isn't enabled by default. It can be enabled with the option
2663 'enable-buildtest-c++'.
2667 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2671 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2675 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2679 * Added KMAC to EVP_MAC.
2683 * Added property based algorithm implementation selection framework to
2688 * Added SCA hardening for modular field inversion in EC_GROUP through
2689 a new dedicated field_inv() pointer in EC_METHOD.
2690 This also addresses a leakage affecting conversions from projective
2691 to affine coordinates.
2693 *Billy Bob Brumley, Nicola Tuveri*
2695 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2696 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2697 those algorithms that were already supported through the EVP_PKEY API
2698 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2699 and scrypt are now wrappers that call EVP_KDF.
2703 * Build devcrypto engine as a dynamic engine.
2705 *Eneas U de Queiroz*
2707 * Add keyed BLAKE2 to EVP_MAC.
2711 * Fix a bug in the computation of the endpoint-pair shared secret used
2712 by DTLS over SCTP. This breaks interoperability with older versions
2713 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2714 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2715 interoperability with such broken implementations. However, enabling
2716 this switch breaks interoperability with correct implementations.
2718 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2719 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2723 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2727 * Changed the license to the Apache License v2.0.
2731 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2733 - Major releases (indicated by incrementing the MAJOR release number)
2734 may introduce incompatible API/ABI changes.
2735 - Minor releases (indicated by incrementing the MINOR release number)
2736 may introduce new features but retain API/ABI compatibility.
2737 - Patch releases (indicated by incrementing the PATCH number)
2738 are intended for bug fixes and other improvements of existing
2739 features only (like improving performance or adding documentation)
2740 and retain API/ABI compatibility.
2744 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2748 * Remove the 'dist' target and add a tarball building script. The
2749 'dist' target has fallen out of use, and it shouldn't be
2750 necessary to configure just to create a source distribution.
2754 * Recreate the OS390-Unix config target. It no longer relies on a
2755 special script like it did for OpenSSL pre-1.1.0.
2759 * Instead of having the source directories listed in Configure, add
2760 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2765 * Add GMAC to EVP_MAC.
2769 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2773 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2774 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2775 to facilitate the continued use of MACs through raw private keys in
2776 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2780 * Deprecate ECDH_KDF_X9_62().
2784 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2785 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2786 are retained for backwards compatibility.
2790 * AES-XTS mode now enforces that its two keys are different to mitigate
2791 the attacked described in "Efficient Instantiations of Tweakable
2792 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2793 Details of this attack can be obtained from:
2794 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2798 * Rename the object files, i.e. give them other names than in previous
2799 versions. Their names now include the name of the final product, as
2800 well as its type mnemonic (bin, lib, shlib).
2804 * Added new option for 'openssl list', '-objects', which will display the
2805 list of built in objects, i.e. OIDs with names.
2809 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2810 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2815 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2816 improves application performance by removing data copies and providing
2817 applications with zero-copy system calls such as sendfile and splice.
2821 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2825 * `PKCS12_parse` now maintains the order of the parsed certificates
2826 when outputting them via `*ca` (rather than reversing it).
2830 * Deprecated pthread fork support methods.
2834 * Added support for FFDHE key exchange in TLS 1.3.
2838 * Added a new concept for OpenSSL plugability: providers. This
2839 functionality is designed to replace the ENGINE API and ENGINE
2840 implementations, and to be much more dynamic, allowing provider
2841 authors to introduce new algorithms among other things, as long as
2842 there's an API that supports the algorithm type.
2844 With this concept comes a new core API for interaction between
2845 libcrypto and provider implementations. Public libcrypto functions
2846 that want to use providers do so through this core API.
2848 The main documentation for this core API is found in
2849 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2850 refer to other manuals describing the API specific for supported
2851 algorithm types (also called operations).
2858 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2860 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2862 * Avoid loading of a dynamic engine twice.
2866 * Prioritise DANE TLSA issuer certs over peer certs
2870 * Fixed random API for MacOS prior to 10.12
2872 These MacOS versions don't support the CommonCrypto APIs
2876 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2878 * Fixed an SM2 Decryption Buffer Overflow.
2880 In order to decrypt SM2 encrypted data an application is expected to
2881 call the API function EVP_PKEY_decrypt(). Typically an application will
2882 call this function twice. The first time, on entry, the "out" parameter
2883 can be NULL and, on exit, the "outlen" parameter is populated with the
2884 buffer size required to hold the decrypted plaintext. The application
2885 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2886 again, but this time passing a non-NULL value for the "out" parameter.
2888 A bug in the implementation of the SM2 decryption code means that the
2889 calculation of the buffer size required to hold the plaintext returned
2890 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2891 size required by the second call. This can lead to a buffer overflow
2892 when EVP_PKEY_decrypt() is called by the application a second time with
2893 a buffer that is too small.
2895 A malicious attacker who is able present SM2 content for decryption to
2896 an application could cause attacker chosen data to overflow the buffer
2897 by up to a maximum of 62 bytes altering the contents of other data held
2898 after the buffer, possibly changing application behaviour or causing
2899 the application to crash. The location of the buffer is application
2900 dependent but is typically heap allocated.
2905 * Fixed various read buffer overruns processing ASN.1 strings
2907 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2908 structure which contains a buffer holding the string data and a field
2909 holding the buffer length. This contrasts with normal C strings which
2910 are represented as a buffer for the string data which is terminated
2911 with a NUL (0) byte.
2913 Although not a strict requirement, ASN.1 strings that are parsed using
2914 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2915 well as any string whose value has been set with the ASN1_STRING_set()
2916 function will additionally NUL terminate the byte array in the
2917 ASN1_STRING structure.
2919 However, it is possible for applications to directly construct valid
2920 ASN1_STRING structures which do not NUL terminate the byte array by
2921 directly setting the "data" and "length" fields in the ASN1_STRING
2922 array. This can also happen by using the ASN1_STRING_set0() function.
2924 Numerous OpenSSL functions that print ASN.1 data have been found to
2925 assume that the ASN1_STRING byte array will be NUL terminated, even
2926 though this is not guaranteed for strings that have been directly
2927 constructed. Where an application requests an ASN.1 structure to be
2928 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2929 been directly constructed by the application without NUL terminating
2930 the "data" field, then a read buffer overrun can occur.
2932 The same thing can also occur during name constraints processing
2933 of certificates (for example if a certificate has been directly
2934 constructed by the application instead of loading it via the OpenSSL
2935 parsing functions, and the certificate contains non NUL terminated
2936 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2937 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2939 If a malicious actor can cause an application to directly construct an
2940 ASN1_STRING and then process it through one of the affected OpenSSL
2941 functions then this issue could be hit. This might result in a crash
2942 (causing a Denial of Service attack). It could also result in the
2943 disclosure of private memory contents (such as private keys, or
2944 sensitive plaintext).
2949 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2951 * Fixed a problem with verifying a certificate chain when using the
2952 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2953 the certificates present in a certificate chain. It is not set by default.
2955 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2956 the chain that have explicitly encoded elliptic curve parameters was added
2957 as an additional strict check.
2959 An error in the implementation of this check meant that the result of a
2960 previous check to confirm that certificates in the chain are valid CA
2961 certificates was overwritten. This effectively bypasses the check
2962 that non-CA certificates must not be able to issue other certificates.
2964 If a "purpose" has been configured then there is a subsequent opportunity
2965 for checks that the certificate is a valid CA. All of the named "purpose"
2966 values implemented in libcrypto perform this check. Therefore, where
2967 a purpose is set the certificate chain will still be rejected even when the
2968 strict flag has been used. A purpose is set by default in libssl client and
2969 server certificate verification routines, but it can be overridden or
2970 removed by an application.
2972 In order to be affected, an application must explicitly set the
2973 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2974 for the certificate verification or, in the case of TLS client or server
2975 applications, override the default purpose.
2980 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2981 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2982 renegotiation ClientHello omits the signature_algorithms extension (where it
2983 was present in the initial ClientHello), but includes a
2984 signature_algorithms_cert extension then a NULL pointer dereference will
2985 result, leading to a crash and a denial of service attack.
2987 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2988 (which is the default configuration). OpenSSL TLS clients are not impacted by
2992 *Peter Kästle and Samuel Sapalski*
2994 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2996 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2997 create a unique hash value based on the issuer and serial number data
2998 contained within an X509 certificate. However, it was failing to correctly
2999 handle any errors that may occur while parsing the issuer field (which might
3000 occur if the issuer field is maliciously constructed). This may subsequently
3001 result in a NULL pointer deref and a crash leading to a potential denial of
3007 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3008 padding mode to correctly check for rollback attacks. This is considered a
3009 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3014 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3015 functions. Previously they could overflow the output length argument in some
3016 cases where the input length is close to the maximum permissible length for
3017 an integer on the platform. In such cases the return value from the function
3018 call would be 1 (indicating success), but the output length value would be
3019 negative. This could cause applications to behave incorrectly or crash.
3024 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3025 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3026 could be exploited in a side channel attack to recover the password. Since
3027 the attack is local host only this is outside of the current OpenSSL
3028 threat model and therefore no CVE is assigned.
3030 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3035 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3037 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3038 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3039 If an attacker can control both items being compared then this could lead
3040 to a possible denial of service attack. OpenSSL itself uses the
3041 GENERAL_NAME_cmp function for two purposes:
3042 1) Comparing CRL distribution point names between an available CRL and a
3043 CRL distribution point embedded in an X509 certificate
3044 2) When verifying that a timestamp response token signer matches the
3045 timestamp authority name (exposed via the API functions
3046 TS_RESP_verify_response and TS_RESP_verify_token)
3051 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3053 * Certificates with explicit curve parameters are now disallowed in
3054 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3058 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3059 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3060 conversely, silently ignore DTLS protocol version bounds when configuring
3061 TLS-based contexts. The commands can be repeated to set bounds of both
3062 types. The same applies with the corresponding "min_protocol" and
3063 "max_protocol" command-line switches, in case some application uses both TLS
3066 SSL_CTX instances that are created for a fixed protocol version (e.g.
3067 TLSv1_server_method()) also silently ignore version bounds. Previously
3068 attempts to apply bounds to these protocol versions would result in an
3069 error. Now only the "version-flexible" SSL_CTX instances are subject to
3070 limits in configuration files in command-line options.
3074 * Handshake now fails if Extended Master Secret extension is dropped
3079 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3081 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3083 * Fixed segmentation fault in SSL_check_chain()
3084 Server or client applications that call the SSL_check_chain() function
3085 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3086 dereference as a result of incorrect handling of the
3087 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3088 or unrecognised signature algorithm is received from the peer. This could
3089 be exploited by a malicious peer in a Denial of Service attack.
3094 * Added AES consttime code for no-asm configurations
3095 an optional constant time support for AES was added
3096 when building openssl for no-asm.
3097 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3098 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3099 At this time this feature is by default disabled.
3100 It will be enabled by default in 3.0.
3104 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3106 * Revert the change of EOF detection while reading in libssl to avoid
3107 regressions in applications depending on the current way of reporting
3108 the EOF. As the existing method is not fully accurate the change to
3109 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3110 branch and will be present in the 3.0 release.
3114 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3115 when primes for RSA keys are computed.
3116 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3117 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3118 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3119 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3120 This avoids possible fingerprinting of newly generated RSA modules.
3124 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3126 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3127 while reading in libssl then we would report an error back to the
3128 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3129 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3130 therefore give a hint as to what went wrong.
3134 * Check that ed25519 and ed448 are allowed by the security level. Previously
3135 signature algorithms not using an MD were not being checked that they were
3136 allowed by the security level.
3140 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3141 was not quite right. The behaviour was not consistent between resumption
3142 and normal handshakes, and also not quite consistent with historical
3143 behaviour. The behaviour in various scenarios has been clarified and
3144 it has been updated to make it match historical behaviour as closely as
3149 * *[VMS only]* The header files that the VMS compilers include automatically,
3150 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3151 that the C++ compiler doesn't understand. This is a shortcoming in the
3152 compiler, but can be worked around with `__cplusplus` guards.
3154 C++ applications that use OpenSSL libraries must be compiled using the
3155 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3156 functions. Otherwise, only functions with symbols of less than 31
3157 characters can be used, as the linker will not be able to successfully
3158 resolve symbols with longer names.
3162 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3163 The presence of this system service is determined at run-time.
3167 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3172 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3174 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3175 number generator (RNG). This was intended to include protection in the
3176 event of a fork() system call in order to ensure that the parent and child
3177 processes did not share the same RNG state. However, this protection was not
3178 being used in the default case.
3180 A partial mitigation for this issue is that the output from a high
3181 precision timer is mixed into the RNG state so the likelihood of a parent
3182 and child process sharing state is significantly reduced.
3184 If an application already calls OPENSSL_init_crypto() explicitly using
3185 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3188 *Matthias St. Pierre*
3190 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3191 used even when parsing explicit parameters, when loading a encoded key
3192 or calling `EC_GROUP_new_from_ecpkparameters()`/
3193 `EC_GROUP_new_from_ecparameters()`.
3194 This prevents bypass of security hardening and performance gains,
3195 especially for curves with specialized EC_METHODs.
3196 By default, if a key encoded with explicit parameters is loaded and later
3197 encoded, the output is still encoded with explicit parameters, even if
3198 internally a "named" EC_GROUP is used for computation.
3202 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3203 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3204 NULL. After this change, only the cofactor parameter can be NULL. It also
3205 does some minimal sanity checks on the passed order.
3210 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3211 An attack is simple, if the first CMS_recipientInfo is valid but the
3212 second CMS_recipientInfo is chosen ciphertext. If the second
3213 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3214 encryption key will be replaced by garbage, and the message cannot be
3215 decoded, but if the RSA decryption fails, the correct encryption key is
3216 used and the recipient will not notice the attack.
3217 As a work around for this potential attack the length of the decrypted
3218 key must be equal to the cipher default key length, in case the
3219 certificate is not given and all recipientInfo are tried out.
3220 The old behaviour can be re-enabled in the CMS code by setting the
3221 CMS_DEBUG_DECRYPT flag.
3226 * Early start up entropy quality from the DEVRANDOM seed source has been
3227 improved for older Linux systems. The RAND subsystem will wait for
3228 /dev/random to be producing output before seeding from /dev/urandom.
3229 The seeded state is stored for future library initialisations using
3230 a system global shared memory segment. The shared memory identifier
3231 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3232 the desired value. The default identifier is 114.
3236 * Correct the extended master secret constant on EBCDIC systems. Without this
3237 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3238 negotiate EMS will fail. Unfortunately this also means that TLS connections
3239 between EBCDIC systems with this fix, and EBCDIC systems without this
3240 fix will fail if they negotiate EMS.
3244 * Use Windows installation paths in the mingw builds
3246 Mingw isn't a POSIX environment per se, which means that Windows
3247 paths should be used for installation.
3252 * Changed DH_check to accept parameters with order q and 2q subgroups.
3253 With order 2q subgroups the bit 0 of the private key is not secret
3254 but DH_generate_key works around that by clearing bit 0 of the
3255 private key for those. This avoids leaking bit 0 of the private key.
3259 * Significantly reduce secure memory usage by the randomness pools.
3263 * Revert the DEVRANDOM_WAIT feature for Linux systems
3265 The DEVRANDOM_WAIT feature added a select() call to wait for the
3266 /dev/random device to become readable before reading from the
3267 /dev/urandom device.
3269 It turned out that this change had negative side effects on
3270 performance which were not acceptable. After some discussion it
3271 was decided to revert this feature and leave it up to the OS
3272 resp. the platform maintainer to ensure a proper initialization
3273 during early boot time.
3275 *Matthias St. Pierre*
3277 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3279 * Add build tests for C++. These are generated files that only do one
3280 thing, to include one public OpenSSL head file each. This tests that
3281 the public header files can be usefully included in a C++ application.
3283 This test isn't enabled by default. It can be enabled with the option
3284 'enable-buildtest-c++'.
3288 * Enable SHA3 pre-hashing for ECDSA and DSA.
3292 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3293 This changes the size when using the `genpkey` command when no size is given.
3294 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3295 generation commands to use 2048 bits by default.
3299 * Reorganize the manual pages to consistently have RETURN VALUES,
3300 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3301 util/fix-doc-nits accordingly.
3303 *Paul Yang, Joshua Lock*
3305 * Add the missing accessor EVP_PKEY_get0_engine()
3309 * Have commands like `s_client` and `s_server` output the signature scheme
3310 along with other cipher suite parameters when debugging.
3314 * Make OPENSSL_config() error agnostic again.
3318 * Do the error handling in RSA decryption constant time.
3322 * Prevent over long nonces in ChaCha20-Poly1305.
3324 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3325 for every encryption operation. RFC 7539 specifies that the nonce value
3326 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3327 and front pads the nonce with 0 bytes if it is less than 12
3328 bytes. However it also incorrectly allows a nonce to be set of up to 16
3329 bytes. In this case only the last 12 bytes are significant and any
3330 additional leading bytes are ignored.
3332 It is a requirement of using this cipher that nonce values are
3333 unique. Messages encrypted using a reused nonce value are susceptible to
3334 serious confidentiality and integrity attacks. If an application changes
3335 the default nonce length to be longer than 12 bytes and then makes a
3336 change to the leading bytes of the nonce expecting the new value to be a
3337 new unique nonce then such an application could inadvertently encrypt
3338 messages with a reused nonce.
3340 Additionally the ignored bytes in a long nonce are not covered by the
3341 integrity guarantee of this cipher. Any application that relies on the
3342 integrity of these ignored leading bytes of a long nonce may be further
3343 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3344 is safe because no such use sets such a long nonce value. However user
3345 applications that use this cipher directly and set a non-default nonce
3346 length to be longer than 12 bytes may be vulnerable.
3348 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3354 * Add DEVRANDOM_WAIT feature for Linux systems
3356 On older Linux systems where the getrandom() system call is not available,
3357 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3358 Contrary to getrandom(), the /dev/urandom device will not block during
3359 early boot when the kernel CSPRNG has not been seeded yet.
3361 To mitigate this known weakness, use select() to wait for /dev/random to
3362 become readable before reading from /dev/urandom.
3364 * Ensure that SM2 only uses SM3 as digest algorithm
3368 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3370 * Change the info callback signals for the start and end of a post-handshake
3371 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3372 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3373 confused by this and assume that a TLSv1.2 renegotiation has started. This
3374 can break KeyUpdate handling. Instead we no longer signal the start and end
3375 of a post handshake message exchange (although the messages themselves are
3376 still signalled). This could break some applications that were expecting
3377 the old signals. However without this KeyUpdate is not usable for many
3382 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3384 * Timing vulnerability in DSA signature generation
3386 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3387 timing side channel attack. An attacker could use variations in the signing
3388 algorithm to recover the private key.
3390 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3395 * Timing vulnerability in ECDSA signature generation
3397 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3398 timing side channel attack. An attacker could use variations in the signing
3399 algorithm to recover the private key.
3401 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3406 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3407 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3408 of two gigabytes and the error handling improved.
3410 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3411 categorized as a normal bug, not a security issue, because the DRBG reseeds
3412 automatically and is fully functional even without additional randomness
3413 provided by the application.
3415 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3417 * Add a new ClientHello callback. Provides a callback interface that gives
3418 the application the ability to adjust the nascent SSL object at the
3419 earliest stage of ClientHello processing, immediately after extensions have
3420 been collected but before they have been processed. In particular, this
3421 callback can adjust the supported TLS versions in response to the contents
3426 * Add SM2 base algorithm support.
3430 * s390x assembly pack: add (improved) hardware-support for the following
3431 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3432 aes-cfb/cfb8, aes-ecb.
3436 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3437 parameter is no longer accepted, as it leads to a corrupt table. NULL
3438 pem_str is reserved for alias entries only.
3442 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3443 step for prime curves. The new implementation is based on formulae from
3444 differential addition-and-doubling in homogeneous projective coordinates
3445 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3446 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3447 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3448 to work in projective coordinates.
3450 *Billy Bob Brumley, Nicola Tuveri*
3452 * Change generating and checking of primes so that the error rate of not
3453 being prime depends on the intended use based on the size of the input.
3454 For larger primes this will result in more rounds of Miller-Rabin.
3455 The maximal error rate for primes with more than 1080 bits is lowered
3458 *Kurt Roeckx, Annie Yousar*
3460 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3464 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3465 moving between systems, and to avoid confusion when a Windows build is
3466 done with mingw vs with MSVC. For POSIX installs, there's still a
3467 symlink or copy named 'tsget' to avoid that confusion as well.
3471 * Revert blinding in ECDSA sign and instead make problematic addition
3472 length-invariant. Switch even to fixed-length Montgomery multiplication.
3476 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3477 step for binary curves. The new implementation is based on formulae from
3478 differential addition-and-doubling in mixed Lopez-Dahab projective
3479 coordinates, modified to independently blind the operands.
3481 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3483 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3484 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3485 EC_METHODs to implement their own specialized "ladder step", to take
3486 advantage of more favorable coordinate systems or more efficient
3487 differential addition-and-doubling algorithms.
3489 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3491 * Modified the random device based seed sources to keep the relevant
3492 file descriptors open rather than reopening them on each access.
3493 This allows such sources to operate in a chroot() jail without
3494 the associated device nodes being available. This behaviour can be
3495 controlled using RAND_keep_random_devices_open().
3499 * Numerous side-channel attack mitigations have been applied. This may have
3500 performance impacts for some algorithms for the benefit of improved
3501 security. Specific changes are noted in this change log by their respective
3506 * AIX shared library support overhaul. Switch to AIX "natural" way of
3507 handling shared libraries, which means collecting shared objects of
3508 different versions and bitnesses in one common archive. This allows to
3509 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3510 doesn't affect the way 3rd party applications are linked, only how
3511 multi-version installation is managed.
3515 * Make ec_group_do_inverse_ord() more robust and available to other
3516 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3517 mitigations are applied to the fallback BN_mod_inverse().
3518 When using this function rather than BN_mod_inverse() directly, new
3519 EC cryptosystem implementations are then safer-by-default.
3523 * Add coordinate blinding for EC_POINT and implement projective
3524 coordinate blinding for generic prime curves as a countermeasure to
3525 chosen point SCA attacks.
3527 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3529 * Add blinding to ECDSA and DSA signatures to protect against side channel
3530 attacks discovered by Keegan Ryan (NCC Group).
3534 * Enforce checking in the `pkeyutl` command to ensure that the input
3535 length does not exceed the maximum supported digest length when performing
3536 a sign, verify or verifyrecover operation.
3540 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3541 I/O in combination with something like select() or poll() will hang. This
3542 can be turned off again using SSL_CTX_clear_mode().
3543 Many applications do not properly handle non-application data records, and
3544 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3545 around the problems in those applications, but can also break some.
3546 It's recommended to read the manpages about SSL_read(), SSL_write(),
3547 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3548 SSL_CTX_set_read_ahead() again.
3552 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3553 now allow empty (zero character) pass phrases.
3557 * Apply blinding to binary field modular inversion and remove patent
3558 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3562 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3563 binary and prime elliptic curves.
3567 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3568 constant time fixed point multiplication.
3572 * Revise elliptic curve scalar multiplication with timing attack
3573 defenses: ec_wNAF_mul redirects to a constant time implementation
3574 when computing fixed point and variable point multiplication (which
3575 in OpenSSL are mostly used with secret scalars in keygen, sign,
3576 ECDH derive operations).
3577 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3580 * Updated CONTRIBUTING
3584 * Updated DRBG / RAND to request nonce and additional low entropy
3585 randomness from the system.
3587 *Matthias St. Pierre*
3589 * Updated 'openssl rehash' to use OpenSSL consistent default.
3593 * Moved the load of the ssl_conf module to libcrypto, which helps
3594 loading engines that libssl uses before libssl is initialised.
3598 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3602 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3604 *Ingo Schwarze, Rich Salz*
3606 * Added output of accepting IP address and port for 'openssl s_server'
3610 * Added a new API for TLSv1.3 ciphersuites:
3611 SSL_CTX_set_ciphersuites()
3612 SSL_set_ciphersuites()
3616 * Memory allocation failures consistently add an error to the error
3621 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3622 in libcrypto when run as setuid/setgid.
3626 * Load any config file by default when libssl is used.
3630 * Added new public header file <openssl/rand_drbg.h> and documentation
3631 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3633 *Matthias St. Pierre*
3635 * QNX support removed (cannot find contributors to get their approval
3636 for the license change).
3640 * TLSv1.3 replay protection for early data has been implemented. See the
3641 SSL_read_early_data() man page for further details.
3645 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3646 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3647 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3648 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3649 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3650 configuration has been separated out. See the ciphers man page or the
3651 SSL_CTX_set_ciphersuites() man page for more information.
3655 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3656 in responder mode now supports the new "-multi" option, which
3657 spawns the specified number of child processes to handle OCSP
3658 requests. The "-timeout" option now also limits the OCSP
3659 responder's patience to wait to receive the full client request
3660 on a newly accepted connection. Child processes are respawned
3661 as needed, and the CA index file is automatically reloaded
3662 when changed. This makes it possible to run the "ocsp" responder
3663 as a long-running service, making the OpenSSL CA somewhat more
3664 feature-complete. In this mode, most diagnostic messages logged
3665 after entering the event loop are logged via syslog(3) rather than
3670 * Added support for X448 and Ed448. Heavily based on original work by
3675 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3676 objects loaded. This adds the functions OSSL_STORE_expect() and
3677 OSSL_STORE_find() as well as needed tools to construct searches and
3678 get the search data out of them.
3682 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3683 version of OpenSSL should review their configuration settings to ensure
3684 that they are still appropriate for TLSv1.3. For further information see:
3685 <https://wiki.openssl.org/index.php/TLS1.3>
3689 * Grand redesign of the OpenSSL random generator
3691 The default RAND method now utilizes an AES-CTR DRBG according to
3692 NIST standard SP 800-90Ar1. The new random generator is essentially
3693 a port of the default random generator from the OpenSSL FIPS 2.0
3694 object module. It is a hybrid deterministic random bit generator
3695 using an AES-CTR bit stream and which seeds and reseeds itself
3696 automatically using trusted system entropy sources.
3698 Some of its new features are:
3699 - Support for multiple DRBG instances with seed chaining.
3700 - The default RAND method makes use of a DRBG.
3701 - There is a public and private DRBG instance.
3702 - The DRBG instances are fork-safe.
3703 - Keep all global DRBG instances on the secure heap if it is enabled.
3704 - The public and private DRBG instance are per thread for lock free
3707 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3709 * Changed Configure so it only says what it does and doesn't dump
3710 so much data. Instead, ./configdata.pm should be used as a script
3711 to display all sorts of configuration data.
3715 * Added processing of "make variables" to Configure.
3719 * Added SHA512/224 and SHA512/256 algorithm support.
3723 * The last traces of Netware support, first removed in 1.1.0, have
3728 * Get rid of Makefile.shared, and in the process, make the processing
3729 of certain files (rc.obj, or the .def/.map/.opt files produced from
3730 the ordinal files) more visible and hopefully easier to trace and
3731 debug (or make silent).
3735 * Make it possible to have environment variable assignments as
3736 arguments to config / Configure.
3740 * Add multi-prime RSA (RFC 8017) support.
3744 * Add SM3 implemented according to GB/T 32905-2016
3745 *Jack Lloyd <jack.lloyd@ribose.com>,*
3746 *Ronald Tse <ronald.tse@ribose.com>,*
3747 *Erick Borsboom <erick.borsboom@ribose.com>*
3749 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3750 as documented in RFC6066.
3751 Based on a patch from Tomasz Moń
3753 *Filipe Raimundo da Silva*
3755 * Add SM4 implemented according to GB/T 32907-2016.
3756 *Jack Lloyd <jack.lloyd@ribose.com>,*
3757 *Ronald Tse <ronald.tse@ribose.com>,*
3758 *Erick Borsboom <erick.borsboom@ribose.com>*
3760 * Reimplement -newreq-nodes and ERR_error_string_n; the
3761 original author does not agree with the license change.
3765 * Add ARIA AEAD TLS support.
3769 * Some macro definitions to support VS6 have been removed. Visual
3770 Studio 6 has not worked since 1.1.0
3774 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3775 without clearing the errors.
3779 * Add "atfork" functions. If building on a system that without
3780 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3781 requirements. The RAND facility now uses/requires this.
3789 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3790 not possible to disable entirely. However, it's still possible to
3791 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3794 To disable, configure with 'no-ui-console'. 'no-ui' is still
3795 possible to use as an alias. Check at compile time with the
3796 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3797 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3801 * Add a STORE module, which implements a uniform and URI based reader of
3802 stores that can contain keys, certificates, CRLs and numerous other
3803 objects. The main API is loosely based on a few stdio functions,
3804 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3805 OSSL_STORE_error and OSSL_STORE_close.
3806 The implementation uses backends called "loaders" to implement arbitrary
3807 URI schemes. There is one built in "loader" for the 'file' scheme.
3811 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3812 then adjusted to work on FreeBSD 8.4 as well.
3813 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3814 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3818 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3819 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3820 error code calls like this:
3822 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3824 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3825 that can be encoded in C. For the foreseeable future, this will only
3828 *Richard Levitte and Tim Hudson*
3830 * Removed BSD cryptodev engine.
3834 * Add a build target 'build_all_generated', to build all generated files
3835 and only that. This can be used to prepare everything that requires
3836 things like perl for a system that lacks perl and then move everything
3837 to that system and do the rest of the build there.
3841 * In the UI interface, make it possible to duplicate the user data. This
3842 can be used by engines that need to retain the data for a longer time
3843 than just the call where this user data is passed.
3847 * Ignore the '-named_curve auto' value for compatibility of applications
3850 *Tomáš Mráz <tmraz@fedoraproject.org>*
3852 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3853 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3854 alerts across multiple records (some of which could be empty). In practice
3855 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3856 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3857 support this at all. Supporting it adds significant complexity to the
3858 record layer, and its removal is unlikely to cause interoperability
3863 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3864 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3865 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3870 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3871 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3873 *Richard Levitte, Andy Polyakov*
3875 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3880 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3881 platform rather than 'mingw'.
3885 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3886 success if they are asked to add an object which already exists
3887 in the store. This change cascades to other functions which load
3888 certificates and CRLs.
3892 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3893 facilitate stack unwinding even from assembly subroutines.
3897 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3898 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3902 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3903 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3904 which is the minimum version we support.
3908 * Certificate time validation (X509_cmp_time) enforces stricter
3909 compliance with RFC 5280. Fractional seconds and timezone offsets
3910 are no longer allowed.
3914 * Add support for ARIA
3918 * s_client will now send the Server Name Indication (SNI) extension by
3919 default unless the new "-noservername" option is used. The server name is
3920 based on the host provided to the "-connect" option unless overridden by
3921 using "-servername".
3925 * Add support for SipHash
3929 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3930 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3931 prevent issues where no progress is being made and the peer continually
3932 sends unrecognised record types, using up resources processing them.
3936 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3937 using the algorithm defined in
3938 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3942 * Heartbeat support has been removed; the ABI is changed for now.
3944 *Richard Levitte, Rich Salz*
3946 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3950 * The RSA "null" method, which was partially supported to avoid patent
3951 issues, has been replaced to always returns NULL.
3958 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3960 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3961 used even when parsing explicit parameters, when loading a encoded key
3962 or calling `EC_GROUP_new_from_ecpkparameters()`/
3963 `EC_GROUP_new_from_ecparameters()`.
3964 This prevents bypass of security hardening and performance gains,
3965 especially for curves with specialized EC_METHODs.
3966 By default, if a key encoded with explicit parameters is loaded and later
3967 encoded, the output is still encoded with explicit parameters, even if
3968 internally a "named" EC_GROUP is used for computation.
3972 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3973 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3974 NULL. After this change, only the cofactor parameter can be NULL. It also
3975 does some minimal sanity checks on the passed order.
3980 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3981 An attack is simple, if the first CMS_recipientInfo is valid but the
3982 second CMS_recipientInfo is chosen ciphertext. If the second
3983 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3984 encryption key will be replaced by garbage, and the message cannot be
3985 decoded, but if the RSA decryption fails, the correct encryption key is
3986 used and the recipient will not notice the attack.
3987 As a work around for this potential attack the length of the decrypted
3988 key must be equal to the cipher default key length, in case the
3989 certificate is not given and all recipientInfo are tried out.
3990 The old behaviour can be re-enabled in the CMS code by setting the
3991 CMS_DEBUG_DECRYPT flag.
3996 * Use Windows installation paths in the mingw builds
3998 Mingw isn't a POSIX environment per se, which means that Windows
3999 paths should be used for installation.
4004 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4006 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4007 This changes the size when using the `genpkey` command when no size is given.
4008 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4009 generation commands to use 2048 bits by default.
4013 * Prevent over long nonces in ChaCha20-Poly1305.
4015 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4016 for every encryption operation. RFC 7539 specifies that the nonce value
4017 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4018 and front pads the nonce with 0 bytes if it is less than 12
4019 bytes. However it also incorrectly allows a nonce to be set of up to 16
4020 bytes. In this case only the last 12 bytes are significant and any
4021 additional leading bytes are ignored.
4023 It is a requirement of using this cipher that nonce values are
4024 unique. Messages encrypted using a reused nonce value are susceptible to
4025 serious confidentiality and integrity attacks. If an application changes
4026 the default nonce length to be longer than 12 bytes and then makes a
4027 change to the leading bytes of the nonce expecting the new value to be a
4028 new unique nonce then such an application could inadvertently encrypt
4029 messages with a reused nonce.
4031 Additionally the ignored bytes in a long nonce are not covered by the
4032 integrity guarantee of this cipher. Any application that relies on the
4033 integrity of these ignored leading bytes of a long nonce may be further
4034 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4035 is safe because no such use sets such a long nonce value. However user
4036 applications that use this cipher directly and set a non-default nonce
4037 length to be longer than 12 bytes may be vulnerable.
4039 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4045 * Added SCA hardening for modular field inversion in EC_GROUP through
4046 a new dedicated field_inv() pointer in EC_METHOD.
4047 This also addresses a leakage affecting conversions from projective
4048 to affine coordinates.
4050 *Billy Bob Brumley, Nicola Tuveri*
4052 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4053 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4057 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4061 * Remove the 'dist' target and add a tarball building script. The
4062 'dist' target has fallen out of use, and it shouldn't be
4063 necessary to configure just to create a source distribution.
4067 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4069 * Timing vulnerability in DSA signature generation
4071 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4072 timing side channel attack. An attacker could use variations in the signing
4073 algorithm to recover the private key.
4075 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4080 * Timing vulnerability in ECDSA signature generation
4082 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4083 timing side channel attack. An attacker could use variations in the signing
4084 algorithm to recover the private key.
4086 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4091 * Add coordinate blinding for EC_POINT and implement projective
4092 coordinate blinding for generic prime curves as a countermeasure to
4093 chosen point SCA attacks.
4095 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4097 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4099 * Client DoS due to large DH parameter
4101 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4102 malicious server can send a very large prime value to the client. This will
4103 cause the client to spend an unreasonably long period of time generating a
4104 key for this prime resulting in a hang until the client has finished. This
4105 could be exploited in a Denial Of Service attack.
4107 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4112 * Cache timing vulnerability in RSA Key Generation
4114 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4115 a cache timing side channel attack. An attacker with sufficient access to
4116 mount cache timing attacks during the RSA key generation process could
4117 recover the private key.
4119 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4120 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4125 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4126 parameter is no longer accepted, as it leads to a corrupt table. NULL
4127 pem_str is reserved for alias entries only.
4131 * Revert blinding in ECDSA sign and instead make problematic addition
4132 length-invariant. Switch even to fixed-length Montgomery multiplication.
4136 * Change generating and checking of primes so that the error rate of not
4137 being prime depends on the intended use based on the size of the input.
4138 For larger primes this will result in more rounds of Miller-Rabin.
4139 The maximal error rate for primes with more than 1080 bits is lowered
4142 *Kurt Roeckx, Annie Yousar*
4144 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4148 * Add blinding to ECDSA and DSA signatures to protect against side channel
4149 attacks discovered by Keegan Ryan (NCC Group).
4153 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4154 now allow empty (zero character) pass phrases.
4158 * Certificate time validation (X509_cmp_time) enforces stricter
4159 compliance with RFC 5280. Fractional seconds and timezone offsets
4160 are no longer allowed.
4164 * Fixed a text canonicalisation bug in CMS
4166 Where a CMS detached signature is used with text content the text goes
4167 through a canonicalisation process first prior to signing or verifying a
4168 signature. This process strips trailing space at the end of lines, converts
4169 line terminators to CRLF and removes additional trailing line terminators
4170 at the end of a file. A bug in the canonicalisation process meant that
4171 some characters, such as form-feed, were incorrectly treated as whitespace
4172 and removed. This is contrary to the specification (RFC5485). This fix
4173 could mean that detached text data signed with an earlier version of
4174 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4175 signed with a fixed OpenSSL may fail to verify with an earlier version of
4176 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4177 and use the "-binary" flag (for the "cms" command line application) or set
4178 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4182 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4184 * Constructed ASN.1 types with a recursive definition could exceed the stack
4186 Constructed ASN.1 types with a recursive definition (such as can be found
4187 in PKCS7) could eventually exceed the stack given malicious input with
4188 excessive recursion. This could result in a Denial Of Service attack. There
4189 are no such structures used within SSL/TLS that come from untrusted sources
4190 so this is considered safe.
4192 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4198 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4200 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4201 effectively reduced to only comparing the least significant bit of each
4202 byte. This allows an attacker to forge messages that would be considered as
4203 authenticated in an amount of tries lower than that guaranteed by the
4204 security claims of the scheme. The module can only be compiled by the
4205 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4207 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4213 * Add a build target 'build_all_generated', to build all generated files
4214 and only that. This can be used to prepare everything that requires
4215 things like perl for a system that lacks perl and then move everything
4216 to that system and do the rest of the build there.
4220 * Backport SSL_OP_NO_RENGOTIATION
4222 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4223 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4224 changes this is no longer possible in 1.1.0. Therefore, the new
4225 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4226 1.1.0 to provide equivalent functionality.
4228 Note that if an application built against 1.1.0h headers (or above) is run
4229 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4230 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4234 * Removed the OS390-Unix config target. It relied on a script that doesn't
4239 * rsaz_1024_mul_avx2 overflow bug on x86_64
4241 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4242 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4243 Analysis suggests that attacks against RSA and DSA as a result of this
4244 defect would be very difficult to perform and are not believed likely.
4245 Attacks against DH1024 are considered just feasible, because most of the
4246 work necessary to deduce information about a private key may be performed
4247 offline. The amount of resources required for such an attack would be
4248 significant. However, for an attack on TLS to be meaningful, the server
4249 would have to share the DH1024 private key among multiple clients, which is
4250 no longer an option since CVE-2016-0701.
4252 This only affects processors that support the AVX2 but not ADX extensions
4253 like Intel Haswell (4th generation).
4255 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4256 was originally found via the OSS-Fuzz project.
4261 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4263 * bn_sqrx8x_internal carry bug on x86_64
4265 There is a carry propagating bug in the x86_64 Montgomery squaring
4266 procedure. No EC algorithms are affected. Analysis suggests that attacks
4267 against RSA and DSA as a result of this defect would be very difficult to
4268 perform and are not believed likely. Attacks against DH are considered just
4269 feasible (although very difficult) because most of the work necessary to
4270 deduce information about a private key may be performed offline. The amount
4271 of resources required for such an attack would be very significant and
4272 likely only accessible to a limited number of attackers. An attacker would
4273 additionally need online access to an unpatched system using the target
4274 private key in a scenario with persistent DH parameters and a private
4275 key that is shared between multiple clients.
4277 This only affects processors that support the BMI1, BMI2 and ADX extensions
4278 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4280 This issue was reported to OpenSSL by the OSS-Fuzz project.
4285 * Malformed X.509 IPAddressFamily could cause OOB read
4287 If an X.509 certificate has a malformed IPAddressFamily extension,
4288 OpenSSL could do a one-byte buffer overread. The most likely result
4289 would be an erroneous display of the certificate in text format.
4291 This issue was reported to OpenSSL by the OSS-Fuzz project.
4296 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4298 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4299 platform rather than 'mingw'.
4303 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4304 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4305 which is the minimum version we support.
4309 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4311 * Encrypt-Then-Mac renegotiation crash
4313 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4314 negotiated where it was not in the original handshake (or vice-versa) then
4315 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4316 and servers are affected.
4318 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4323 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4325 * Truncated packet could crash via OOB read
4327 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4328 cipher is being used, then a truncated packet can cause that host to
4329 perform an out-of-bounds read, usually resulting in a crash.
4331 This issue was reported to OpenSSL by Robert Święcki of Google.
4336 * Bad (EC)DHE parameters cause a client crash
4338 If a malicious server supplies bad parameters for a DHE or ECDHE key
4339 exchange then this can result in the client attempting to dereference a
4340 NULL pointer leading to a client crash. This could be exploited in a Denial
4343 This issue was reported to OpenSSL by Guido Vranken.
4348 * BN_mod_exp may produce incorrect results on x86_64
4350 There is a carry propagating bug in the x86_64 Montgomery squaring
4351 procedure. No EC algorithms are affected. Analysis suggests that attacks
4352 against RSA and DSA as a result of this defect would be very difficult to
4353 perform and are not believed likely. Attacks against DH are considered just
4354 feasible (although very difficult) because most of the work necessary to
4355 deduce information about a private key may be performed offline. The amount
4356 of resources required for such an attack would be very significant and
4357 likely only accessible to a limited number of attackers. An attacker would
4358 additionally need online access to an unpatched system using the target
4359 private key in a scenario with persistent DH parameters and a private
4360 key that is shared between multiple clients. For example this can occur by
4361 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4362 similar to CVE-2015-3193 but must be treated as a separate problem.
4364 This issue was reported to OpenSSL by the OSS-Fuzz project.
4369 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4371 * ChaCha20/Poly1305 heap-buffer-overflow
4373 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4374 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4375 crash. This issue is not considered to be exploitable beyond a DoS.
4377 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4382 * CMS Null dereference
4384 Applications parsing invalid CMS structures can crash with a NULL pointer
4385 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4386 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4387 structure callback if an attempt is made to free certain invalid encodings.
4388 Only CHOICE structures using a callback which do not handle NULL value are
4391 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4396 * Montgomery multiplication may produce incorrect results
4398 There is a carry propagating bug in the Broadwell-specific Montgomery
4399 multiplication procedure that handles input lengths divisible by, but
4400 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4401 and DH private keys are impossible. This is because the subroutine in
4402 question is not used in operations with the private key itself and an input
4403 of the attacker's direct choice. Otherwise the bug can manifest itself as
4404 transient authentication and key negotiation failures or reproducible
4405 erroneous outcome of public-key operations with specially crafted input.
4406 Among EC algorithms only Brainpool P-512 curves are affected and one
4407 presumably can attack ECDH key negotiation. Impact was not analyzed in
4408 detail, because pre-requisites for attack are considered unlikely. Namely
4409 multiple clients have to choose the curve in question and the server has to
4410 share the private key among them, neither of which is default behaviour.
4411 Even then only clients that chose the curve will be affected.
4413 This issue was publicly reported as transient failures and was not
4414 initially recognized as a security issue. Thanks to Richard Morgan for
4415 providing reproducible case.
4420 * Removed automatic addition of RPATH in shared libraries and executables,
4421 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4425 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4427 * Fix Use After Free for large message sizes
4429 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4430 message larger than approx 16k is received then the underlying buffer to
4431 store the incoming message is reallocated and moved. Unfortunately a
4432 dangling pointer to the old location is left which results in an attempt to
4433 write to the previously freed location. This is likely to result in a
4434 crash, however it could potentially lead to execution of arbitrary code.
4436 This issue only affects OpenSSL 1.1.0a.
4438 This issue was reported to OpenSSL by Robert Święcki.
4443 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4445 * OCSP Status Request extension unbounded memory growth
4447 A malicious client can send an excessively large OCSP Status Request
4448 extension. If that client continually requests renegotiation, sending a
4449 large OCSP Status Request extension each time, then there will be unbounded
4450 memory growth on the server. This will eventually lead to a Denial Of
4451 Service attack through memory exhaustion. Servers with a default
4452 configuration are vulnerable even if they do not support OCSP. Builds using
4453 the "no-ocsp" build time option are not affected.
4455 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4460 * SSL_peek() hang on empty record
4462 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4463 sends an empty record. This could be exploited by a malicious peer in a
4464 Denial Of Service attack.
4466 This issue was reported to OpenSSL by Alex Gaynor.
4471 * Excessive allocation of memory in tls_get_message_header() and
4472 dtls1_preprocess_fragment()
4474 A (D)TLS message includes 3 bytes for its length in the header for the
4475 message. This would allow for messages up to 16Mb in length. Messages of
4476 this length are excessive and OpenSSL includes a check to ensure that a
4477 peer is sending reasonably sized messages in order to avoid too much memory
4478 being consumed to service a connection. A flaw in the logic of version
4479 1.1.0 means that memory for the message is allocated too early, prior to
4480 the excessive message length check. Due to way memory is allocated in
4481 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4482 to service a connection. This could lead to a Denial of Service through
4483 memory exhaustion. However, the excessive message length check still takes
4484 place, and this would cause the connection to immediately fail. Assuming
4485 that the application calls SSL_free() on the failed connection in a timely
4486 manner then the 21Mb of allocated memory will then be immediately freed
4487 again. Therefore, the excessive memory allocation will be transitory in
4488 nature. This then means that there is only a security impact if:
4490 1) The application does not call SSL_free() in a timely manner in the event
4491 that the connection fails
4493 2) The application is working in a constrained environment where there is
4494 very little free memory
4496 3) The attacker initiates multiple connection attempts such that there are
4497 multiple connections in a state where memory has been allocated for the
4498 connection; SSL_free() has not yet been called; and there is insufficient
4499 memory to service the multiple requests.
4501 Except in the instance of (1) above any Denial Of Service is likely to be
4502 transitory because as soon as the connection fails the memory is
4503 subsequently freed again in the SSL_free() call. However there is an
4504 increased risk during this period of application crashes due to the lack of
4505 memory - which would then mean a more serious Denial of Service.
4507 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4508 (CVE-2016-6307 and CVE-2016-6308)
4512 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4513 had to be removed. Primary reason is that vendor assembler can't
4514 assemble our modules with -KPIC flag. As result it, assembly
4515 support, was not even available as option. But its lack means
4516 lack of side-channel resistant code, which is incompatible with
4517 security by todays standards. Fortunately gcc is readily available
4518 prepackaged option, which we firmly point at...
4522 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4524 * Windows command-line tool supports UTF-8 opt-in option for arguments
4525 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4526 (to any value) allows Windows user to access PKCS#12 file generated
4527 with Windows CryptoAPI and protected with non-ASCII password, as well
4528 as files generated under UTF-8 locale on Linux also protected with
4533 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4534 have been disabled by default and removed from DEFAULT, just like RC4.
4535 See the RC4 item below to re-enable both.
4539 * The method for finding the storage location for the Windows RAND seed file
4540 has changed. First we check %RANDFILE%. If that is not set then we check
4541 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4542 all else fails we fall back to C:\.
4546 * The EVP_EncryptUpdate() function has had its return type changed from void
4547 to int. A return of 0 indicates and error while a return of 1 indicates
4552 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4553 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4554 off the constant time implementation for RSA, DSA and DH have been made
4555 no-ops and deprecated.
4559 * Windows RAND implementation was simplified to only get entropy by
4560 calling CryptGenRandom(). Various other RAND-related tickets
4563 *Joseph Wylie Yandle, Rich Salz*
4565 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4566 and `OPENSSL_LH_`, respectively. The old names are available
4567 with API compatibility. They new names are now completely documented.
4571 * Unify TYPE_up_ref(obj) methods signature.
4572 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4573 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4574 int (instead of void) like all others TYPE_up_ref() methods.
4575 So now these methods also check the return value of CRYPTO_atomic_add(),
4576 and the validity of object reference counter.
4578 *fdasilvayy@gmail.com*
4580 * With Windows Visual Studio builds, the .pdb files are installed
4581 alongside the installed libraries and executables. For a static
4582 library installation, ossl_static.pdb is the associate compiler
4583 generated .pdb file to be used when linking programs.
4587 * Remove openssl.spec. Packaging files belong with the packagers.
4591 * Automatic Darwin/OSX configuration has had a refresh, it will now
4592 recognise x86_64 architectures automatically. You can still decide
4593 to build for a different bitness with the environment variable
4594 KERNEL_BITS (can be 32 or 64), for example:
4596 KERNEL_BITS=32 ./config
4600 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4601 256 bit AES and HMAC with SHA256.
4605 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4609 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4613 * To enable users to have their own config files and build file templates,
4614 Configure looks in the directory indicated by the environment variable
4615 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4616 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4617 name and is used as is.
4621 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4622 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4623 X509_CERT_FILE_CTX was removed.
4627 * "shared" builds are now the default. To create only static libraries use
4628 the "no-shared" Configure option.
4632 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4633 All of these option have not worked for some while and are fundamental
4638 * Make various cleanup routines no-ops and mark them as deprecated. Most
4639 global cleanup functions are no longer required because they are handled
4640 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4641 Explicitly de-initing can cause problems (e.g. where a library that uses
4642 OpenSSL de-inits, but an application is still using it). The affected
4643 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4644 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4645 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4646 COMP_zlib_cleanup().
4650 * --strict-warnings no longer enables runtime debugging options
4651 such as REF_DEBUG. Instead, debug options are automatically
4652 enabled with '--debug' builds.
4654 *Andy Polyakov, Emilia Käsper*
4656 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4657 have been moved out of the public header files. New functions for managing
4658 these have been added.
4662 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4663 objects have been moved out of the public header files. New
4664 functions for managing these have been added.
4668 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4669 have been moved out of the public header files. New functions for managing
4670 these have been added.
4674 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4675 moved out of the public header files. New functions for managing these
4680 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4684 * Removed the mk1mf build scripts.
4688 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4689 it is always safe to #include a header now.
4693 * Removed the aged BC-32 config and all its supporting scripts
4697 * Removed support for Ultrix, Netware, and OS/2.
4701 * Add support for HKDF.
4703 *Alessandro Ghedini*
4705 * Add support for blake2b and blake2s
4709 * Added support for "pipelining". Ciphers that have the
4710 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4711 encryptions/decryptions simultaneously. There are currently no built-in
4712 ciphers with this property but the expectation is that engines will be able
4713 to offer it to significantly improve throughput. Support has been extended
4714 into libssl so that multiple records for a single connection can be
4715 processed in one go (for >=TLS 1.1).
4719 * Added the AFALG engine. This is an async capable engine which is able to
4720 offload work to the Linux kernel. In this initial version it only supports
4721 AES128-CBC. The kernel must be version 4.1.0 or greater.
4725 * OpenSSL now uses a new threading API. It is no longer necessary to
4726 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4727 are two supported threading models: pthreads and windows threads. It is
4728 also possible to configure OpenSSL at compile time for "no-threads". The
4729 old threading API should no longer be used. The functions have been
4730 replaced with "no-op" compatibility macros.
4732 *Alessandro Ghedini, Matt Caswell*
4734 * Modify behavior of ALPN to invoke callback after SNI/servername
4735 callback, such that updates to the SSL_CTX affect ALPN.
4739 * Add SSL_CIPHER queries for authentication and key-exchange.
4743 * Changes to the DEFAULT cipherlist:
4744 - Prefer (EC)DHE handshakes over plain RSA.
4745 - Prefer AEAD ciphers over legacy ciphers.
4746 - Prefer ECDSA over RSA when both certificates are available.
4747 - Prefer TLSv1.2 ciphers/PRF.
4748 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4753 * Change the ECC default curve list to be this, in order: x25519,
4754 secp256r1, secp521r1, secp384r1.
4758 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4759 disabled by default. They can be re-enabled using the
4760 enable-weak-ssl-ciphers option to Configure.
4764 * If the server has ALPN configured, but supports no protocols that the
4765 client advertises, send a fatal "no_application_protocol" alert.
4766 This behaviour is SHALL in RFC 7301, though it isn't universally
4767 implemented by other servers.
4771 * Add X25519 support.
4772 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4773 for public and private key encoding using the format documented in
4774 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4775 key generation and key derivation.
4777 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4782 * Deprecate SRP_VBASE_get_by_user.
4783 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4784 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4785 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4786 seed, even if the seed is configured.
4788 Users should use SRP_VBASE_get1_by_user instead. Note that in
4789 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4790 also that even though configuring the SRP seed attempts to hide
4791 invalid usernames by continuing the handshake with fake
4792 credentials, this behaviour is not constant time and no strong
4793 guarantees are made that the handshake is indistinguishable from
4794 that of a valid user.
4798 * Configuration change; it's now possible to build dynamic engines
4799 without having to build shared libraries and vice versa. This
4800 only applies to the engines in `engines/`, those in `crypto/engine/`
4801 will always be built into libcrypto (i.e. "static").
4803 Building dynamic engines is enabled by default; to disable, use
4804 the configuration option "disable-dynamic-engine".
4806 The only requirements for building dynamic engines are the
4807 presence of the DSO module and building with position independent
4808 code, so they will also automatically be disabled if configuring
4809 with "disable-dso" or "disable-pic".
4811 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4812 are also taken away from openssl/opensslconf.h, as they are
4817 * Configuration change; if there is a known flag to compile
4818 position independent code, it will always be applied on the
4819 libcrypto and libssl object files, and never on the application
4820 object files. This means other libraries that use routines from
4821 libcrypto / libssl can be made into shared libraries regardless
4822 of how OpenSSL was configured.
4824 If this isn't desirable, the configuration options "disable-pic"
4825 or "no-pic" can be used to disable the use of PIC. This will
4826 also disable building shared libraries and dynamic engines.
4830 * Removed JPAKE code. It was experimental and has no wide use.
4834 * The INSTALL_PREFIX Makefile variable has been renamed to
4835 DESTDIR. That makes for less confusion on what this variable
4836 is for. Also, the configuration option --install_prefix is
4841 * Heartbeat for TLS has been removed and is disabled by default
4842 for DTLS; configure with enable-heartbeats. Code that uses the
4843 old #define's might need to be updated.
4845 *Emilia Käsper, Rich Salz*
4847 * Rename REF_CHECK to REF_DEBUG.
4851 * New "unified" build system
4853 The "unified" build system is aimed to be a common system for all
4854 platforms we support. With it comes new support for VMS.
4856 This system builds supports building in a different directory tree
4857 than the source tree. It produces one Makefile (for unix family
4858 or lookalikes), or one descrip.mms (for VMS).
4860 The source of information to make the Makefile / descrip.mms is
4861 small files called 'build.info', holding the necessary
4862 information for each directory with source to compile, and a
4863 template in Configurations, like unix-Makefile.tmpl or
4866 With this change, the library names were also renamed on Windows
4867 and on VMS. They now have names that are closer to the standard
4868 on Unix, and include the major version number, and in certain
4869 cases, the architecture they are built for. See "Notes on shared
4870 libraries" in INSTALL.
4872 We rely heavily on the perl module Text::Template.
4876 * Added support for auto-initialisation and de-initialisation of the library.
4877 OpenSSL no longer requires explicit init or deinit routines to be called,
4878 except in certain circumstances. See the OPENSSL_init_crypto() and
4879 OPENSSL_init_ssl() man pages for further information.
4883 * The arguments to the DTLSv1_listen function have changed. Specifically the
4884 "peer" argument is now expected to be a BIO_ADDR object.
4886 * Rewrite of BIO networking library. The BIO library lacked consistent
4887 support of IPv6, and adding it required some more extensive
4888 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4889 which hold all types of addresses and chains of address information.
4890 It also introduces a new API, with functions like BIO_socket,
4891 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4892 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4893 have been adapted accordingly.
4897 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4902 * CRIME protection: disable compression by default, even if OpenSSL is
4903 compiled with zlib enabled. Applications can still enable compression
4904 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4905 using the SSL_CONF library to configure compression.
4909 * The signature of the session callback configured with
4910 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4911 was explicitly marked as `const unsigned char*` instead of
4916 * Always DPURIFY. Remove the use of uninitialized memory in the
4917 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4921 * Removed many obsolete configuration items, including
4922 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4923 MD2_CHAR, MD2_INT, MD2_LONG
4925 IDEA_SHORT, IDEA_LONG
4926 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4928 *Rich Salz, with advice from Andy Polyakov*
4930 * Many BN internals have been moved to an internal header file.
4932 *Rich Salz with help from Andy Polyakov*
4934 * Configuration and writing out the results from it has changed.
4935 Files such as Makefile include/openssl/opensslconf.h and are now
4936 produced through general templates, such as Makefile.in and
4937 crypto/opensslconf.h.in and some help from the perl module
4940 Also, the center of configuration information is no longer
4941 Makefile. Instead, Configure produces a perl module in
4942 configdata.pm which holds most of the config data (in the hash
4943 table %config), the target data that comes from the target
4944 configuration in one of the `Configurations/*.conf` files (in
4949 * To clarify their intended purposes, the Configure options
4950 --prefix and --openssldir change their semantics, and become more
4951 straightforward and less interdependent.
4953 --prefix shall be used exclusively to give the location INSTALLTOP
4954 where programs, scripts, libraries, include files and manuals are
4955 going to be installed. The default is now /usr/local.
4957 --openssldir shall be used exclusively to give the default
4958 location OPENSSLDIR where certificates, private keys, CRLs are
4959 managed. This is also where the default openssl.cnf gets
4961 If the directory given with this option is a relative path, the
4962 values of both the --prefix value and the --openssldir value will
4963 be combined to become OPENSSLDIR.
4964 The default for --openssldir is INSTALLTOP/ssl.
4966 Anyone who uses --openssldir to specify where OpenSSL is to be
4967 installed MUST change to use --prefix instead.
4971 * The GOST engine was out of date and therefore it has been removed. An up
4972 to date GOST engine is now being maintained in an external repository.
4973 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4974 support for GOST ciphersuites (these are only activated if a GOST engine
4979 * EGD is no longer supported by default; use enable-egd when
4982 *Ben Kaduk and Rich Salz*
4984 * The distribution now has Makefile.in files, which are used to
4985 create Makefile's when Configure is run. *Configure must be run
4986 before trying to build now.*
4990 * The return value for SSL_CIPHER_description() for error conditions
4995 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4997 Obtaining and performing DNSSEC validation of TLSA records is
4998 the application's responsibility. The application provides
4999 the TLSA records of its choice to OpenSSL, and these are then
5000 used to authenticate the peer.
5002 The TLSA records need not even come from DNS. They can, for
5003 example, be used to implement local end-entity certificate or
5004 trust-anchor "pinning", where the "pin" data takes the form
5005 of TLSA records, which can augment or replace verification
5006 based on the usual WebPKI public certification authorities.
5010 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5011 continues to support deprecated interfaces in default builds.
5012 However, applications are strongly advised to compile their
5013 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5014 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5015 or the 1.1.0 releases.
5017 In environments in which all applications have been ported to
5018 not use any deprecated interfaces OpenSSL's Configure script
5019 should be used with the --api=1.1.0 option to entirely remove
5020 support for the deprecated features from the library and
5021 unconditionally disable them in the installed headers.
5022 Essentially the same effect can be achieved with the "no-deprecated"
5023 argument to Configure, except that this will always restrict
5024 the build to just the latest API, rather than a fixed API
5027 As applications are ported to future revisions of the API,
5028 they should update their compile-time OPENSSL_API_COMPAT define
5029 accordingly, but in most cases should be able to continue to
5030 compile with later releases.
5032 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5033 0x10000000L and 0x00908000L, respectively. However those
5034 versions did not support the OPENSSL_API_COMPAT feature, and
5035 so applications are not typically tested for explicit support
5036 of just the undeprecated features of either release.
5040 * Add support for setting the minimum and maximum supported protocol.
5041 It can bet set via the SSL_set_min_proto_version() and
5042 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5043 MaxProtocol. It's recommended to use the new APIs to disable
5044 protocols instead of disabling individual protocols using
5045 SSL_set_options() or SSL_CONF's Protocol. This change also
5046 removes support for disabling TLS 1.2 in the OpenSSL TLS
5047 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5051 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5055 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5056 and integrates ECDSA and ECDH functionality into EC. Implementations can
5057 now redirect key generation and no longer need to convert to or from
5060 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5061 include the ec.h header file instead.
5065 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5066 ciphers who are no longer supported and drops support the ephemeral RSA key
5067 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5071 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5072 opaque. For HMAC_CTX, the following constructors and destructors
5075 HMAC_CTX *HMAC_CTX_new(void);
5076 void HMAC_CTX_free(HMAC_CTX *ctx);
5078 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5079 destroy such methods has been added. See EVP_MD_meth_new(3) and
5080 EVP_CIPHER_meth_new(3) for documentation.
5083 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5084 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5085 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5086 an already created structure.
5087 2) For consistency with the majority of our object creators and
5088 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5089 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5090 for deprecated builds.
5094 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5095 cryptographic operations to be performed asynchronously as long as an
5096 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5097 further details. Libssl has also had this capability integrated with the
5098 introduction of the new mode SSL_MODE_ASYNC and associated error
5099 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5100 pages. This work was developed in partnership with Intel Corp.
5104 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5105 always enabled now. If you want to disable the support you should
5106 exclude it using the list of supported ciphers. This also means that the
5107 "-no_ecdhe" option has been removed from s_server.
5111 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5112 SSL_{CTX_}set1_curves() which can set a list.
5116 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5117 curve you want to support using SSL_{CTX_}set1_curves().
5121 * State machine rewrite. The state machine code has been significantly
5122 refactored in order to remove much duplication of code and solve issues
5123 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5124 further details). This change does have some associated API changes.
5125 Notably the SSL_state() function has been removed and replaced by
5126 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5127 SSL_set_state() has been removed altogether. The previous handshake states
5128 defined in ssl.h and ssl3.h have also been removed.
5132 * All instances of the string "ssleay" in the public API were replaced
5133 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5134 Some error codes related to internal RSA_eay API's were renamed.
5138 * The demo files in crypto/threads were moved to demo/threads.
5142 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5145 *Matt Caswell, Rich Salz*
5147 * New ASN.1 embed macro.
5149 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5150 structure is not allocated: it is part of the parent. That is instead of
5158 This reduces memory fragmentation and make it impossible to accidentally
5159 set a mandatory field to NULL.
5161 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5162 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5163 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5168 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5172 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5173 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5174 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5175 DES and RC4 ciphersuites.
5179 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5180 This changes the decoding behaviour for some invalid messages,
5181 though the change is mostly in the more lenient direction, and
5182 legacy behaviour is preserved as much as possible.
5186 * Fix no-stdio build.
5187 *David Woodhouse <David.Woodhouse@intel.com> and also*
5188 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5190 * New testing framework
5191 The testing framework has been largely rewritten and is now using
5192 perl and the perl modules Test::Harness and an extended variant of
5193 Test::More called OpenSSL::Test to do its work. All test scripts in
5194 test/ have been rewritten into test recipes, and all direct calls to
5195 executables in test/Makefile have become individual recipes using the
5196 simplified testing OpenSSL::Test::Simple.
5198 For documentation on our testing modules, do:
5200 perldoc test/testlib/OpenSSL/Test/Simple.pm
5201 perldoc test/testlib/OpenSSL/Test.pm
5205 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5206 are used; the latter aborts on memory leaks (usually checked on exit).
5207 Some undocumented "set malloc, etc., hooks" functions were removed
5208 and others were changed. All are now documented.
5212 * In DSA_generate_parameters_ex, if the provided seed is too short,
5215 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5217 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5218 from RFC4279, RFC4785, RFC5487, RFC5489.
5220 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5221 original RSA_PSK patch.
5225 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5226 era flag was never set throughout the codebase (only read). Also removed
5227 SSL3_FLAGS_POP_BUFFER which was only used if
5228 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5232 * Changed the default name options in the "ca", "crl", "req" and "x509"
5233 to be "oneline" instead of "compat".
5237 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5238 not aware of clients that still exhibit this bug, and the workaround
5239 hasn't been working properly for a while.
5243 * The return type of BIO_number_read() and BIO_number_written() as well as
5244 the corresponding num_read and num_write members in the BIO structure has
5245 changed from unsigned long to uint64_t. On platforms where an unsigned
5246 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5251 * Given the pervasive nature of TLS extensions it is inadvisable to run
5252 OpenSSL without support for them. It also means that maintaining
5253 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5254 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5258 * Removed support for the two export grade static DH ciphersuites
5259 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5260 were newly added (along with a number of other static DH ciphersuites) to
5261 1.0.2. However the two export ones have *never* worked since they were
5262 introduced. It seems strange in any case to be adding new export
5263 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5267 * Version negotiation has been rewritten. In particular SSLv23_method(),
5268 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5269 and turned into macros which simply call the new preferred function names
5270 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5271 should use the new names instead. Also as part of this change the ssl23.h
5272 header file has been removed.
5276 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5277 code and the associated standard is no longer considered fit-for-purpose.
5281 * RT2547 was closed. When generating a private key, try to make the
5282 output file readable only by the owner. This behavior change might
5283 be noticeable when interacting with other software.
5285 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5290 * Added HTTP GET support to the ocsp command.
5294 * Changed default digest for the dgst and enc commands from MD5 to
5299 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5303 * Added support for TLS extended master secret from
5304 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5305 initial patch which was a great help during development.
5309 * All libssl internal structures have been removed from the public header
5310 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5311 now redundant). Users should not attempt to access internal structures
5312 directly. Instead they should use the provided API functions.
5316 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5317 Access to deprecated functions can be re-enabled by running config with
5318 "enable-deprecated". In addition applications wishing to use deprecated
5319 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5320 will, by default, disable some transitive includes that previously existed
5321 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5325 * Added support for OCB mode. OpenSSL has been granted a patent license
5326 compatible with the OpenSSL license for use of OCB. Details are available
5327 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5328 for OCB can be removed by calling config with no-ocb.
5332 * SSLv2 support has been removed. It still supports receiving an SSLv2
5333 compatible client hello.
5337 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5338 done while fixing the error code for the key-too-small case.
5340 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5342 * CA.sh has been removed; use CA.pl instead.
5346 * Removed old DES API.
5350 * Remove various unsupported platforms:
5356 Sinix/ReliantUNIX RM400
5361 16-bit platforms such as WIN16
5365 * Clean up OPENSSL_NO_xxx #define's
5366 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5367 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5368 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5369 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5370 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5371 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5372 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5373 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5374 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5375 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5379 * Cleaned up dead code
5380 Remove all but one '#ifdef undef' which is to be looked at.
5384 * Clean up calling of xxx_free routines.
5385 Just like free(), fix most of the xxx_free routines to accept
5386 NULL. Remove the non-null checks from callers. Save much code.
5390 * Add secure heap for storage of private keys (when possible).
5391 Add BIO_s_secmem(), CBIGNUM, etc.
5392 Contributed by Akamai Technologies under our Corporate CLA.
5396 * Experimental support for a new, fast, unbiased prime candidate generator,
5397 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5399 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5401 * New output format NSS in the sess_id command line tool. This allows
5402 exporting the session id and the master key in NSS keylog format.
5404 *Martin Kaiser <martin@kaiser.cx>*
5406 * Harmonize version and its documentation. -f flag is used to display
5409 *mancha <mancha1@zoho.com>*
5411 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5412 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5414 *mancha <mancha1@zoho.com>*
5416 * Fix some double frees. These are not thought to be exploitable.
5418 *mancha <mancha1@zoho.com>*
5420 * A missing bounds check in the handling of the TLS heartbeat extension
5421 can be used to reveal up to 64k of memory to a connected client or
5424 Thanks for Neel Mehta of Google Security for discovering this bug and to
5425 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5426 preparing the fix ([CVE-2014-0160])
5428 *Adam Langley, Bodo Moeller*
5430 * Fix for the attack described in the paper "Recovering OpenSSL
5431 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5432 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5433 <http://eprint.iacr.org/2014/140>
5435 Thanks to Yuval Yarom and Naomi Benger for discovering this
5436 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5438 *Yuval Yarom and Naomi Benger*
5440 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5441 this fixes a limitation in previous versions of OpenSSL.
5445 * Experimental encrypt-then-mac support.
5447 Experimental support for encrypt then mac from
5448 draft-gutmann-tls-encrypt-then-mac-02.txt
5450 To enable it set the appropriate extension number (0x42 for the test
5451 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5453 For non-compliant peers (i.e. just about everything) this should have no
5456 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5460 * Add EVP support for key wrapping algorithms, to avoid problems with
5461 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5462 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5463 algorithms and include tests cases.
5467 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5472 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5473 MGF1 digest and OAEP label.
5477 * Make openssl verify return errors.
5479 *Chris Palmer <palmer@google.com> and Ben Laurie*
5481 * New function ASN1_TIME_diff to calculate the difference between two
5482 ASN1_TIME structures or one structure and the current time.
5486 * Update fips_test_suite to support multiple command line options. New
5487 test to induce all self test errors in sequence and check expected
5492 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5493 sign or verify all in one operation.
5497 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5498 test programs and fips_test_suite. Includes functionality to parse
5499 the minimal script output of fipsalgest.pl directly.
5503 * Add authorisation parameter to FIPS_module_mode_set().
5507 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5511 * Use separate DRBG fields for internal and external flags. New function
5512 FIPS_drbg_health_check() to perform on demand health checking. Add
5513 generation tests to fips_test_suite with reduced health check interval to
5514 demonstrate periodic health checking. Add "nodh" option to
5515 fips_test_suite to skip very slow DH test.
5519 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5524 * More extensive health check for DRBG checking many more failure modes.
5525 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5526 combination: call this in fips_test_suite.
5530 * Add support for canonical generation of DSA parameter 'g'. See
5533 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5534 POST to handle HMAC cases.
5538 * Add functions FIPS_module_version() and FIPS_module_version_text()
5539 to return numerical and string versions of the FIPS module number.
5543 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5544 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5545 outside the validated module in the FIPS capable OpenSSL.
5549 * Minor change to DRBG entropy callback semantics. In some cases
5550 there is no multiple of the block length between min_len and
5551 max_len. Allow the callback to return more than max_len bytes
5552 of entropy but discard any extra: it is the callback's responsibility
5553 to ensure that the extra data discarded does not impact the
5554 requested amount of entropy.
5558 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5559 information in FIPS186-3, SP800-57 and SP800-131A.
5563 * CCM support via EVP. Interface is very similar to GCM case except we
5564 must supply all data in one chunk (i.e. no update, final) and the
5565 message length must be supplied if AAD is used. Add algorithm test
5570 * Initial version of POST overhaul. Add POST callback to allow the status
5571 of POST to be monitored and/or failures induced. Modify fips_test_suite
5572 to use callback. Always run all selftests even if one fails.
5576 * XTS support including algorithm test driver in the fips_gcmtest program.
5577 Note: this does increase the maximum key length from 32 to 64 bytes but
5578 there should be no binary compatibility issues as existing applications
5579 will never use XTS mode.
5583 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5584 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5585 performs algorithm blocking for unapproved PRNG types. Also do not
5586 set PRNG type in FIPS_mode_set(): leave this to the application.
5587 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5588 the standard OpenSSL PRNG: set additional data to a date time vector.
5592 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5593 This shouldn't present any incompatibility problems because applications
5594 shouldn't be using these directly and any that are will need to rethink
5595 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5599 * Extensive self tests and health checking required by SP800-90 DRBG.
5600 Remove strength parameter from FIPS_drbg_instantiate and always
5601 instantiate at maximum supported strength.
5605 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5609 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5613 * New function DH_compute_key_padded() to compute a DH key and pad with
5614 leading zeroes if needed: this complies with SP800-56A et al.
5618 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5619 anything, incomplete, subject to change and largely untested at present.
5623 * Modify fipscanisteronly build option to only build the necessary object
5624 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5628 * Add experimental option FIPSSYMS to give all symbols in
5629 fipscanister.o and FIPS or fips prefix. This will avoid
5630 conflicts with future versions of OpenSSL. Add perl script
5631 util/fipsas.pl to preprocess assembly language source files
5632 and rename any affected symbols.
5636 * Add selftest checks and algorithm block of non-fips algorithms in
5637 FIPS mode. Remove DES2 from selftests.
5641 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5642 return internal method without any ENGINE dependencies. Add new
5643 tiny fips sign and verify functions.
5647 * New build option no-ec2m to disable characteristic 2 code.
5651 * New build option "fipscanisteronly". This only builds fipscanister.o
5652 and (currently) associated fips utilities. Uses the file Makefile.fips
5653 instead of Makefile.org as the prototype.
5657 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5658 Update fips_gcmtest to use IV generator.
5662 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5663 setting output buffer to NULL. The `*Final` function must be
5664 called although it will not retrieve any additional data. The tag
5665 can be set or retrieved with a ctrl. The IV length is by default 12
5666 bytes (96 bits) but can be set to an alternative value. If the IV
5667 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5672 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5673 underlying do_cipher function handles all cipher semantics itself
5674 including padding and finalisation. This is useful if (for example)
5675 an ENGINE cipher handles block padding itself. The behaviour of
5676 do_cipher is subtly changed if this flag is set: the return value
5677 is the number of characters written to the output buffer (zero is
5678 no longer an error code) or a negative error code. Also if the
5679 input buffer is NULL and length 0 finalisation should be performed.
5683 * If a candidate issuer certificate is already part of the constructed
5684 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5688 * Improve forward-security support: add functions
5690 void SSL_CTX_set_not_resumable_session_callback(
5691 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5692 void SSL_set_not_resumable_session_callback(
5693 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5695 for use by SSL/TLS servers; the callback function will be called whenever a
5696 new session is created, and gets to decide whether the session may be
5697 cached to make it resumable (return 0) or not (return 1). (As by the
5698 SSL/TLS protocol specifications, the session_id sent by the server will be
5699 empty to indicate that the session is not resumable; also, the server will
5700 not generate RFC 4507 (RFC 5077) session tickets.)
5702 A simple reasonable callback implementation is to return is_forward_secure.
5703 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5704 by the SSL/TLS server library, indicating whether it can provide forward
5707 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5709 * New -verify_name option in command line utilities to set verification
5714 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5715 Add CMAC pkey methods.
5719 * Experimental renegotiation in s_server -www mode. If the client
5720 browses /reneg connection is renegotiated. If /renegcert it is
5721 renegotiated requesting a certificate.
5725 * Add an "external" session cache for debugging purposes to s_server. This
5726 should help trace issues which normally are only apparent in deployed
5727 multi-process servers.
5731 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5732 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5733 BIO_set_cipher() and some obscure PEM functions were changed so they
5734 can now return an error. The RAND changes required a change to the
5735 RAND_METHOD structure.
5739 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5740 a gcc attribute to warn if the result of a function is ignored. This
5741 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5742 whose return value is often ignored.
5746 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5747 These allow SCTs (signed certificate timestamps) to be requested and
5748 validated when establishing a connection.
5750 *Rob Percival <robpercival@google.com>*
5755 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5757 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5758 used even when parsing explicit parameters, when loading a encoded key
5759 or calling `EC_GROUP_new_from_ecpkparameters()`/
5760 `EC_GROUP_new_from_ecparameters()`.
5761 This prevents bypass of security hardening and performance gains,
5762 especially for curves with specialized EC_METHODs.
5763 By default, if a key encoded with explicit parameters is loaded and later
5764 encoded, the output is still encoded with explicit parameters, even if
5765 internally a "named" EC_GROUP is used for computation.
5769 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5770 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5771 NULL. After this change, only the cofactor parameter can be NULL. It also
5772 does some minimal sanity checks on the passed order.
5777 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5778 An attack is simple, if the first CMS_recipientInfo is valid but the
5779 second CMS_recipientInfo is chosen ciphertext. If the second
5780 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5781 encryption key will be replaced by garbage, and the message cannot be
5782 decoded, but if the RSA decryption fails, the correct encryption key is
5783 used and the recipient will not notice the attack.
5784 As a work around for this potential attack the length of the decrypted
5785 key must be equal to the cipher default key length, in case the
5786 certificate is not given and all recipientInfo are tried out.
5787 The old behaviour can be re-enabled in the CMS code by setting the
5788 CMS_DEBUG_DECRYPT flag.
5793 * Document issue with installation paths in diverse Windows builds
5795 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5796 binaries and run-time config file.
5801 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5803 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5804 This changes the size when using the `genpkey` command when no size is given.
5805 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5806 generation commands to use 2048 bits by default.
5810 * Add FIPS support for Android Arm 64-bit
5812 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5813 Module in Version 2.0.10. For some reason, the corresponding target
5814 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5815 built with FIPS support on Android Arm 64-bit. This omission has been
5818 *Matthias St. Pierre*
5820 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5822 * 0-byte record padding oracle
5824 If an application encounters a fatal protocol error and then calls
5825 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5826 then OpenSSL can respond differently to the calling application if a 0 byte
5827 record is received with invalid padding compared to if a 0 byte record is
5828 received with an invalid MAC. If the application then behaves differently
5829 based on that in a way that is detectable to the remote peer, then this
5830 amounts to a padding oracle that could be used to decrypt data.
5832 In order for this to be exploitable "non-stitched" ciphersuites must be in
5833 use. Stitched ciphersuites are optimised implementations of certain
5834 commonly used ciphersuites. Also the application must call SSL_shutdown()
5835 twice even if a protocol error has occurred (applications should not do
5836 this but some do anyway).
5838 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5839 Aviram, with additional investigation by Steven Collison and Andrew
5840 Hourselt. It was reported to OpenSSL on 10th December 2018.
5845 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5849 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5851 * Microarchitecture timing vulnerability in ECC scalar multiplication
5853 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5854 shown to be vulnerable to a microarchitecture timing side channel attack.
5855 An attacker with sufficient access to mount local timing attacks during
5856 ECDSA signature generation could recover the private key.
5858 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5859 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5865 * Timing vulnerability in DSA signature generation
5867 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5868 timing side channel attack. An attacker could use variations in the signing
5869 algorithm to recover the private key.
5871 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5876 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5877 Module, accidentally introduced while backporting security fixes from the
5878 development branch and hindering the use of ECC in FIPS mode.
5882 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5884 * Client DoS due to large DH parameter
5886 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5887 malicious server can send a very large prime value to the client. This will
5888 cause the client to spend an unreasonably long period of time generating a
5889 key for this prime resulting in a hang until the client has finished. This
5890 could be exploited in a Denial Of Service attack.
5892 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5897 * Cache timing vulnerability in RSA Key Generation
5899 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5900 a cache timing side channel attack. An attacker with sufficient access to
5901 mount cache timing attacks during the RSA key generation process could
5902 recover the private key.
5904 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5905 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5910 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5911 parameter is no longer accepted, as it leads to a corrupt table. NULL
5912 pem_str is reserved for alias entries only.
5916 * Revert blinding in ECDSA sign and instead make problematic addition
5917 length-invariant. Switch even to fixed-length Montgomery multiplication.
5921 * Change generating and checking of primes so that the error rate of not
5922 being prime depends on the intended use based on the size of the input.
5923 For larger primes this will result in more rounds of Miller-Rabin.
5924 The maximal error rate for primes with more than 1080 bits is lowered
5927 *Kurt Roeckx, Annie Yousar*
5929 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5933 * Add blinding to ECDSA and DSA signatures to protect against side channel
5934 attacks discovered by Keegan Ryan (NCC Group).
5938 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5939 now allow empty (zero character) pass phrases.
5943 * Certificate time validation (X509_cmp_time) enforces stricter
5944 compliance with RFC 5280. Fractional seconds and timezone offsets
5945 are no longer allowed.
5949 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5951 * Constructed ASN.1 types with a recursive definition could exceed the stack
5953 Constructed ASN.1 types with a recursive definition (such as can be found
5954 in PKCS7) could eventually exceed the stack given malicious input with
5955 excessive recursion. This could result in a Denial Of Service attack. There
5956 are no such structures used within SSL/TLS that come from untrusted sources
5957 so this is considered safe.
5959 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5965 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5967 * Read/write after SSL object in error state
5969 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5970 mechanism. The intent was that if a fatal error occurred during a handshake
5971 then OpenSSL would move into the error state and would immediately fail if
5972 you attempted to continue the handshake. This works as designed for the
5973 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5974 SSL_connect()), however due to a bug it does not work correctly if
5975 SSL_read() or SSL_write() is called directly. In that scenario, if the
5976 handshake fails then a fatal error will be returned in the initial function
5977 call. If SSL_read()/SSL_write() is subsequently called by the application
5978 for the same SSL object then it will succeed and the data is passed without
5979 being decrypted/encrypted directly from the SSL/TLS record layer.
5981 In order to exploit this issue an application bug would have to be present
5982 that resulted in a call to SSL_read()/SSL_write() being issued after having
5983 already received a fatal error.
5985 This issue was reported to OpenSSL by David Benjamin (Google).
5990 * rsaz_1024_mul_avx2 overflow bug on x86_64
5992 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5993 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5994 Analysis suggests that attacks against RSA and DSA as a result of this
5995 defect would be very difficult to perform and are not believed likely.
5996 Attacks against DH1024 are considered just feasible, because most of the
5997 work necessary to deduce information about a private key may be performed
5998 offline. The amount of resources required for such an attack would be
5999 significant. However, for an attack on TLS to be meaningful, the server
6000 would have to share the DH1024 private key among multiple clients, which is
6001 no longer an option since CVE-2016-0701.
6003 This only affects processors that support the AVX2 but not ADX extensions
6004 like Intel Haswell (4th generation).
6006 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6007 was originally found via the OSS-Fuzz project.
6012 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6014 * bn_sqrx8x_internal carry bug on x86_64
6016 There is a carry propagating bug in the x86_64 Montgomery squaring
6017 procedure. No EC algorithms are affected. Analysis suggests that attacks
6018 against RSA and DSA as a result of this defect would be very difficult to
6019 perform and are not believed likely. Attacks against DH are considered just
6020 feasible (although very difficult) because most of the work necessary to
6021 deduce information about a private key may be performed offline. The amount
6022 of resources required for such an attack would be very significant and
6023 likely only accessible to a limited number of attackers. An attacker would
6024 additionally need online access to an unpatched system using the target
6025 private key in a scenario with persistent DH parameters and a private
6026 key that is shared between multiple clients.
6028 This only affects processors that support the BMI1, BMI2 and ADX extensions
6029 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6031 This issue was reported to OpenSSL by the OSS-Fuzz project.
6036 * Malformed X.509 IPAddressFamily could cause OOB read
6038 If an X.509 certificate has a malformed IPAddressFamily extension,
6039 OpenSSL could do a one-byte buffer overread. The most likely result
6040 would be an erroneous display of the certificate in text format.
6042 This issue was reported to OpenSSL by the OSS-Fuzz project.
6046 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6048 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6049 platform rather than 'mingw'.
6053 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6055 * Truncated packet could crash via OOB read
6057 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6058 cipher is being used, then a truncated packet can cause that host to
6059 perform an out-of-bounds read, usually resulting in a crash.
6061 This issue was reported to OpenSSL by Robert Święcki of Google.
6066 * BN_mod_exp may produce incorrect results on x86_64
6068 There is a carry propagating bug in the x86_64 Montgomery squaring
6069 procedure. No EC algorithms are affected. Analysis suggests that attacks
6070 against RSA and DSA as a result of this defect would be very difficult to
6071 perform and are not believed likely. Attacks against DH are considered just
6072 feasible (although very difficult) because most of the work necessary to
6073 deduce information about a private key may be performed offline. The amount
6074 of resources required for such an attack would be very significant and
6075 likely only accessible to a limited number of attackers. An attacker would
6076 additionally need online access to an unpatched system using the target
6077 private key in a scenario with persistent DH parameters and a private
6078 key that is shared between multiple clients. For example this can occur by
6079 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6080 similar to CVE-2015-3193 but must be treated as a separate problem.
6082 This issue was reported to OpenSSL by the OSS-Fuzz project.
6087 * Montgomery multiplication may produce incorrect results
6089 There is a carry propagating bug in the Broadwell-specific Montgomery
6090 multiplication procedure that handles input lengths divisible by, but
6091 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6092 and DH private keys are impossible. This is because the subroutine in
6093 question is not used in operations with the private key itself and an input
6094 of the attacker's direct choice. Otherwise the bug can manifest itself as
6095 transient authentication and key negotiation failures or reproducible
6096 erroneous outcome of public-key operations with specially crafted input.
6097 Among EC algorithms only Brainpool P-512 curves are affected and one
6098 presumably can attack ECDH key negotiation. Impact was not analyzed in
6099 detail, because pre-requisites for attack are considered unlikely. Namely
6100 multiple clients have to choose the curve in question and the server has to
6101 share the private key among them, neither of which is default behaviour.
6102 Even then only clients that chose the curve will be affected.
6104 This issue was publicly reported as transient failures and was not
6105 initially recognized as a security issue. Thanks to Richard Morgan for
6106 providing reproducible case.
6111 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6112 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6113 prevent issues where no progress is being made and the peer continually
6114 sends unrecognised record types, using up resources processing them.
6118 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6120 * Missing CRL sanity check
6122 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6123 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6124 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6126 This issue only affects the OpenSSL 1.0.2i
6131 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6133 * OCSP Status Request extension unbounded memory growth
6135 A malicious client can send an excessively large OCSP Status Request
6136 extension. If that client continually requests renegotiation, sending a
6137 large OCSP Status Request extension each time, then there will be unbounded
6138 memory growth on the server. This will eventually lead to a Denial Of
6139 Service attack through memory exhaustion. Servers with a default
6140 configuration are vulnerable even if they do not support OCSP. Builds using
6141 the "no-ocsp" build time option are not affected.
6143 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6148 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6151 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6157 * OOB write in MDC2_Update()
6159 An overflow can occur in MDC2_Update() either if called directly or
6160 through the EVP_DigestUpdate() function using MDC2. If an attacker
6161 is able to supply very large amounts of input data after a previous
6162 call to EVP_EncryptUpdate() with a partial block then a length check
6163 can overflow resulting in a heap corruption.
6165 The amount of data needed is comparable to SIZE_MAX which is impractical
6168 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6173 * Malformed SHA512 ticket DoS
6175 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6176 DoS attack where a malformed ticket will result in an OOB read which will
6179 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6180 a custom server callback and ticket lookup mechanism.
6182 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6187 * OOB write in BN_bn2dec()
6189 The function BN_bn2dec() does not check the return value of BN_div_word().
6190 This can cause an OOB write if an application uses this function with an
6191 overly large BIGNUM. This could be a problem if an overly large certificate
6192 or CRL is printed out from an untrusted source. TLS is not affected because
6193 record limits will reject an oversized certificate before it is parsed.
6195 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6200 * OOB read in TS_OBJ_print_bio()
6202 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6203 the total length the OID text representation would use and not the amount
6204 of data written. This will result in OOB reads when large OIDs are
6207 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6212 * Pointer arithmetic undefined behaviour
6214 Avoid some undefined pointer arithmetic
6216 A common idiom in the codebase is to check limits in the following manner:
6219 Where "p" points to some malloc'd data of SIZE bytes and
6222 "len" here could be from some externally supplied data (e.g. from a TLS
6225 The rules of C pointer arithmetic are such that "p + len" is only well
6226 defined where len <= SIZE. Therefore the above idiom is actually
6227 undefined behaviour.
6229 For example this could cause problems if some malloc implementation
6230 provides an address for "p" such that "p + len" actually overflows for
6231 values of len that are too big and therefore p + len < limit.
6233 This issue was reported to OpenSSL by Guido Vranken
6238 * Constant time flag not preserved in DSA signing
6240 Operations in the DSA signing algorithm should run in constant time in
6241 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6242 implementation means that a non-constant time codepath is followed for
6243 certain operations. This has been demonstrated through a cache-timing
6244 attack to be sufficient for an attacker to recover the private DSA key.
6246 This issue was reported by César Pereida (Aalto University), Billy Brumley
6247 (Tampere University of Technology), and Yuval Yarom (The University of
6248 Adelaide and NICTA).
6253 * DTLS buffered message DoS
6255 In a DTLS connection where handshake messages are delivered out-of-order
6256 those messages that OpenSSL is not yet ready to process will be buffered
6257 for later use. Under certain circumstances, a flaw in the logic means that
6258 those messages do not get removed from the buffer even though the handshake
6259 has been completed. An attacker could force up to approx. 15 messages to
6260 remain in the buffer when they are no longer required. These messages will
6261 be cleared when the DTLS connection is closed. The default maximum size for
6262 a message is 100k. Therefore, the attacker could force an additional 1500k
6263 to be consumed per connection. By opening many simultaneous connections an
6264 attacker could cause a DoS attack through memory exhaustion.
6266 This issue was reported to OpenSSL by Quan Luo.
6271 * DTLS replay protection DoS
6273 A flaw in the DTLS replay attack protection mechanism means that records
6274 that arrive for future epochs update the replay protection "window" before
6275 the MAC for the record has been validated. This could be exploited by an
6276 attacker by sending a record for the next epoch (which does not have to
6277 decrypt or have a valid MAC), with a very large sequence number. This means
6278 that all subsequent legitimate packets are dropped causing a denial of
6279 service for a specific DTLS connection.
6281 This issue was reported to OpenSSL by the OCAP audit team.
6286 * Certificate message OOB reads
6288 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6289 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6290 theoretical DoS risk but this has not been observed in practice on common
6293 The messages affected are client certificate, client certificate request
6294 and server certificate. As a result the attack can only be performed
6295 against a client or a server which enables client authentication.
6297 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6302 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6304 * Prevent padding oracle in AES-NI CBC MAC check
6306 A MITM attacker can use a padding oracle attack to decrypt traffic
6307 when the connection uses an AES CBC cipher and the server support
6310 This issue was introduced as part of the fix for Lucky 13 padding
6311 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6312 constant time by making sure that always the same bytes are read and
6313 compared against either the MAC or padding bytes. But it no longer
6314 checked that there was enough data to have both the MAC and padding
6317 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6321 * Fix EVP_EncodeUpdate overflow
6323 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6324 Base64 encoding of binary data. If an attacker is able to supply very large
6325 amounts of input data then a length check can overflow resulting in a heap
6328 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6329 the `PEM_write_bio*` family of functions. These are mainly used within the
6330 OpenSSL command line applications, so any application which processes data
6331 from an untrusted source and outputs it as a PEM file should be considered
6332 vulnerable to this issue. User applications that call these APIs directly
6333 with large amounts of untrusted data may also be vulnerable.
6335 This issue was reported by Guido Vranken.
6340 * Fix EVP_EncryptUpdate overflow
6342 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6343 is able to supply very large amounts of input data after a previous call to
6344 EVP_EncryptUpdate() with a partial block then a length check can overflow
6345 resulting in a heap corruption. Following an analysis of all OpenSSL
6346 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6347 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6348 the first called function after an EVP_EncryptInit(), and therefore that
6349 specific call must be safe. The second form is where the length passed to
6350 EVP_EncryptUpdate() can be seen from the code to be some small value and
6351 therefore there is no possibility of an overflow. Since all instances are
6352 one of these two forms, it is believed that there can be no overflows in
6353 internal code due to this problem. It should be noted that
6354 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6355 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6356 of these calls have also been analysed too and it is believed there are no
6357 instances in internal usage where an overflow could occur.
6359 This issue was reported by Guido Vranken.
6364 * Prevent ASN.1 BIO excessive memory allocation
6366 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6367 a short invalid encoding can cause allocation of large amounts of memory
6368 potentially consuming excessive resources or exhausting memory.
6370 Any application parsing untrusted data through d2i BIO functions is
6371 affected. The memory based functions such as d2i_X509() are *not* affected.
6372 Since the memory based functions are used by the TLS library, TLS
6373 applications are not affected.
6375 This issue was reported by Brian Carpenter.
6382 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6383 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6384 in arbitrary stack data being returned in the buffer.
6386 This issue was reported by Guido Vranken.
6391 * Modify behavior of ALPN to invoke callback after SNI/servername
6392 callback, such that updates to the SSL_CTX affect ALPN.
6396 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6401 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6402 methods are enabled and ssl2 is disabled the methods return NULL.
6406 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6408 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6409 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6410 provide any "EXPORT" or "LOW" strength ciphers.
6414 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6415 is by default disabled at build-time. Builds that are not configured with
6416 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6417 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6418 will need to explicitly call either of:
6420 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6422 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6424 as appropriate. Even if either of those is used, or the application
6425 explicitly uses the version-specific SSLv2_method() or its client and
6426 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6427 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6428 ciphers, and SSLv2 56-bit DES are no longer available.
6433 * Fix a double-free in DSA code
6435 A double free bug was discovered when OpenSSL parses malformed DSA private
6436 keys and could lead to a DoS attack or memory corruption for applications
6437 that receive DSA private keys from untrusted sources. This scenario is
6440 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6446 * Disable SRP fake user seed to address a server memory leak.
6448 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6450 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6451 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6452 was changed to ignore the "fake user" SRP seed, even if the seed
6455 Users should use SRP_VBASE_get1_by_user instead. Note that in
6456 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6457 also that even though configuring the SRP seed attempts to hide
6458 invalid usernames by continuing the handshake with fake
6459 credentials, this behaviour is not constant time and no strong
6460 guarantees are made that the handshake is indistinguishable from
6461 that of a valid user.
6466 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6468 In the BN_hex2bn function the number of hex digits is calculated using an
6469 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6470 large values of `i` this can result in `bn_expand` not allocating any
6471 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6472 field as NULL leading to a subsequent NULL ptr deref. For very large values
6473 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6474 In this case memory is allocated to the internal BIGNUM data field, but it
6475 is insufficiently sized leading to heap corruption. A similar issue exists
6476 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6477 is ever called by user applications with very large untrusted hex/dec data.
6478 This is anticipated to be a rare occurrence.
6480 All OpenSSL internal usage of these functions use data that is not expected
6481 to be untrusted, e.g. config file data or application command line
6482 arguments. If user developed applications generate config file data based
6483 on untrusted data then it is possible that this could also lead to security
6484 consequences. This is also anticipated to be rare.
6486 This issue was reported to OpenSSL by Guido Vranken.
6491 * Fix memory issues in `BIO_*printf` functions
6493 The internal `fmtstr` function used in processing a "%s" format string in
6494 the `BIO_*printf` functions could overflow while calculating the length of a
6495 string and cause an OOB read when printing very long strings.
6497 Additionally the internal `doapr_outch` function can attempt to write to an
6498 OOB memory location (at an offset from the NULL pointer) in the event of a
6499 memory allocation failure. In 1.0.2 and below this could be caused where
6500 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6501 could be in processing a very long "%s" format string. Memory leaks can
6504 The first issue may mask the second issue dependent on compiler behaviour.
6505 These problems could enable attacks where large amounts of untrusted data
6506 is passed to the `BIO_*printf` functions. If applications use these functions
6507 in this way then they could be vulnerable. OpenSSL itself uses these
6508 functions when printing out human-readable dumps of ASN.1 data. Therefore
6509 applications that print this data could be vulnerable if the data is from
6510 untrusted sources. OpenSSL command line applications could also be
6511 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6512 as command line arguments.
6514 Libssl is not considered directly vulnerable. Additionally certificates etc
6515 received via remote connections via libssl are also unlikely to be able to
6516 trigger these issues because of message size limits enforced within libssl.
6518 This issue was reported to OpenSSL Guido Vranken.
6523 * Side channel attack on modular exponentiation
6525 A side-channel attack was found which makes use of cache-bank conflicts on
6526 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6527 of RSA keys. The ability to exploit this issue is limited as it relies on
6528 an attacker who has control of code in a thread running on the same
6529 hyper-threaded core as the victim thread which is performing decryptions.
6531 This issue was reported to OpenSSL by Yuval Yarom, The University of
6532 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6533 Nadia Heninger, University of Pennsylvania with more information at
6534 <http://cachebleed.info>.
6539 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6540 if no keysize is specified with default_bits. This fixes an
6541 omission in an earlier change that changed all RSA/DSA key generation
6542 commands to use 2048 bits by default.
6546 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6548 * DH small subgroups
6550 Historically OpenSSL only ever generated DH parameters based on "safe"
6551 primes. More recently (in version 1.0.2) support was provided for
6552 generating X9.42 style parameter files such as those required for RFC 5114
6553 support. The primes used in such files may not be "safe". Where an
6554 application is using DH configured with parameters based on primes that are
6555 not "safe" then an attacker could use this fact to find a peer's private
6556 DH exponent. This attack requires that the attacker complete multiple
6557 handshakes in which the peer uses the same private DH exponent. For example
6558 this could be used to discover a TLS server's private DH exponent if it's
6559 reusing the private DH exponent or it's using a static DH ciphersuite.
6561 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6562 TLS. It is not on by default. If the option is not set then the server
6563 reuses the same private DH exponent for the life of the server process and
6564 would be vulnerable to this attack. It is believed that many popular
6565 applications do set this option and would therefore not be at risk.
6567 The fix for this issue adds an additional check where a "q" parameter is
6568 available (as is the case in X9.42 based parameters). This detects the
6569 only known attack, and is the only possible defense for static DH
6570 ciphersuites. This could have some performance impact.
6572 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6573 default and cannot be disabled. This could have some performance impact.
6575 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6580 * SSLv2 doesn't block disabled ciphers
6582 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6583 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6584 been disabled, provided that the SSLv2 protocol was not also disabled via
6587 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6588 and Sebastian Schinzel.
6593 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6595 * BN_mod_exp may produce incorrect results on x86_64
6597 There is a carry propagating bug in the x86_64 Montgomery squaring
6598 procedure. No EC algorithms are affected. Analysis suggests that attacks
6599 against RSA and DSA as a result of this defect would be very difficult to
6600 perform and are not believed likely. Attacks against DH are considered just
6601 feasible (although very difficult) because most of the work necessary to
6602 deduce information about a private key may be performed offline. The amount
6603 of resources required for such an attack would be very significant and
6604 likely only accessible to a limited number of attackers. An attacker would
6605 additionally need online access to an unpatched system using the target
6606 private key in a scenario with persistent DH parameters and a private
6607 key that is shared between multiple clients. For example this can occur by
6608 default in OpenSSL DHE based SSL/TLS ciphersuites.
6610 This issue was reported to OpenSSL by Hanno Böck.
6615 * Certificate verify crash with missing PSS parameter
6617 The signature verification routines will crash with a NULL pointer
6618 dereference if presented with an ASN.1 signature using the RSA PSS
6619 algorithm and absent mask generation function parameter. Since these
6620 routines are used to verify certificate signature algorithms this can be
6621 used to crash any certificate verification operation and exploited in a
6622 DoS attack. Any application which performs certificate verification is
6623 vulnerable including OpenSSL clients and servers which enable client
6626 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6631 * X509_ATTRIBUTE memory leak
6633 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6634 memory. This structure is used by the PKCS#7 and CMS routines so any
6635 application which reads PKCS#7 or CMS data from untrusted sources is
6636 affected. SSL/TLS is not affected.
6638 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6644 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6645 This changes the decoding behaviour for some invalid messages,
6646 though the change is mostly in the more lenient direction, and
6647 legacy behaviour is preserved as much as possible.
6651 * In DSA_generate_parameters_ex, if the provided seed is too short,
6654 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6656 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6658 * Alternate chains certificate forgery
6660 During certificate verification, OpenSSL will attempt to find an
6661 alternative certificate chain if the first attempt to build such a chain
6662 fails. An error in the implementation of this logic can mean that an
6663 attacker could cause certain checks on untrusted certificates to be
6664 bypassed, such as the CA flag, enabling them to use a valid leaf
6665 certificate to act as a CA and "issue" an invalid certificate.
6667 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6672 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6674 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6675 incompatibility in the handling of HMAC. The previous ABI has now been
6680 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6682 * Malformed ECParameters causes infinite loop
6684 When processing an ECParameters structure OpenSSL enters an infinite loop
6685 if the curve specified is over a specially malformed binary polynomial
6688 This can be used to perform denial of service against any
6689 system which processes public keys, certificate requests or
6690 certificates. This includes TLS clients and TLS servers with
6691 client authentication enabled.
6693 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6698 * Exploitable out-of-bounds read in X509_cmp_time
6700 X509_cmp_time does not properly check the length of the ASN1_TIME
6701 string and can read a few bytes out of bounds. In addition,
6702 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6705 An attacker can use this to craft malformed certificates and CRLs of
6706 various sizes and potentially cause a segmentation fault, resulting in
6707 a DoS on applications that verify certificates or CRLs. TLS clients
6708 that verify CRLs are affected. TLS clients and servers with client
6709 authentication enabled may be affected if they use custom verification
6712 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6713 independently by Hanno Böck.
6718 * PKCS7 crash with missing EnvelopedContent
6720 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6721 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6722 with missing content and trigger a NULL pointer dereference on parsing.
6724 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6725 structures from untrusted sources are affected. OpenSSL clients and
6726 servers are not affected.
6728 This issue was reported to OpenSSL by Michal Zalewski (Google).
6733 * CMS verify infinite loop with unknown hash function
6735 When verifying a signedData message the CMS code can enter an infinite loop
6736 if presented with an unknown hash function OID. This can be used to perform
6737 denial of service against any system which verifies signedData messages using
6739 This issue was reported to OpenSSL by Johannes Bauer.
6744 * Race condition handling NewSessionTicket
6746 If a NewSessionTicket is received by a multi-threaded client when attempting to
6747 reuse a previous ticket then a race condition can occur potentially leading to
6748 a double free of the ticket data.
6753 * Only support 256-bit or stronger elliptic curves with the
6754 'ecdh_auto' setting (server) or by default (client). Of supported
6755 curves, prefer P-256 (both).
6759 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6761 * ClientHello sigalgs DoS fix
6763 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6764 invalid signature algorithms extension a NULL pointer dereference will
6765 occur. This can be exploited in a DoS attack against the server.
6767 This issue was was reported to OpenSSL by David Ramos of Stanford
6771 *Stephen Henson and Matt Caswell*
6773 * Multiblock corrupted pointer fix
6775 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6776 feature only applies on 64 bit x86 architecture platforms that support AES
6777 NI instructions. A defect in the implementation of "multiblock" can cause
6778 OpenSSL's internal write buffer to become incorrectly set to NULL when
6779 using non-blocking IO. Typically, when the user application is using a
6780 socket BIO for writing, this will only result in a failed connection.
6781 However if some other BIO is used then it is likely that a segmentation
6782 fault will be triggered, thus enabling a potential DoS attack.
6784 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6789 * Segmentation fault in DTLSv1_listen fix
6791 The DTLSv1_listen function is intended to be stateless and processes the
6792 initial ClientHello from many peers. It is common for user code to loop
6793 over the call to DTLSv1_listen until a valid ClientHello is received with
6794 an associated cookie. A defect in the implementation of DTLSv1_listen means
6795 that state is preserved in the SSL object from one invocation to the next
6796 that can lead to a segmentation fault. Errors processing the initial
6797 ClientHello can trigger this scenario. An example of such an error could be
6798 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6801 This issue was reported to OpenSSL by Per Allansson.
6806 * Segmentation fault in ASN1_TYPE_cmp fix
6808 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6809 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6810 certificate signature algorithm consistency this can be used to crash any
6811 certificate verification operation and exploited in a DoS attack. Any
6812 application which performs certificate verification is vulnerable including
6813 OpenSSL clients and servers which enable client authentication.
6818 * Segmentation fault for invalid PSS parameters fix
6820 The signature verification routines will crash with a NULL pointer
6821 dereference if presented with an ASN.1 signature using the RSA PSS
6822 algorithm and invalid parameters. Since these routines are used to verify
6823 certificate signature algorithms this can be used to crash any
6824 certificate verification operation and exploited in a DoS attack. Any
6825 application which performs certificate verification is vulnerable including
6826 OpenSSL clients and servers which enable client authentication.
6828 This issue was was reported to OpenSSL by Brian Carpenter.
6833 * ASN.1 structure reuse memory corruption fix
6835 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6836 memory corruption via an invalid write. Such reuse is and has been
6837 strongly discouraged and is believed to be rare.
6839 Applications that parse structures containing CHOICE or ANY DEFINED BY
6840 components may be affected. Certificate parsing (d2i_X509 and related
6841 functions) are however not affected. OpenSSL clients and servers are
6847 * PKCS7 NULL pointer dereferences fix
6849 The PKCS#7 parsing code does not handle missing outer ContentInfo
6850 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6851 missing content and trigger a NULL pointer dereference on parsing.
6853 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6854 otherwise parse PKCS#7 structures from untrusted sources are
6855 affected. OpenSSL clients and servers are not affected.
6857 This issue was reported to OpenSSL by Michal Zalewski (Google).
6862 * DoS via reachable assert in SSLv2 servers fix
6864 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6865 servers that both support SSLv2 and enable export cipher suites by sending
6866 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6868 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6869 (OpenSSL development team).
6874 * Empty CKE with client auth and DHE fix
6876 If client auth is used then a server can seg fault in the event of a DHE
6877 ciphersuite being selected and a zero length ClientKeyExchange message
6878 being sent by the client. This could be exploited in a DoS attack.
6883 * Handshake with unseeded PRNG fix
6885 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6886 with an unseeded PRNG. The conditions are:
6887 - The client is on a platform where the PRNG has not been seeded
6888 automatically, and the user has not seeded manually
6889 - A protocol specific client method version has been used (i.e. not
6890 SSL_client_methodv23)
6891 - A ciphersuite is used that does not require additional random data from
6892 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6894 If the handshake succeeds then the client random that has been used will
6895 have been generated from a PRNG with insufficient entropy and therefore the
6896 output may be predictable.
6898 For example using the following command with an unseeded openssl will
6899 succeed on an unpatched platform:
6901 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6906 * Use After Free following d2i_ECPrivatekey error fix
6908 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6909 could cause a use after free condition. This, in turn, could cause a double
6910 free in several private key parsing functions (such as d2i_PrivateKey
6911 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6912 for applications that receive EC private keys from untrusted
6913 sources. This scenario is considered rare.
6915 This issue was discovered by the BoringSSL project and fixed in their
6921 * X509_to_X509_REQ NULL pointer deref fix
6923 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6924 the certificate key is invalid. This function is rarely used in practice.
6926 This issue was discovered by Brian Carpenter.
6931 * Removed the export ciphers from the DEFAULT ciphers
6935 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6937 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6938 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6939 So far those who have to target multiple platforms would compromise
6940 and argue that binary targeting say ARMv5 would still execute on
6941 ARMv8. "Universal" build resolves this compromise by providing
6942 near-optimal performance even on newer platforms.
6946 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6947 (other platforms pending).
6949 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6951 * Add support for the SignedCertificateTimestampList certificate and
6952 OCSP response extensions from RFC6962.
6956 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6957 for corner cases. (Certain input points at infinity could lead to
6958 bogus results, with non-infinity inputs mapped to infinity too.)
6962 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6963 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6964 common cases are optimized and there still is room for further
6965 improvements. Vector Permutation AES for Altivec is also added.
6969 * Add support for little-endian ppc64 Linux target.
6971 *Marcelo Cerri (IBM)*
6973 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6974 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6975 are optimized and there still is room for further improvements.
6976 Both 32- and 64-bit modes are supported.
6978 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6980 * Improved ARMv7 NEON support.
6984 * Support for SPARC Architecture 2011 crypto extensions, first
6985 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6986 SHA256/512, MD5, GHASH and modular exponentiation.
6988 *Andy Polyakov, David Miller*
6990 * Accelerated modular exponentiation for Intel processors, a.k.a.
6993 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6995 * Support for new and upcoming Intel processors, including AVX2,
6996 BMI and SHA ISA extensions. This includes additional "stitched"
6997 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7000 This work was sponsored by Intel Corp.
7004 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7005 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7006 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7010 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7011 this fixes a limitation in previous versions of OpenSSL.
7015 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7016 MGF1 digest and OAEP label.
7020 * Add EVP support for key wrapping algorithms, to avoid problems with
7021 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7022 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7023 algorithms and include tests cases.
7027 * Add functions to allocate and set the fields of an ECDSA_METHOD
7030 *Douglas E. Engert, Steve Henson*
7032 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7033 difference in days and seconds between two tm or ASN1_TIME structures.
7037 * Add -rev test option to s_server to just reverse order of characters
7038 received by client and send back to server. Also prints an abbreviated
7039 summary of the connection parameters.
7043 * New option -brief for s_client and s_server to print out a brief summary
7044 of connection parameters.
7048 * Add callbacks for arbitrary TLS extensions.
7050 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7052 * New option -crl_download in several openssl utilities to download CRLs
7053 from CRLDP extension in certificates.
7057 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7061 * New function X509_CRL_diff to generate a delta CRL from the difference
7062 of two full CRLs. Add support to "crl" utility.
7066 * New functions to set lookup_crls function and to retrieve
7067 X509_STORE from X509_STORE_CTX.
7071 * Print out deprecated issuer and subject unique ID fields in
7076 * Extend OCSP I/O functions so they can be used for simple general purpose
7077 HTTP as well as OCSP. New wrapper function which can be used to download
7078 CRLs using the OCSP API.
7082 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7086 * `SSL_CONF*` functions. These provide a common framework for application
7087 configuration using configuration files or command lines.
7091 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7092 message callback and prints the results. Needs compile time option
7093 "enable-ssl-trace". New options to s_client and s_server to enable
7098 * New ctrl and macro to retrieve supported points extensions.
7099 Print out extension in s_server and s_client.
7103 * New functions to retrieve certificate signature and signature
7108 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7113 * New Suite B modes for TLS code. These use and enforce the requirements
7114 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7115 only use Suite B curves. The Suite B modes can be set by using the
7116 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7120 * New chain verification flags for Suite B levels of security. Check
7121 algorithms are acceptable when flags are set in X509_verify_cert.
7125 * Make tls1_check_chain return a set of flags indicating checks passed
7126 by a certificate chain. Add additional tests to handle client
7127 certificates: checks for matching certificate type and issuer name
7132 * If an attempt is made to use a signature algorithm not in the peer
7133 preference list abort the handshake. If client has no suitable
7134 signature algorithms in response to a certificate request do not
7135 use the certificate.
7139 * If server EC tmp key is not in client preference list abort handshake.
7143 * Add support for certificate stores in CERT structure. This makes it
7144 possible to have different stores per SSL structure or one store in
7145 the parent SSL_CTX. Include distinct stores for certificate chain
7146 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7147 to build and store a certificate chain in CERT structure: returning
7148 an error if the chain cannot be built: this will allow applications
7149 to test if a chain is correctly configured.
7151 Note: if the CERT based stores are not set then the parent SSL_CTX
7152 store is used to retain compatibility with existing behaviour.
7156 * New function ssl_set_client_disabled to set a ciphersuite disabled
7157 mask based on the current session, check mask when sending client
7158 hello and checking the requested ciphersuite.
7162 * New ctrls to retrieve and set certificate types in a certificate
7163 request message. Print out received values in s_client. If certificate
7164 types is not set with custom values set sensible values based on
7165 supported signature algorithms.
7169 * Support for distinct client and server supported signature algorithms.
7173 * Add certificate callback. If set this is called whenever a certificate
7174 is required by client or server. An application can decide which
7175 certificate chain to present based on arbitrary criteria: for example
7176 supported signature algorithms. Add very simple example to s_server.
7177 This fixes many of the problems and restrictions of the existing client
7178 certificate callback: for example you can now clear an existing
7179 certificate and specify the whole chain.
7183 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7184 the certificate can be used for (if anything). Set valid_flags field
7185 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7186 to have similar checks in it.
7188 Add new "cert_flags" field to CERT structure and include a "strict mode".
7189 This enforces some TLS certificate requirements (such as only permitting
7190 certificate signature algorithms contained in the supported algorithms
7191 extension) which some implementations ignore: this option should be used
7192 with caution as it could cause interoperability issues.
7196 * Update and tidy signature algorithm extension processing. Work out
7197 shared signature algorithms based on preferences and peer algorithms
7198 and print them out in s_client and s_server. Abort handshake if no
7199 shared signature algorithms.
7203 * Add new functions to allow customised supported signature algorithms
7204 for SSL and SSL_CTX structures. Add options to s_client and s_server
7209 * New function SSL_certs_clear() to delete all references to certificates
7210 from an SSL structure. Before this once a certificate had been added
7211 it couldn't be removed.
7215 * Integrate hostname, email address and IP address checking with certificate
7216 verification. New verify options supporting checking in openssl utility.
7220 * Fixes and wildcard matching support to hostname and email checking
7221 functions. Add manual page.
7223 *Florian Weimer (Red Hat Product Security Team)*
7225 * New functions to check a hostname email or IP address against a
7226 certificate. Add options x509 utility to print results of checks against
7231 * Fix OCSP checking.
7233 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7235 * Initial experimental support for explicitly trusted non-root CAs.
7236 OpenSSL still tries to build a complete chain to a root but if an
7237 intermediate CA has a trust setting included that is used. The first
7238 setting is used: whether to trust (e.g., -addtrust option to the x509
7243 * Add -trusted_first option which attempts to find certificates in the
7244 trusted store even if an untrusted chain is also supplied.
7248 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7249 platform support for Linux and Android.
7253 * Support for linux-x32, ILP32 environment in x86_64 framework.
7257 * Experimental multi-implementation support for FIPS capable OpenSSL.
7258 When in FIPS mode the approved implementations are used as normal,
7259 when not in FIPS mode the internal unapproved versions are used instead.
7260 This means that the FIPS capable OpenSSL isn't forced to use the
7261 (often lower performance) FIPS implementations outside FIPS mode.
7265 * Transparently support X9.42 DH parameters when calling
7266 PEM_read_bio_DHparameters. This means existing applications can handle
7267 the new parameter format automatically.
7271 * Initial experimental support for X9.42 DH parameter format: mainly
7272 to support use of 'q' parameter for RFC5114 parameters.
7276 * Add DH parameters from RFC5114 including test data to dhtest.
7280 * Support for automatic EC temporary key parameter selection. If enabled
7281 the most preferred EC parameters are automatically used instead of
7282 hardcoded fixed parameters. Now a server just has to call:
7283 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7284 support ECDH and use the most appropriate parameters.
7288 * Enhance and tidy EC curve and point format TLS extension code. Use
7289 static structures instead of allocation if default values are used.
7290 New ctrls to set curves we wish to support and to retrieve shared curves.
7291 Print out shared curves in s_server. New options to s_server and s_client
7292 to set list of supported curves.
7296 * New ctrls to retrieve supported signature algorithms and
7297 supported curve values as an array of NIDs. Extend openssl utility
7298 to print out received values.
7302 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7303 between NIDs and the more common NIST names such as "P-256". Enhance
7304 ecparam utility and ECC method to recognise the NIST names for curves.
7308 * Enhance SSL/TLS certificate chain handling to support different
7309 chains for each certificate instead of one chain in the parent SSL_CTX.
7313 * Support for fixed DH ciphersuite client authentication: where both
7314 server and client use DH certificates with common parameters.
7318 * Support for fixed DH ciphersuites: those requiring DH server
7323 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7325 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7326 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7327 X509_CINF_get_signature were reverted post internal team review.
7332 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7334 * OCSP Status Request extension unbounded memory growth
7336 A malicious client can send an excessively large OCSP Status Request
7337 extension. If that client continually requests renegotiation, sending a
7338 large OCSP Status Request extension each time, then there will be unbounded
7339 memory growth on the server. This will eventually lead to a Denial Of
7340 Service attack through memory exhaustion. Servers with a default
7341 configuration are vulnerable even if they do not support OCSP. Builds using
7342 the "no-ocsp" build time option are not affected.
7344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7349 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7352 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7358 * OOB write in MDC2_Update()
7360 An overflow can occur in MDC2_Update() either if called directly or
7361 through the EVP_DigestUpdate() function using MDC2. If an attacker
7362 is able to supply very large amounts of input data after a previous
7363 call to EVP_EncryptUpdate() with a partial block then a length check
7364 can overflow resulting in a heap corruption.
7366 The amount of data needed is comparable to SIZE_MAX which is impractical
7369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7374 * Malformed SHA512 ticket DoS
7376 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7377 DoS attack where a malformed ticket will result in an OOB read which will
7380 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7381 a custom server callback and ticket lookup mechanism.
7383 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7388 * OOB write in BN_bn2dec()
7390 The function BN_bn2dec() does not check the return value of BN_div_word().
7391 This can cause an OOB write if an application uses this function with an
7392 overly large BIGNUM. This could be a problem if an overly large certificate
7393 or CRL is printed out from an untrusted source. TLS is not affected because
7394 record limits will reject an oversized certificate before it is parsed.
7396 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7401 * OOB read in TS_OBJ_print_bio()
7403 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7404 the total length the OID text representation would use and not the amount
7405 of data written. This will result in OOB reads when large OIDs are
7408 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7413 * Pointer arithmetic undefined behaviour
7415 Avoid some undefined pointer arithmetic
7417 A common idiom in the codebase is to check limits in the following manner:
7420 Where "p" points to some malloc'd data of SIZE bytes and
7423 "len" here could be from some externally supplied data (e.g. from a TLS
7426 The rules of C pointer arithmetic are such that "p + len" is only well
7427 defined where len <= SIZE. Therefore, the above idiom is actually
7428 undefined behaviour.
7430 For example this could cause problems if some malloc implementation
7431 provides an address for "p" such that "p + len" actually overflows for
7432 values of len that are too big and therefore p + len < limit.
7434 This issue was reported to OpenSSL by Guido Vranken
7439 * Constant time flag not preserved in DSA signing
7441 Operations in the DSA signing algorithm should run in constant time in
7442 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7443 implementation means that a non-constant time codepath is followed for
7444 certain operations. This has been demonstrated through a cache-timing
7445 attack to be sufficient for an attacker to recover the private DSA key.
7447 This issue was reported by César Pereida (Aalto University), Billy Brumley
7448 (Tampere University of Technology), and Yuval Yarom (The University of
7449 Adelaide and NICTA).
7454 * DTLS buffered message DoS
7456 In a DTLS connection where handshake messages are delivered out-of-order
7457 those messages that OpenSSL is not yet ready to process will be buffered
7458 for later use. Under certain circumstances, a flaw in the logic means that
7459 those messages do not get removed from the buffer even though the handshake
7460 has been completed. An attacker could force up to approx. 15 messages to
7461 remain in the buffer when they are no longer required. These messages will
7462 be cleared when the DTLS connection is closed. The default maximum size for
7463 a message is 100k. Therefore, the attacker could force an additional 1500k
7464 to be consumed per connection. By opening many simultaneous connections an
7465 attacker could cause a DoS attack through memory exhaustion.
7467 This issue was reported to OpenSSL by Quan Luo.
7472 * DTLS replay protection DoS
7474 A flaw in the DTLS replay attack protection mechanism means that records
7475 that arrive for future epochs update the replay protection "window" before
7476 the MAC for the record has been validated. This could be exploited by an
7477 attacker by sending a record for the next epoch (which does not have to
7478 decrypt or have a valid MAC), with a very large sequence number. This means
7479 that all subsequent legitimate packets are dropped causing a denial of
7480 service for a specific DTLS connection.
7482 This issue was reported to OpenSSL by the OCAP audit team.
7487 * Certificate message OOB reads
7489 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7490 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7491 theoretical DoS risk but this has not been observed in practice on common
7494 The messages affected are client certificate, client certificate request
7495 and server certificate. As a result the attack can only be performed
7496 against a client or a server which enables client authentication.
7498 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7503 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7505 * Prevent padding oracle in AES-NI CBC MAC check
7507 A MITM attacker can use a padding oracle attack to decrypt traffic
7508 when the connection uses an AES CBC cipher and the server support
7511 This issue was introduced as part of the fix for Lucky 13 padding
7512 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7513 constant time by making sure that always the same bytes are read and
7514 compared against either the MAC or padding bytes. But it no longer
7515 checked that there was enough data to have both the MAC and padding
7518 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7523 * Fix EVP_EncodeUpdate overflow
7525 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7526 Base64 encoding of binary data. If an attacker is able to supply very large
7527 amounts of input data then a length check can overflow resulting in a heap
7530 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7531 the `PEM_write_bio*` family of functions. These are mainly used within the
7532 OpenSSL command line applications, so any application which processes data
7533 from an untrusted source and outputs it as a PEM file should be considered
7534 vulnerable to this issue. User applications that call these APIs directly
7535 with large amounts of untrusted data may also be vulnerable.
7537 This issue was reported by Guido Vranken.
7542 * Fix EVP_EncryptUpdate overflow
7544 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7545 is able to supply very large amounts of input data after a previous call to
7546 EVP_EncryptUpdate() with a partial block then a length check can overflow
7547 resulting in a heap corruption. Following an analysis of all OpenSSL
7548 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7549 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7550 the first called function after an EVP_EncryptInit(), and therefore that
7551 specific call must be safe. The second form is where the length passed to
7552 EVP_EncryptUpdate() can be seen from the code to be some small value and
7553 therefore there is no possibility of an overflow. Since all instances are
7554 one of these two forms, it is believed that there can be no overflows in
7555 internal code due to this problem. It should be noted that
7556 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7557 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7558 of these calls have also been analysed too and it is believed there are no
7559 instances in internal usage where an overflow could occur.
7561 This issue was reported by Guido Vranken.
7566 * Prevent ASN.1 BIO excessive memory allocation
7568 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7569 a short invalid encoding can casuse allocation of large amounts of memory
7570 potentially consuming excessive resources or exhausting memory.
7572 Any application parsing untrusted data through d2i BIO functions is
7573 affected. The memory based functions such as d2i_X509() are *not* affected.
7574 Since the memory based functions are used by the TLS library, TLS
7575 applications are not affected.
7577 This issue was reported by Brian Carpenter.
7584 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7585 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7586 in arbitrary stack data being returned in the buffer.
7588 This issue was reported by Guido Vranken.
7593 * Modify behavior of ALPN to invoke callback after SNI/servername
7594 callback, such that updates to the SSL_CTX affect ALPN.
7598 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7603 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7604 methods are enabled and ssl2 is disabled the methods return NULL.
7608 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7610 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7611 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7612 provide any "EXPORT" or "LOW" strength ciphers.
7616 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7617 is by default disabled at build-time. Builds that are not configured with
7618 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7619 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7620 will need to explicitly call either of:
7622 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7624 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7626 as appropriate. Even if either of those is used, or the application
7627 explicitly uses the version-specific SSLv2_method() or its client and
7628 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7629 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7630 ciphers, and SSLv2 56-bit DES are no longer available.
7635 * Fix a double-free in DSA code
7637 A double free bug was discovered when OpenSSL parses malformed DSA private
7638 keys and could lead to a DoS attack or memory corruption for applications
7639 that receive DSA private keys from untrusted sources. This scenario is
7642 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7648 * Disable SRP fake user seed to address a server memory leak.
7650 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7652 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7653 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7654 was changed to ignore the "fake user" SRP seed, even if the seed
7657 Users should use SRP_VBASE_get1_by_user instead. Note that in
7658 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7659 also that even though configuring the SRP seed attempts to hide
7660 invalid usernames by continuing the handshake with fake
7661 credentials, this behaviour is not constant time and no strong
7662 guarantees are made that the handshake is indistinguishable from
7663 that of a valid user.
7668 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7670 In the BN_hex2bn function the number of hex digits is calculated using an
7671 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7672 large values of `i` this can result in `bn_expand` not allocating any
7673 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7674 field as NULL leading to a subsequent NULL ptr deref. For very large values
7675 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7676 In this case memory is allocated to the internal BIGNUM data field, but it
7677 is insufficiently sized leading to heap corruption. A similar issue exists
7678 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7679 is ever called by user applications with very large untrusted hex/dec data.
7680 This is anticipated to be a rare occurrence.
7682 All OpenSSL internal usage of these functions use data that is not expected
7683 to be untrusted, e.g. config file data or application command line
7684 arguments. If user developed applications generate config file data based
7685 on untrusted data then it is possible that this could also lead to security
7686 consequences. This is also anticipated to be rare.
7688 This issue was reported to OpenSSL by Guido Vranken.
7693 * Fix memory issues in `BIO_*printf` functions
7695 The internal `fmtstr` function used in processing a "%s" format string in
7696 the `BIO_*printf` functions could overflow while calculating the length of a
7697 string and cause an OOB read when printing very long strings.
7699 Additionally the internal `doapr_outch` function can attempt to write to an
7700 OOB memory location (at an offset from the NULL pointer) in the event of a
7701 memory allocation failure. In 1.0.2 and below this could be caused where
7702 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7703 could be in processing a very long "%s" format string. Memory leaks can
7706 The first issue may mask the second issue dependent on compiler behaviour.
7707 These problems could enable attacks where large amounts of untrusted data
7708 is passed to the `BIO_*printf` functions. If applications use these functions
7709 in this way then they could be vulnerable. OpenSSL itself uses these
7710 functions when printing out human-readable dumps of ASN.1 data. Therefore
7711 applications that print this data could be vulnerable if the data is from
7712 untrusted sources. OpenSSL command line applications could also be
7713 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7714 as command line arguments.
7716 Libssl is not considered directly vulnerable. Additionally certificates etc
7717 received via remote connections via libssl are also unlikely to be able to
7718 trigger these issues because of message size limits enforced within libssl.
7720 This issue was reported to OpenSSL Guido Vranken.
7725 * Side channel attack on modular exponentiation
7727 A side-channel attack was found which makes use of cache-bank conflicts on
7728 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7729 of RSA keys. The ability to exploit this issue is limited as it relies on
7730 an attacker who has control of code in a thread running on the same
7731 hyper-threaded core as the victim thread which is performing decryptions.
7733 This issue was reported to OpenSSL by Yuval Yarom, The University of
7734 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7735 Nadia Heninger, University of Pennsylvania with more information at
7736 <http://cachebleed.info>.
7741 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7742 if no keysize is specified with default_bits. This fixes an
7743 omission in an earlier change that changed all RSA/DSA key generation
7744 commands to use 2048 bits by default.
7748 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7750 * Protection for DH small subgroup attacks
7752 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7753 switched on by default and cannot be disabled. This could have some
7758 * SSLv2 doesn't block disabled ciphers
7760 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7761 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7762 been disabled, provided that the SSLv2 protocol was not also disabled via
7765 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7766 and Sebastian Schinzel.
7771 * Reject DH handshakes with parameters shorter than 1024 bits.
7775 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7777 * Certificate verify crash with missing PSS parameter
7779 The signature verification routines will crash with a NULL pointer
7780 dereference if presented with an ASN.1 signature using the RSA PSS
7781 algorithm and absent mask generation function parameter. Since these
7782 routines are used to verify certificate signature algorithms this can be
7783 used to crash any certificate verification operation and exploited in a
7784 DoS attack. Any application which performs certificate verification is
7785 vulnerable including OpenSSL clients and servers which enable client
7788 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7793 * X509_ATTRIBUTE memory leak
7795 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7796 memory. This structure is used by the PKCS#7 and CMS routines so any
7797 application which reads PKCS#7 or CMS data from untrusted sources is
7798 affected. SSL/TLS is not affected.
7800 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7806 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7807 This changes the decoding behaviour for some invalid messages,
7808 though the change is mostly in the more lenient direction, and
7809 legacy behaviour is preserved as much as possible.
7813 * In DSA_generate_parameters_ex, if the provided seed is too short,
7814 use a random seed, as already documented.
7816 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7818 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7820 * Alternate chains certificate forgery
7822 During certificate verification, OpenSSL will attempt to find an
7823 alternative certificate chain if the first attempt to build such a chain
7824 fails. An error in the implementation of this logic can mean that an
7825 attacker could cause certain checks on untrusted certificates to be
7826 bypassed, such as the CA flag, enabling them to use a valid leaf
7827 certificate to act as a CA and "issue" an invalid certificate.
7829 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7835 * Race condition handling PSK identify hint
7837 If PSK identity hints are received by a multi-threaded client then
7838 the values are wrongly updated in the parent SSL_CTX structure. This can
7839 result in a race condition potentially leading to a double free of the
7845 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7847 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7848 incompatibility in the handling of HMAC. The previous ABI has now been
7851 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7853 * Malformed ECParameters causes infinite loop
7855 When processing an ECParameters structure OpenSSL enters an infinite loop
7856 if the curve specified is over a specially malformed binary polynomial
7859 This can be used to perform denial of service against any
7860 system which processes public keys, certificate requests or
7861 certificates. This includes TLS clients and TLS servers with
7862 client authentication enabled.
7864 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7869 * Exploitable out-of-bounds read in X509_cmp_time
7871 X509_cmp_time does not properly check the length of the ASN1_TIME
7872 string and can read a few bytes out of bounds. In addition,
7873 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7876 An attacker can use this to craft malformed certificates and CRLs of
7877 various sizes and potentially cause a segmentation fault, resulting in
7878 a DoS on applications that verify certificates or CRLs. TLS clients
7879 that verify CRLs are affected. TLS clients and servers with client
7880 authentication enabled may be affected if they use custom verification
7883 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7884 independently by Hanno Böck.
7889 * PKCS7 crash with missing EnvelopedContent
7891 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7892 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7893 with missing content and trigger a NULL pointer dereference on parsing.
7895 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7896 structures from untrusted sources are affected. OpenSSL clients and
7897 servers are not affected.
7899 This issue was reported to OpenSSL by Michal Zalewski (Google).
7904 * CMS verify infinite loop with unknown hash function
7906 When verifying a signedData message the CMS code can enter an infinite loop
7907 if presented with an unknown hash function OID. This can be used to perform
7908 denial of service against any system which verifies signedData messages using
7910 This issue was reported to OpenSSL by Johannes Bauer.
7915 * Race condition handling NewSessionTicket
7917 If a NewSessionTicket is received by a multi-threaded client when attempting to
7918 reuse a previous ticket then a race condition can occur potentially leading to
7919 a double free of the ticket data.
7924 * Reject DH handshakes with parameters shorter than 768 bits.
7926 *Kurt Roeckx and Emilia Kasper*
7928 * dhparam: generate 2048-bit parameters by default.
7930 *Kurt Roeckx and Emilia Kasper*
7932 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7934 * Segmentation fault in ASN1_TYPE_cmp fix
7936 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7937 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7938 certificate signature algorithm consistency this can be used to crash any
7939 certificate verification operation and exploited in a DoS attack. Any
7940 application which performs certificate verification is vulnerable including
7941 OpenSSL clients and servers which enable client authentication.
7946 * ASN.1 structure reuse memory corruption fix
7948 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7949 memory corruption via an invalid write. Such reuse is and has been
7950 strongly discouraged and is believed to be rare.
7952 Applications that parse structures containing CHOICE or ANY DEFINED BY
7953 components may be affected. Certificate parsing (d2i_X509 and related
7954 functions) are however not affected. OpenSSL clients and servers are
7960 * PKCS7 NULL pointer dereferences fix
7962 The PKCS#7 parsing code does not handle missing outer ContentInfo
7963 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7964 missing content and trigger a NULL pointer dereference on parsing.
7966 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7967 otherwise parse PKCS#7 structures from untrusted sources are
7968 affected. OpenSSL clients and servers are not affected.
7970 This issue was reported to OpenSSL by Michal Zalewski (Google).
7975 * DoS via reachable assert in SSLv2 servers fix
7977 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7978 servers that both support SSLv2 and enable export cipher suites by sending
7979 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7981 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7982 (OpenSSL development team).
7987 * Use After Free following d2i_ECPrivatekey error fix
7989 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7990 could cause a use after free condition. This, in turn, could cause a double
7991 free in several private key parsing functions (such as d2i_PrivateKey
7992 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7993 for applications that receive EC private keys from untrusted
7994 sources. This scenario is considered rare.
7996 This issue was discovered by the BoringSSL project and fixed in their
8002 * X509_to_X509_REQ NULL pointer deref fix
8004 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8005 the certificate key is invalid. This function is rarely used in practice.
8007 This issue was discovered by Brian Carpenter.
8012 * Removed the export ciphers from the DEFAULT ciphers
8016 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8018 * Build fixes for the Windows and OpenVMS platforms
8020 *Matt Caswell and Richard Levitte*
8022 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8024 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8025 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8026 dereference. This could lead to a Denial Of Service attack. Thanks to
8027 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8032 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8033 dtls1_buffer_record function under certain conditions. In particular this
8034 could occur if an attacker sent repeated DTLS records with the same
8035 sequence number but for the next epoch. The memory leak could be exploited
8036 by an attacker in a Denial of Service attack through memory exhaustion.
8037 Thanks to Chris Mueller for reporting this issue.
8042 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8043 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8044 method would be set to NULL which could later result in a NULL pointer
8045 dereference. Thanks to Frank Schmirler for reporting this issue.
8050 * Abort handshake if server key exchange message is omitted for ephemeral
8053 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8054 reporting this issue.
8059 * Remove non-export ephemeral RSA code on client and server. This code
8060 violated the TLS standard by allowing the use of temporary RSA keys in
8061 non-export ciphersuites and could be used by a server to effectively
8062 downgrade the RSA key length used to a value smaller than the server
8063 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8064 INRIA or reporting this issue.
8069 * Fixed issue where DH client certificates are accepted without verification.
8070 An OpenSSL server will accept a DH certificate for client authentication
8071 without the certificate verify message. This effectively allows a client to
8072 authenticate without the use of a private key. This only affects servers
8073 which trust a client certificate authority which issues certificates
8074 containing DH keys: these are extremely rare and hardly ever encountered.
8075 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8081 * Ensure that the session ID context of an SSL is updated when its
8082 SSL_CTX is updated via SSL_set_SSL_CTX.
8084 The session ID context is typically set from the parent SSL_CTX,
8085 and can vary with the CTX.
8089 * Fix various certificate fingerprint issues.
8091 By using non-DER or invalid encodings outside the signed portion of a
8092 certificate the fingerprint can be changed without breaking the signature.
8093 Although no details of the signed portion of the certificate can be changed
8094 this can cause problems with some applications: e.g. those using the
8095 certificate fingerprint for blacklists.
8097 1. Reject signatures with non zero unused bits.
8099 If the BIT STRING containing the signature has non zero unused bits reject
8100 the signature. All current signature algorithms require zero unused bits.
8102 2. Check certificate algorithm consistency.
8104 Check the AlgorithmIdentifier inside TBS matches the one in the
8105 certificate signature. NB: this will result in signature failure
8106 errors for some broken certificates.
8108 Thanks to Konrad Kraszewski from Google for reporting this issue.
8110 3. Check DSA/ECDSA signatures use DER.
8112 Re-encode DSA/ECDSA signatures and compare with the original received
8113 signature. Return an error if there is a mismatch.
8115 This will reject various cases including garbage after signature
8116 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8117 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8118 (negative or with leading zeroes).
8120 Further analysis was conducted and fixes were developed by Stephen Henson
8121 of the OpenSSL core team.
8127 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8128 results on some platforms, including x86_64. This bug occurs at random
8129 with a very low probability, and is not known to be exploitable in any
8130 way, though its exact impact is difficult to determine. Thanks to Pieter
8131 Wuille (Blockstream) who reported this issue and also suggested an initial
8132 fix. Further analysis was conducted by the OpenSSL development team and
8133 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8134 the OpenSSL core team.
8139 * Do not resume sessions on the server if the negotiated protocol
8140 version does not match the session's version. Resuming with a different
8141 version, while not strictly forbidden by the RFC, is of questionable
8142 sanity and breaks all known clients.
8144 *David Benjamin, Emilia Käsper*
8146 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8147 early CCS messages during renegotiation. (Note that because
8148 renegotiation is encrypted, this early CCS was not exploitable.)
8152 * Tighten client-side session ticket handling during renegotiation:
8153 ensure that the client only accepts a session ticket if the server sends
8154 the extension anew in the ServerHello. Previously, a TLS client would
8155 reuse the old extension state and thus accept a session ticket if one was
8156 announced in the initial ServerHello.
8158 Similarly, ensure that the client requires a session ticket if one
8159 was advertised in the ServerHello. Previously, a TLS client would
8160 ignore a missing NewSessionTicket message.
8164 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8168 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8169 sends a carefully crafted handshake message, to cause OpenSSL to fail
8170 to free up to 64k of memory causing a memory leak. This could be
8171 exploited in a Denial Of Service attack. This issue affects OpenSSL
8172 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8173 whether SRTP is used or configured. Implementations of OpenSSL that
8174 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8176 The fix was developed by the OpenSSL team.
8181 * Session Ticket Memory Leak.
8183 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8184 integrity of that ticket is first verified. In the event of a session
8185 ticket integrity check failing, OpenSSL will fail to free memory
8186 causing a memory leak. By sending a large number of invalid session
8187 tickets an attacker could exploit this issue in a Denial Of Service
8193 * Build option no-ssl3 is incomplete.
8195 When OpenSSL is configured with "no-ssl3" as a build option, servers
8196 could accept and complete an SSL 3.0 handshake, and clients could be
8197 configured to send them.
8200 *Akamai and the OpenSSL team*
8202 * Add support for TLS_FALLBACK_SCSV.
8203 Client applications doing fallback retries should call
8204 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8207 *Adam Langley, Bodo Moeller*
8209 * Add additional DigestInfo checks.
8211 Re-encode DigestInto in DER and check against the original when
8212 verifying RSA signature: this will reject any improperly encoded
8213 DigestInfo structures.
8215 Note: this is a precautionary measure and no attacks are currently known.
8219 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8221 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8222 SRP code can be overrun an internal buffer. Add sanity check that
8223 g, A, B < N to SRP code.
8225 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8226 Group for discovering this issue.
8231 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8232 TLS 1.0 instead of higher protocol versions when the ClientHello message
8233 is badly fragmented. This allows a man-in-the-middle attacker to force a
8234 downgrade to TLS 1.0 even if both the server and the client support a
8235 higher protocol version, by modifying the client's TLS records.
8237 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8238 researching this issue.
8243 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8244 to a denial of service attack. A malicious server can crash the client
8245 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8246 ciphersuite and sending carefully crafted handshake messages.
8248 Thanks to Felix Gröbert (Google) for discovering and researching this
8254 * By sending carefully crafted DTLS packets an attacker could cause openssl
8255 to leak memory. This can be exploited through a Denial of Service attack.
8256 Thanks to Adam Langley for discovering and researching this issue.
8261 * An attacker can force openssl to consume large amounts of memory whilst
8262 processing DTLS handshake messages. This can be exploited through a
8263 Denial of Service attack.
8264 Thanks to Adam Langley for discovering and researching this issue.
8269 * An attacker can force an error condition which causes openssl to crash
8270 whilst processing DTLS packets due to memory being freed twice. This
8271 can be exploited through a Denial of Service attack.
8272 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8278 * If a multithreaded client connects to a malicious server using a resumed
8279 session and the server sends an ec point format extension it could write
8280 up to 255 bytes to freed memory.
8282 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8288 * A malicious server can crash an OpenSSL client with a null pointer
8289 dereference (read) by specifying an SRP ciphersuite even though it was not
8290 properly negotiated with the client. This can be exploited through a
8291 Denial of Service attack.
8293 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8294 discovering and researching this issue.
8299 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8300 X509_name_oneline, X509_name_print_ex et al. to leak some information
8301 from the stack. Applications may be affected if they echo pretty printing
8302 output to the attacker.
8304 Thanks to Ivan Fratric (Google) for discovering this issue.
8307 *Emilia Käsper, and Steve Henson*
8309 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8310 for corner cases. (Certain input points at infinity could lead to
8311 bogus results, with non-infinity inputs mapped to infinity too.)
8315 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8317 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8318 handshake can force the use of weak keying material in OpenSSL
8319 SSL/TLS clients and servers.
8321 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8322 researching this issue. ([CVE-2014-0224])
8324 *KIKUCHI Masashi, Steve Henson*
8326 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8327 OpenSSL DTLS client the code can be made to recurse eventually crashing
8330 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8333 *Imre Rad, Steve Henson*
8335 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8336 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8337 client or server. This is potentially exploitable to run arbitrary
8338 code on a vulnerable client or server.
8340 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8342 *Jüri Aedla, Steve Henson*
8344 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8345 are subject to a denial of service attack.
8347 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8348 this issue. ([CVE-2014-3470])
8350 *Felix Gröbert, Ivan Fratric, Steve Henson*
8352 * Harmonize version and its documentation. -f flag is used to display
8355 *mancha <mancha1@zoho.com>*
8357 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8358 in i2d_ECPrivateKey.
8360 *mancha <mancha1@zoho.com>*
8362 * Fix some double frees. These are not thought to be exploitable.
8364 *mancha <mancha1@zoho.com>*
8366 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8368 * A missing bounds check in the handling of the TLS heartbeat extension
8369 can be used to reveal up to 64k of memory to a connected client or
8372 Thanks for Neel Mehta of Google Security for discovering this bug and to
8373 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8374 preparing the fix ([CVE-2014-0160])
8376 *Adam Langley, Bodo Moeller*
8378 * Fix for the attack described in the paper "Recovering OpenSSL
8379 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8380 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8381 <http://eprint.iacr.org/2014/140>
8383 Thanks to Yuval Yarom and Naomi Benger for discovering this
8384 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8386 *Yuval Yarom and Naomi Benger*
8388 * TLS pad extension: draft-agl-tls-padding-03
8390 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8391 TLS client Hello record length value would otherwise be > 255 and
8392 less that 512 pad with a dummy extension containing zeroes so it
8393 is at least 512 bytes long.
8395 *Adam Langley, Steve Henson*
8397 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8399 * Fix for TLS record tampering bug. A carefully crafted invalid
8400 handshake could crash OpenSSL with a NULL pointer exception.
8401 Thanks to Anton Johansson for reporting this issues.
8404 * Keep original DTLS digest and encryption contexts in retransmission
8405 structures so we can use the previous session parameters if they need
8406 to be resent. ([CVE-2013-6450])
8410 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8411 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8412 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8413 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8414 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8415 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8417 *Rob Stradling, Adam Langley*
8419 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8421 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8422 supporting platforms or when small records were transferred.
8424 *Andy Polyakov, Steve Henson*
8426 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8428 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8430 This addresses the flaw in CBC record processing discovered by
8431 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8432 at: <http://www.isg.rhul.ac.uk/tls/>
8434 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8435 Security Group at Royal Holloway, University of London
8436 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8437 Emilia Käsper for the initial patch.
8440 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8442 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8443 ciphersuites which can be exploited in a denial of service attack.
8444 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8445 and detecting this bug and to Wolfgang Ettlinger
8446 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8451 * Return an error when checking OCSP signatures when key is NULL.
8452 This fixes a DoS attack. ([CVE-2013-0166])
8456 * Make openssl verify return errors.
8458 *Chris Palmer <palmer@google.com> and Ben Laurie*
8460 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8461 the right response is stapled. Also change SSL_get_certificate()
8462 so it returns the certificate actually sent.
8463 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8465 *Rob Stradling <rob.stradling@comodo.com>*
8467 * Fix possible deadlock when decoding public keys.
8471 * Don't use TLS 1.0 record version number in initial client hello
8476 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8478 * Sanity check record length before skipping explicit IV in TLS
8479 1.2, 1.1 and DTLS to fix DoS attack.
8481 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8482 fuzzing as a service testing platform.
8487 * Initialise tkeylen properly when encrypting CMS messages.
8488 Thanks to Solar Designer of Openwall for reporting this issue.
8492 * In FIPS mode don't try to use composite ciphers as they are not
8497 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8499 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8500 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8501 mean any application compiled against OpenSSL 1.0.0 headers setting
8502 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8503 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8504 0x10000000L Any application which was previously compiled against
8505 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8506 will need to be recompiled as a result. Letting be results in
8507 inability to disable specifically TLS 1.1 and in client context,
8508 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8512 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8513 disable just protocol X, but all protocols above X *if* there are
8514 protocols *below* X still enabled. In more practical terms it means
8515 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8516 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8517 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8522 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8524 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8525 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8526 in CRYPTO_realloc_clean.
8528 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8529 issue and to Adam Langley <agl@chromium.org> for fixing it.
8532 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8534 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8538 * Workarounds for some broken servers that "hang" if a client hello
8539 record length exceeds 255 bytes.
8541 1. Do not use record version number > TLS 1.0 in initial client
8542 hello: some (but not all) hanging servers will now work.
8543 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8544 the number of ciphers sent in the client hello. This should be
8545 set to an even number, such as 50, for example by passing:
8546 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8547 Most broken servers should now work.
8548 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8549 TLS 1.2 client support entirely.
8553 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8557 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8559 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8560 STRING form instead of a DigestInfo.
8564 * The format used for MDC2 RSA signatures is inconsistent between EVP
8565 and the RSA_sign/RSA_verify functions. This was made more apparent when
8566 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8567 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8568 the correct format in RSA_verify so both forms transparently work.
8572 * Some servers which support TLS 1.0 can choke if we initially indicate
8573 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8574 encrypted premaster secret. As a workaround use the maximum permitted
8575 client version in client hello, this should keep such servers happy
8576 and still work with previous versions of OpenSSL.
8580 * Add support for TLS/DTLS heartbeats.
8582 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8584 * Add support for SCTP.
8586 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8588 * Improved PRNG seeding for VOS.
8590 *Paul Green <Paul.Green@stratus.com>*
8592 * Extensive assembler packs updates, most notably:
8594 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8595 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8596 - x86_64: bit-sliced AES implementation;
8597 - ARM: NEON support, contemporary platforms optimizations;
8598 - s390x: z196 support;
8599 - `*`: GHASH and GF(2^m) multiplication implementations;
8603 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8604 (removal of unnecessary code)
8606 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8608 * Add TLS key material exporter from RFC 5705.
8612 * Add DTLS-SRTP negotiation from RFC 5764.
8616 * Add Next Protocol Negotiation,
8617 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8618 disabled with a no-npn flag to config or Configure. Code donated
8621 *Adam Langley <agl@google.com> and Ben Laurie*
8623 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8624 NIST-P256, NIST-P521, with constant-time single point multiplication on
8625 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8626 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8627 Code made available under Apache License version 2.0.
8629 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8630 line to include this in your build of OpenSSL, and run "make depend" (or
8631 "make update"). This enables the following EC_METHODs:
8633 EC_GFp_nistp224_method()
8634 EC_GFp_nistp256_method()
8635 EC_GFp_nistp521_method()
8637 EC_GROUP_new_by_curve_name() will automatically use these (while
8638 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8641 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8643 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8644 all platforms. Move ssize_t definition from e_os.h to the public
8645 header file e_os2.h as it now appears in public header file cms.h
8649 * New -sigopt option to the ca, req and x509 utilities. Additional
8650 signature parameters can be passed using this option and in
8655 * Add RSA PSS signing function. This will generate and set the
8656 appropriate AlgorithmIdentifiers for PSS based on those in the
8657 corresponding EVP_MD_CTX structure. No application support yet.
8661 * Support for companion algorithm specific ASN1 signing routines.
8662 New function ASN1_item_sign_ctx() signs a pre-initialised
8663 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8664 the appropriate parameters.
8668 * Add new algorithm specific ASN1 verification initialisation function
8669 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8670 handling will be the same no matter what EVP_PKEY_METHOD is used.
8671 Add a PSS handler to support verification of PSS signatures: checked
8672 against a number of sample certificates.
8676 * Add signature printing for PSS. Add PSS OIDs.
8678 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8680 * Add algorithm specific signature printing. An individual ASN1 method
8681 can now print out signatures instead of the standard hex dump.
8683 More complex signatures (e.g. PSS) can print out more meaningful
8684 information. Include DSA version that prints out the signature
8689 * Password based recipient info support for CMS library: implementing
8694 * Split password based encryption into PBES2 and PBKDF2 functions. This
8695 neatly separates the code into cipher and PBE sections and is required
8696 for some algorithms that split PBES2 into separate pieces (such as
8697 password based CMS).
8701 * Session-handling fixes:
8702 - Fix handling of connections that are resuming with a session ID,
8703 but also support Session Tickets.
8704 - Fix a bug that suppressed issuing of a new ticket if the client
8705 presented a ticket with an expired session.
8706 - Try to set the ticket lifetime hint to something reasonable.
8707 - Make tickets shorter by excluding irrelevant information.
8708 - On the client side, don't ignore renewed tickets.
8710 *Adam Langley, Bodo Moeller (Google)*
8712 * Fix PSK session representation.
8716 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8718 This work was sponsored by Intel.
8722 * Add GCM support to TLS library. Some custom code is needed to split
8723 the IV between the fixed (from PRF) and explicit (from TLS record)
8724 portions. This adds all GCM ciphersuites supported by RFC5288 and
8725 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8726 add a special AESGCM string for GCM only.
8730 * Expand range of ctrls for AES GCM. Permit setting invocation
8731 field on decrypt and retrieval of invocation field only on encrypt.
8735 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8736 As required by RFC5289 these ciphersuites cannot be used if for
8737 versions of TLS earlier than 1.2.
8741 * For FIPS capable OpenSSL interpret a NULL default public key method
8742 as unset and return the appropriate default but do *not* set the default.
8743 This means we can return the appropriate method in applications that
8744 switch between FIPS and non-FIPS modes.
8748 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8749 ENGINE is used then we cannot handle that in the FIPS module so we
8750 keep original code iff non-FIPS operations are allowed.
8754 * Add -attime option to openssl utilities.
8756 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8758 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8762 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8763 FIPS EC methods unconditionally for now.
8767 * New build option no-ec2m to disable characteristic 2 code.
8771 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8772 all cases can be covered as some introduce binary incompatibilities.
8776 * Redirect RSA operations to FIPS module including keygen,
8777 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8781 * Add similar low-level API blocking to ciphers.
8785 * low-level digest APIs are not approved in FIPS mode: any attempt
8786 to use these will cause a fatal error. Applications that *really* want
8787 to use them can use the `private_*` version instead.
8791 * Redirect cipher operations to FIPS module for FIPS builds.
8795 * Redirect digest operations to FIPS module for FIPS builds.
8799 * Update build system to add "fips" flag which will link in fipscanister.o
8800 for static and shared library builds embedding a signature if needed.
8804 * Output TLS supported curves in preference order instead of numerical
8805 order. This is currently hardcoded for the highest order curves first.
8806 This should be configurable so applications can judge speed vs strength.
8810 * Add TLS v1.2 server support for client authentication.
8814 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8819 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8820 FIPS modules versions.
8824 * Add TLS v1.2 client side support for client authentication. Keep cache
8825 of handshake records longer as we don't know the hash algorithm to use
8826 until after the certificate request message is received.
8830 * Initial TLS v1.2 client support. Add a default signature algorithms
8831 extension including all the algorithms we support. Parse new signature
8832 format in client key exchange. Relax some ECC signing restrictions for
8833 TLS v1.2 as indicated in RFC5246.
8837 * Add server support for TLS v1.2 signature algorithms extension. Switch
8838 to new signature format when needed using client digest preference.
8839 All server ciphersuites should now work correctly in TLS v1.2. No client
8840 support yet and no support for client certificates.
8844 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8845 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8846 ciphersuites. At present only RSA key exchange ciphersuites work with
8847 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8848 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8849 and version checking.
8853 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8854 with this defined it will not be affected by any changes to ssl internal
8855 structures. Add several utility functions to allow openssl application
8856 to work with OPENSSL_NO_SSL_INTERN defined.
8860 * A long standing patch to add support for SRP from EdelWeb (Peter
8861 Sylvester and Christophe Renou) was integrated.
8862 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8863 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8866 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8870 * Permit abbreviated handshakes when renegotiating using the function
8871 SSL_renegotiate_abbreviated().
8873 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8875 * Add call to ENGINE_register_all_complete() to
8876 ENGINE_load_builtin_engines(), so some implementations get used
8877 automatically instead of needing explicit application support.
8881 * Add support for TLS key exporter as described in RFC5705.
8883 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8885 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8886 a few changes are required:
8888 Add SSL_OP_NO_TLSv1_1 flag.
8889 Add TLSv1_1 methods.
8890 Update version checking logic to handle version 1.1.
8891 Add explicit IV handling (ported from DTLS code).
8892 Add command line options to s_client/s_server.
8899 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8901 * X509_ATTRIBUTE memory leak
8903 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8904 memory. This structure is used by the PKCS#7 and CMS routines so any
8905 application which reads PKCS#7 or CMS data from untrusted sources is
8906 affected. SSL/TLS is not affected.
8908 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8914 * Race condition handling PSK identify hint
8916 If PSK identity hints are received by a multi-threaded client then
8917 the values are wrongly updated in the parent SSL_CTX structure. This can
8918 result in a race condition potentially leading to a double free of the
8924 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8926 * Malformed ECParameters causes infinite loop
8928 When processing an ECParameters structure OpenSSL enters an infinite loop
8929 if the curve specified is over a specially malformed binary polynomial
8932 This can be used to perform denial of service against any
8933 system which processes public keys, certificate requests or
8934 certificates. This includes TLS clients and TLS servers with
8935 client authentication enabled.
8937 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8942 * Exploitable out-of-bounds read in X509_cmp_time
8944 X509_cmp_time does not properly check the length of the ASN1_TIME
8945 string and can read a few bytes out of bounds. In addition,
8946 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8949 An attacker can use this to craft malformed certificates and CRLs of
8950 various sizes and potentially cause a segmentation fault, resulting in
8951 a DoS on applications that verify certificates or CRLs. TLS clients
8952 that verify CRLs are affected. TLS clients and servers with client
8953 authentication enabled may be affected if they use custom verification
8956 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8957 independently by Hanno Böck.
8962 * PKCS7 crash with missing EnvelopedContent
8964 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8965 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8966 with missing content and trigger a NULL pointer dereference on parsing.
8968 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8969 structures from untrusted sources are affected. OpenSSL clients and
8970 servers are not affected.
8972 This issue was reported to OpenSSL by Michal Zalewski (Google).
8977 * CMS verify infinite loop with unknown hash function
8979 When verifying a signedData message the CMS code can enter an infinite loop
8980 if presented with an unknown hash function OID. This can be used to perform
8981 denial of service against any system which verifies signedData messages using
8983 This issue was reported to OpenSSL by Johannes Bauer.
8988 * Race condition handling NewSessionTicket
8990 If a NewSessionTicket is received by a multi-threaded client when attempting to
8991 reuse a previous ticket then a race condition can occur potentially leading to
8992 a double free of the ticket data.
8997 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8999 * Segmentation fault in ASN1_TYPE_cmp fix
9001 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9002 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9003 certificate signature algorithm consistency this can be used to crash any
9004 certificate verification operation and exploited in a DoS attack. Any
9005 application which performs certificate verification is vulnerable including
9006 OpenSSL clients and servers which enable client authentication.
9011 * ASN.1 structure reuse memory corruption fix
9013 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9014 memory corruption via an invalid write. Such reuse is and has been
9015 strongly discouraged and is believed to be rare.
9017 Applications that parse structures containing CHOICE or ANY DEFINED BY
9018 components may be affected. Certificate parsing (d2i_X509 and related
9019 functions) are however not affected. OpenSSL clients and servers are
9025 * PKCS7 NULL pointer dereferences fix
9027 The PKCS#7 parsing code does not handle missing outer ContentInfo
9028 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9029 missing content and trigger a NULL pointer dereference on parsing.
9031 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9032 otherwise parse PKCS#7 structures from untrusted sources are
9033 affected. OpenSSL clients and servers are not affected.
9035 This issue was reported to OpenSSL by Michal Zalewski (Google).
9040 * DoS via reachable assert in SSLv2 servers fix
9042 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9043 servers that both support SSLv2 and enable export cipher suites by sending
9044 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9046 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9047 (OpenSSL development team).
9052 * Use After Free following d2i_ECPrivatekey error fix
9054 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9055 could cause a use after free condition. This, in turn, could cause a double
9056 free in several private key parsing functions (such as d2i_PrivateKey
9057 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9058 for applications that receive EC private keys from untrusted
9059 sources. This scenario is considered rare.
9061 This issue was discovered by the BoringSSL project and fixed in their
9067 * X509_to_X509_REQ NULL pointer deref fix
9069 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9070 the certificate key is invalid. This function is rarely used in practice.
9072 This issue was discovered by Brian Carpenter.
9077 * Removed the export ciphers from the DEFAULT ciphers
9081 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9083 * Build fixes for the Windows and OpenVMS platforms
9085 *Matt Caswell and Richard Levitte*
9087 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9089 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9090 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9091 dereference. This could lead to a Denial Of Service attack. Thanks to
9092 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9097 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9098 dtls1_buffer_record function under certain conditions. In particular this
9099 could occur if an attacker sent repeated DTLS records with the same
9100 sequence number but for the next epoch. The memory leak could be exploited
9101 by an attacker in a Denial of Service attack through memory exhaustion.
9102 Thanks to Chris Mueller for reporting this issue.
9107 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9108 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9109 method would be set to NULL which could later result in a NULL pointer
9110 dereference. Thanks to Frank Schmirler for reporting this issue.
9115 * Abort handshake if server key exchange message is omitted for ephemeral
9118 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9119 reporting this issue.
9124 * Remove non-export ephemeral RSA code on client and server. This code
9125 violated the TLS standard by allowing the use of temporary RSA keys in
9126 non-export ciphersuites and could be used by a server to effectively
9127 downgrade the RSA key length used to a value smaller than the server
9128 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9129 INRIA or reporting this issue.
9134 * Fixed issue where DH client certificates are accepted without verification.
9135 An OpenSSL server will accept a DH certificate for client authentication
9136 without the certificate verify message. This effectively allows a client to
9137 authenticate without the use of a private key. This only affects servers
9138 which trust a client certificate authority which issues certificates
9139 containing DH keys: these are extremely rare and hardly ever encountered.
9140 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9146 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9147 results on some platforms, including x86_64. This bug occurs at random
9148 with a very low probability, and is not known to be exploitable in any
9149 way, though its exact impact is difficult to determine. Thanks to Pieter
9150 Wuille (Blockstream) who reported this issue and also suggested an initial
9151 fix. Further analysis was conducted by the OpenSSL development team and
9152 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9153 the OpenSSL core team.
9158 * Fix various certificate fingerprint issues.
9160 By using non-DER or invalid encodings outside the signed portion of a
9161 certificate the fingerprint can be changed without breaking the signature.
9162 Although no details of the signed portion of the certificate can be changed
9163 this can cause problems with some applications: e.g. those using the
9164 certificate fingerprint for blacklists.
9166 1. Reject signatures with non zero unused bits.
9168 If the BIT STRING containing the signature has non zero unused bits reject
9169 the signature. All current signature algorithms require zero unused bits.
9171 2. Check certificate algorithm consistency.
9173 Check the AlgorithmIdentifier inside TBS matches the one in the
9174 certificate signature. NB: this will result in signature failure
9175 errors for some broken certificates.
9177 Thanks to Konrad Kraszewski from Google for reporting this issue.
9179 3. Check DSA/ECDSA signatures use DER.
9181 Re-encode DSA/ECDSA signatures and compare with the original received
9182 signature. Return an error if there is a mismatch.
9184 This will reject various cases including garbage after signature
9185 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9186 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9187 (negative or with leading zeroes).
9189 Further analysis was conducted and fixes were developed by Stephen Henson
9190 of the OpenSSL core team.
9196 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9198 * Session Ticket Memory Leak.
9200 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9201 integrity of that ticket is first verified. In the event of a session
9202 ticket integrity check failing, OpenSSL will fail to free memory
9203 causing a memory leak. By sending a large number of invalid session
9204 tickets an attacker could exploit this issue in a Denial Of Service
9210 * Build option no-ssl3 is incomplete.
9212 When OpenSSL is configured with "no-ssl3" as a build option, servers
9213 could accept and complete an SSL 3.0 handshake, and clients could be
9214 configured to send them.
9217 *Akamai and the OpenSSL team*
9219 * Add support for TLS_FALLBACK_SCSV.
9220 Client applications doing fallback retries should call
9221 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9224 *Adam Langley, Bodo Moeller*
9226 * Add additional DigestInfo checks.
9228 Re-encode DigestInto in DER and check against the original when
9229 verifying RSA signature: this will reject any improperly encoded
9230 DigestInfo structures.
9232 Note: this is a precautionary measure and no attacks are currently known.
9236 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9238 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9239 to a denial of service attack. A malicious server can crash the client
9240 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9241 ciphersuite and sending carefully crafted handshake messages.
9243 Thanks to Felix Gröbert (Google) for discovering and researching this
9249 * By sending carefully crafted DTLS packets an attacker could cause openssl
9250 to leak memory. This can be exploited through a Denial of Service attack.
9251 Thanks to Adam Langley for discovering and researching this issue.
9256 * An attacker can force openssl to consume large amounts of memory whilst
9257 processing DTLS handshake messages. This can be exploited through a
9258 Denial of Service attack.
9259 Thanks to Adam Langley for discovering and researching this issue.
9264 * An attacker can force an error condition which causes openssl to crash
9265 whilst processing DTLS packets due to memory being freed twice. This
9266 can be exploited through a Denial of Service attack.
9267 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9273 * If a multithreaded client connects to a malicious server using a resumed
9274 session and the server sends an ec point format extension it could write
9275 up to 255 bytes to freed memory.
9277 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9283 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9284 X509_name_oneline, X509_name_print_ex et al. to leak some information
9285 from the stack. Applications may be affected if they echo pretty printing
9286 output to the attacker.
9288 Thanks to Ivan Fratric (Google) for discovering this issue.
9291 *Emilia Käsper, and Steve Henson*
9293 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9294 for corner cases. (Certain input points at infinity could lead to
9295 bogus results, with non-infinity inputs mapped to infinity too.)
9299 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9301 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9302 handshake can force the use of weak keying material in OpenSSL
9303 SSL/TLS clients and servers.
9305 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9306 researching this issue. ([CVE-2014-0224])
9308 *KIKUCHI Masashi, Steve Henson*
9310 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9311 OpenSSL DTLS client the code can be made to recurse eventually crashing
9314 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9317 *Imre Rad, Steve Henson*
9319 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9320 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9321 client or server. This is potentially exploitable to run arbitrary
9322 code on a vulnerable client or server.
9324 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9326 *Jüri Aedla, Steve Henson*
9328 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9329 are subject to a denial of service attack.
9331 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9332 this issue. ([CVE-2014-3470])
9334 *Felix Gröbert, Ivan Fratric, Steve Henson*
9336 * Harmonize version and its documentation. -f flag is used to display
9339 *mancha <mancha1@zoho.com>*
9341 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9342 in i2d_ECPrivateKey.
9344 *mancha <mancha1@zoho.com>*
9346 * Fix some double frees. These are not thought to be exploitable.
9348 *mancha <mancha1@zoho.com>*
9350 * Fix for the attack described in the paper "Recovering OpenSSL
9351 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9352 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9353 <http://eprint.iacr.org/2014/140>
9355 Thanks to Yuval Yarom and Naomi Benger for discovering this
9356 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9358 *Yuval Yarom and Naomi Benger*
9360 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9362 * Keep original DTLS digest and encryption contexts in retransmission
9363 structures so we can use the previous session parameters if they need
9364 to be resent. ([CVE-2013-6450])
9368 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9369 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9370 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9371 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9372 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9373 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9375 *Rob Stradling, Adam Langley*
9377 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9379 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9381 This addresses the flaw in CBC record processing discovered by
9382 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9383 at: <http://www.isg.rhul.ac.uk/tls/>
9385 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9386 Security Group at Royal Holloway, University of London
9387 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9388 Emilia Käsper for the initial patch.
9391 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9393 * Return an error when checking OCSP signatures when key is NULL.
9394 This fixes a DoS attack. ([CVE-2013-0166])
9398 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9399 the right response is stapled. Also change SSL_get_certificate()
9400 so it returns the certificate actually sent.
9401 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9402 (This is a backport)
9404 *Rob Stradling <rob.stradling@comodo.com>*
9406 * Fix possible deadlock when decoding public keys.
9410 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9412 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9415 * Sanity check record length before skipping explicit IV in DTLS
9418 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9419 fuzzing as a service testing platform.
9424 * Initialise tkeylen properly when encrypting CMS messages.
9425 Thanks to Solar Designer of Openwall for reporting this issue.
9429 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9431 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9432 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9433 in CRYPTO_realloc_clean.
9435 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9436 issue and to Adam Langley <agl@chromium.org> for fixing it.
9439 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9441 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9443 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9444 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9445 content decryption and always return the same error. Note: this attack
9446 needs on average 2^20 messages so it only affects automated senders. The
9447 old behaviour can be re-enabled in the CMS code by setting the
9448 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9449 an MMA defence is not necessary.
9450 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9451 this issue. ([CVE-2012-0884])
9455 * Fix CVE-2011-4619: make sure we really are receiving a
9456 client hello before rejecting multiple SGC restarts. Thanks to
9457 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9461 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9463 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9464 Thanks to Antonio Martin, Enterprise Secure Access Research and
9465 Development, Cisco Systems, Inc. for discovering this bug and
9466 preparing a fix. ([CVE-2012-0050])
9470 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9472 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9473 of the Vaudenay padding oracle attack on CBC mode encryption
9474 which enables an efficient plaintext recovery attack against
9475 the OpenSSL implementation of DTLS. Their attack exploits timing
9476 differences arising during decryption processing. A research
9477 paper describing this attack can be found at:
9478 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9479 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9480 Security Group at Royal Holloway, University of London
9481 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9482 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9483 for preparing the fix. ([CVE-2011-4108])
9485 *Robin Seggelmann, Michael Tuexen*
9487 * Clear bytes used for block padding of SSL 3.0 records.
9490 *Adam Langley (Google)*
9492 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9493 Kadianakis <desnacked@gmail.com> for discovering this issue and
9494 Adam Langley for preparing the fix. ([CVE-2011-4619])
9496 *Adam Langley (Google)*
9498 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9500 *Andrey Kulikov <amdeich@gmail.com>*
9502 * Prevent malformed RFC3779 data triggering an assertion failure.
9503 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9504 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9506 *Rob Austein <sra@hactrn.net>*
9508 * Improved PRNG seeding for VOS.
9510 *Paul Green <Paul.Green@stratus.com>*
9512 * Fix ssl_ciph.c set-up race.
9514 *Adam Langley (Google)*
9516 * Fix spurious failures in ecdsatest.c.
9518 *Emilia Käsper (Google)*
9520 * Fix the BIO_f_buffer() implementation (which was mixing different
9521 interpretations of the `..._len` fields).
9523 *Adam Langley (Google)*
9525 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9526 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9527 threads won't reuse the same blinding coefficients.
9529 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9530 lock to call BN_BLINDING_invert_ex, and avoids one use of
9531 BN_BLINDING_update for each BN_BLINDING structure (previously,
9532 the last update always remained unused).
9534 *Emilia Käsper (Google)*
9536 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9538 *Bob Buckholz (Google)*
9540 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9542 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9543 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9545 *Kaspar Brand <ossl@velox.ch>*
9547 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9548 for multi-threaded use of ECDH. ([CVE-2011-3210])
9550 *Adam Langley (Google)*
9552 * Fix x509_name_ex_d2i memory leak on bad inputs.
9556 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9557 signature public key algorithm by using OID xref utilities instead.
9558 Before this you could only use some ECC ciphersuites with SHA1 only.
9562 * Add protection against ECDSA timing attacks as mentioned in the paper
9563 by Billy Bob Brumley and Nicola Tuveri, see:
9564 <http://eprint.iacr.org/2011/232.pdf>
9566 *Billy Bob Brumley and Nicola Tuveri*
9568 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9570 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9572 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9574 * Fix bug in string printing code: if *any* escaping is enabled we must
9575 escape the escape character (backslash) or the resulting string is
9580 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9582 * Disable code workaround for ancient and obsolete Netscape browsers
9583 and servers: an attacker can use it in a ciphersuite downgrade attack.
9584 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9588 * Fixed J-PAKE implementation error, originally discovered by
9589 Sebastien Martini, further info and confirmation from Stefan
9590 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9594 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9596 * Fix extension code to avoid race conditions which can result in a buffer
9597 overrun vulnerability: resumed sessions must not be modified as they can
9598 be shared by multiple threads. CVE-2010-3864
9602 * Fix WIN32 build system to correctly link an ENGINE directory into
9607 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9609 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9612 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9614 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9616 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9617 context. The operation can be customised via the ctrl mechanism in
9618 case ENGINEs want to include additional functionality.
9622 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9626 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9627 output hashes compatible with older versions of OpenSSL.
9629 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9631 * Fix compression algorithm handling: if resuming a session use the
9632 compression algorithm of the resumed session instead of determining
9633 it from client hello again. Don't allow server to change algorithm.
9637 * Add load_crls() function to commands tidying load_certs() too. Add option
9638 to verify utility to allow additional CRLs to be included.
9642 * Update OCSP request code to permit adding custom headers to the request:
9643 some responders need this.
9647 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9650 *Julia Lawall <julia@diku.dk>*
9652 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9653 needlessly dereferenced structures, used obsolete functions and
9654 didn't handle all updated verify codes correctly.
9658 * Disable MD2 in the default configuration.
9662 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9663 indicate the initial BIO being pushed or popped. This makes it possible
9664 to determine whether the BIO is the one explicitly called or as a result
9665 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9666 it handles reference counts correctly and doesn't zero out the I/O bio
9667 when it is not being explicitly popped. WARNING: applications which
9668 included workarounds for the old buggy behaviour will need to be modified
9669 or they could free up already freed BIOs.
9673 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9674 renaming to all platforms (within the 0.9.8 branch, this was
9675 done conditionally on Netware platforms to avoid a name clash).
9677 *Guenter <lists@gknw.net>*
9679 * Add ECDHE and PSK support to DTLS.
9681 *Michael Tuexen <tuexen@fh-muenster.de>*
9683 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9688 * Add "missing" function EVP_MD_flags() (without this the only way to
9689 retrieve a digest flags is by accessing the structure directly. Update
9690 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9691 or cipher is registered as in the "from" argument. Print out all
9692 registered digests in the dgst usage message instead of manually
9693 attempting to work them out.
9697 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9698 this allows the use of compression and extensions. Change default cipher
9699 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9700 by default unless an application cipher string requests it.
9704 * Alter match criteria in PKCS12_parse(). It used to try to use local
9705 key ids to find matching certificates and keys but some PKCS#12 files
9706 don't follow the (somewhat unwritten) rules and this strategy fails.
9707 Now just gather all certificates together and the first private key
9708 then look for the first certificate that matches the key.
9712 * Support use of registered digest and cipher names for dgst and cipher
9713 commands instead of having to add each one as a special case. So now
9720 openssl dgst -sha256 foo
9722 and this works for ENGINE based algorithms too.
9726 * Update Gost ENGINE to support parameter files.
9728 *Victor B. Wagner <vitus@cryptocom.ru>*
9730 * Support GeneralizedTime in ca utility.
9732 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9734 * Enhance the hash format used for certificate directory links. The new
9735 form uses the canonical encoding (meaning equivalent names will work
9736 even if they aren't identical) and uses SHA1 instead of MD5. This form
9737 is incompatible with the older format and as a result c_rehash should
9738 be used to rebuild symbolic links.
9742 * Make PKCS#8 the default write format for private keys, replacing the
9743 traditional format. This form is standardised, more secure and doesn't
9744 include an implicit MD5 dependency.
9748 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9749 committed to OpenSSL should pass this lot as a minimum.
9753 * Add session ticket override functionality for use by EAP-FAST.
9755 *Jouni Malinen <j@w1.fi>*
9757 * Modify HMAC functions to return a value. Since these can be implemented
9758 in an ENGINE errors can occur.
9762 * Type-checked OBJ_bsearch_ex.
9766 * Type-checked OBJ_bsearch. Also some constification necessitated
9767 by type-checking. Still to come: TXT_DB, bsearch(?),
9768 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9773 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9774 seconds to a tm structure directly, instead of going through OS
9775 specific date routines. This avoids any issues with OS routines such
9776 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9777 and X509_time_adj_ex() to cover the extended range. The existing
9778 X509_time_adj() is still usable and will no longer have any date issues.
9782 * Delta CRL support. New use deltas option which will attempt to locate
9783 and search any appropriate delta CRLs available.
9785 This work was sponsored by Google.
9789 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9790 code and add additional score elements. Validate alternate CRL paths
9791 as part of the CRL checking and indicate a new error "CRL path validation
9792 error" in this case. Applications wanting additional details can use
9793 the verify callback and check the new "parent" field. If this is not
9794 NULL CRL path validation is taking place. Existing applications won't
9795 see this because it requires extended CRL support which is off by
9798 This work was sponsored by Google.
9802 * Support for freshest CRL extension.
9804 This work was sponsored by Google.
9808 * Initial indirect CRL support. Currently only supported in the CRLs
9809 passed directly and not via lookup. Process certificate issuer
9810 CRL entry extension and lookup CRL entries by bother issuer name
9811 and serial number. Check and process CRL issuer entry in IDP extension.
9813 This work was sponsored by Google.
9817 * Add support for distinct certificate and CRL paths. The CRL issuer
9818 certificate is validated separately in this case. Only enabled if
9819 an extended CRL support flag is set: this flag will enable additional
9820 CRL functionality in future.
9822 This work was sponsored by Google.
9826 * Add support for policy mappings extension.
9828 This work was sponsored by Google.
9832 * Fixes to pathlength constraint, self issued certificate handling,
9833 policy processing to align with RFC3280 and PKITS tests.
9835 This work was sponsored by Google.
9839 * Support for name constraints certificate extension. DN, email, DNS
9840 and URI types are currently supported.
9842 This work was sponsored by Google.
9846 * To cater for systems that provide a pointer-based thread ID rather
9847 than numeric, deprecate the current numeric thread ID mechanism and
9848 replace it with a structure and associated callback type. This
9849 mechanism allows a numeric "hash" to be extracted from a thread ID in
9850 either case, and on platforms where pointers are larger than 'long',
9851 mixing is done to help ensure the numeric 'hash' is usable even if it
9852 can't be guaranteed unique. The default mechanism is to use "&errno"
9853 as a pointer-based thread ID to distinguish between threads.
9855 Applications that want to provide their own thread IDs should now use
9856 CRYPTO_THREADID_set_callback() to register a callback that will call
9857 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9859 Note that ERR_remove_state() is now deprecated, because it is tied
9860 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9861 to free the current thread's error state should be replaced by
9862 ERR_remove_thread_state(NULL).
9864 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9865 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9866 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9867 application was previously providing a numeric thread callback that
9868 was inappropriate for distinguishing threads, then uniqueness might
9869 have been obtained with &errno that happened immediately in the
9870 intermediate development versions of OpenSSL; this is no longer the
9871 case, the numeric thread callback will now override the automatic use
9874 *Geoff Thorpe, with help from Bodo Moeller*
9876 * Initial support for different CRL issuing certificates. This covers a
9877 simple case where the self issued certificates in the chain exist and
9878 the real CRL issuer is higher in the existing chain.
9880 This work was sponsored by Google.
9884 * Removed effectively defunct crypto/store from the build.
9888 * Revamp of STACK to provide stronger type-checking. Still to come:
9889 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9890 ASN1_STRING, CONF_VALUE.
9894 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9895 RAM on SSL connections. This option can save about 34k per idle SSL.
9899 * Revamp of LHASH to provide stronger type-checking. Still to come:
9900 STACK, TXT_DB, bsearch, qsort.
9904 * Initial support for Cryptographic Message Syntax (aka CMS) based
9905 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9906 support for data, signedData, compressedData, digestedData and
9907 encryptedData, envelopedData types included. Scripts to check against
9908 RFC4134 examples draft and interop and consistency checks of many
9909 content types and variants.
9913 * Add options to enc utility to support use of zlib compression BIO.
9917 * Extend mk1mf to support importing of options and assembly language
9918 files from Configure script, currently only included in VC-WIN32.
9919 The assembly language rules can now optionally generate the source
9920 files from the associated perl scripts.
9924 * Implement remaining functionality needed to support GOST ciphersuites.
9925 Interop testing has been performed using CryptoPro implementations.
9927 *Victor B. Wagner <vitus@cryptocom.ru>*
9929 * s390x assembler pack.
9933 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9938 * Implement Opaque PRF Input TLS extension as specified in
9939 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9940 official specification yet and no extension type assignment by
9941 IANA exists, this extension (for now) will have to be explicitly
9942 enabled when building OpenSSL by providing the extension number
9943 to use. For example, specify an option
9945 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9947 to the "config" or "Configure" script to enable the extension,
9948 assuming extension number 0x9527 (which is a completely arbitrary
9949 and unofficial assignment based on the MD5 hash of the Internet
9950 Draft). Note that by doing so, you potentially lose
9951 interoperability with other TLS implementations since these might
9952 be using the same extension number for other purposes.
9954 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9955 opaque PRF input value to use in the handshake. This will create
9956 an internal copy of the length-'len' string at 'src', and will
9957 return non-zero for success.
9959 To get more control and flexibility, provide a callback function
9962 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9963 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9967 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9970 Callback function 'cb' will be called in handshakes, and is
9971 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9972 Argument 'arg' is for application purposes (the value as given to
9973 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9974 be provided to the callback function). The callback function
9975 has to return non-zero to report success: usually 1 to use opaque
9976 PRF input just if possible, or 2 to enforce use of the opaque PRF
9977 input. In the latter case, the library will abort the handshake
9978 if opaque PRF input is not successfully negotiated.
9980 Arguments 'peerinput' and 'len' given to the callback function
9981 will always be NULL and 0 in the case of a client. A server will
9982 see the client's opaque PRF input through these variables if
9983 available (NULL and 0 otherwise). Note that if the server
9984 provides an opaque PRF input, the length must be the same as the
9985 length of the client's opaque PRF input.
9987 Note that the callback function will only be called when creating
9988 a new session (session resumption can resume whatever was
9989 previously negotiated), and will not be called in SSL 2.0
9990 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9991 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9992 for applications that need to enforce opaque PRF input.
9996 * Update ssl code to support digests other than SHA1+MD5 for handshake
9999 *Victor B. Wagner <vitus@cryptocom.ru>*
10001 * Add RFC4507 support to OpenSSL. This includes the corrections in
10002 RFC4507bis. The encrypted ticket format is an encrypted encoded
10003 SSL_SESSION structure, that way new session features are automatically
10006 If a client application caches session in an SSL_SESSION structure
10007 support is transparent because tickets are now stored in the encoded
10010 The SSL_CTX structure automatically generates keys for ticket
10011 protection in servers so again support should be possible
10012 with no application modification.
10014 If a client or server wishes to disable RFC4507 support then the option
10015 SSL_OP_NO_TICKET can be set.
10017 Add a TLS extension debugging callback to allow the contents of any client
10018 or server extensions to be examined.
10020 This work was sponsored by Google.
10024 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10025 OpenSSL should now compile cleanly on gcc 4.2
10027 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10029 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10030 support including streaming MAC support: this is required for GOST
10031 ciphersuite support.
10033 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10035 * Add option -stream to use PKCS#7 streaming in smime utility. New
10036 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10037 to output in BER and PEM format.
10041 * Experimental support for use of HMAC via EVP_PKEY interface. This
10042 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10043 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10044 ENGINE support for HMAC keys which are unextractable. New -mac and
10045 -macopt options to dgst utility.
10049 * New option -sigopt to dgst utility. Update dgst to use
10050 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10051 alternative signing parameters such as X9.31 or PSS in the dgst
10056 * Change ssl_cipher_apply_rule(), the internal function that does
10057 the work each time a ciphersuite string requests enabling
10058 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10059 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10060 the order of disabled ciphersuites such that those ciphersuites
10061 that most recently went from enabled to disabled not only stay
10062 in order with respect to each other, but also have higher priority
10063 than other disabled ciphersuites the next time ciphersuites are
10066 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10067 the same ciphersuites as with "HIGH" alone, but in a specific
10068 order where the PSK ciphersuites come first (since they are the
10069 most recently disabled ciphersuites when "HIGH" is parsed).
10071 Also, change ssl_create_cipher_list() (using this new
10072 functionality) such that between otherwise identical
10073 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10078 * Change ssl_create_cipher_list() so that it automatically
10079 arranges the ciphersuites in reasonable order before starting
10080 to process the rule string. Thus, the definition for "DEFAULT"
10081 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10082 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10083 This makes it much easier to arrive at a reasonable default order
10084 in applications for which anonymous ciphers are OK (meaning
10085 that you can't actually use DEFAULT).
10087 *Bodo Moeller; suggested by Victor Duchovni*
10089 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10090 processing) into multiple integers instead of setting
10091 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10092 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10093 (These masks as well as the individual bit definitions are hidden
10094 away into the non-exported interface ssl/ssl_locl.h, so this
10095 change to the definition of the SSL_CIPHER structure shouldn't
10096 affect applications.) This give us more bits for each of these
10097 categories, so there is no longer a need to coagulate AES128 and
10098 AES256 into a single algorithm bit, and to coagulate Camellia128
10099 and Camellia256 into a single algorithm bit, which has led to all
10102 Thus, among other things, the kludge introduced in 0.9.7m and
10103 0.9.8e for masking out AES256 independently of AES128 or masking
10104 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10106 With the change, we also introduce new ciphersuite aliases that
10107 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10112 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10113 Use the leftmost N bytes of the signature input if the input is
10114 larger than the prime q (with N being the size in bytes of q).
10118 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10119 it yet and it is largely untested.
10123 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10127 * Initial incomplete changes to avoid need for function casts in OpenSSL
10128 some compilers (gcc 4.2 and later) reject their use. Safestack is
10129 reimplemented. Update ASN1 to avoid use of legacy functions.
10133 * Win32/64 targets are linked with Winsock2.
10137 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10138 to external functions. This can be used to increase CRL handling
10139 efficiency especially when CRLs are very large by (for example) storing
10140 the CRL revoked certificates in a database.
10144 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10145 new CRLs added to a directory can be used. New command line option
10146 -verify_return_error to s_client and s_server. This causes real errors
10147 to be returned by the verify callback instead of carrying on no matter
10148 what. This reflects the way a "real world" verify callback would behave.
10152 * GOST engine, supporting several GOST algorithms and public key formats.
10153 Kindly donated by Cryptocom.
10157 * Partial support for Issuing Distribution Point CRL extension. CRLs
10158 partitioned by DP are handled but no indirect CRL or reason partitioning
10159 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10160 selected via a scoring technique which handles IDP and AKID in CRLs.
10164 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10165 will ultimately be used for all verify operations: this will remove the
10166 X509_STORE dependency on certificate verification and allow alternative
10167 lookup methods. X509_STORE based implementations of these two callbacks.
10171 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10172 Modify get_crl() to find a valid (unexpired) CRL if possible.
10176 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10177 this would be called X509_CRL_cmp() but that name is already used by
10178 a function that just compares CRL issuer names. Cache several CRL
10179 extensions in X509_CRL structure and cache CRLDP in X509.
10183 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10184 this maps equivalent X509_NAME structures into a consistent structure.
10185 Name comparison can then be performed rapidly using memcmp().
10189 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10194 * Allow digests to supply their own micalg string for S/MIME type using
10195 the ctrl EVP_MD_CTRL_MICALG.
10199 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10200 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10201 ctrl. It can then customise the structure before and/or after signing
10206 * New function OBJ_add_sigid() to allow application defined signature OIDs
10207 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10208 to free up any added signature OIDs.
10212 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10213 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10214 digest and cipher tables. New options added to openssl utility:
10215 list-message-digest-algorithms and list-cipher-algorithms.
10219 * Change the array representation of binary polynomials: the list
10220 of degrees of non-zero coefficients is now terminated with -1.
10221 Previously it was terminated with 0, which was also part of the
10222 value; thus, the array representation was not applicable to
10223 polynomials where t^0 has coefficient zero. This change makes
10224 the array representation useful in a more general context.
10228 * Various modifications and fixes to SSL/TLS cipher string
10229 handling. For ECC, the code now distinguishes between fixed ECDH
10230 with RSA certificates on the one hand and with ECDSA certificates
10231 on the other hand, since these are separate ciphersuites. The
10232 unused code for Fortezza ciphersuites has been removed.
10234 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10235 (not "ECDHE"). For consistency with the code for DH
10236 certificates, use of ECDH certificates is now considered ECDH
10237 authentication, not RSA or ECDSA authentication (the latter is
10238 merely the CA's signing algorithm and not actively used in the
10241 The temporary ciphersuite alias "ECCdraft" is no longer
10242 available, and ECC ciphersuites are no longer excluded from "ALL"
10243 and "DEFAULT". The following aliases now exist for RFC 4492
10244 ciphersuites, most of these by analogy with the DH case:
10246 kECDHr - ECDH cert, signed with RSA
10247 kECDHe - ECDH cert, signed with ECDSA
10248 kECDH - ECDH cert (signed with either RSA or ECDSA)
10249 kEECDH - ephemeral ECDH
10250 ECDH - ECDH cert or ephemeral ECDH
10253 aECDSA - ECDSA cert
10256 AECDH - anonymous ECDH
10257 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10261 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10262 Use correct micalg parameters depending on digest(s) in signed message.
10266 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10267 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10271 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10272 an engine to register a method. Add ENGINE lookups for methods and
10273 functional reference processing.
10277 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10278 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10283 * New -resign option to smime utility. This adds one or more signers
10284 to an existing PKCS#7 signedData structure. Also -md option to use an
10285 alternative message digest algorithm for signing.
10289 * Tidy up PKCS#7 routines and add new functions to make it easier to
10290 create PKCS7 structures containing multiple signers. Update smime
10291 application to support multiple signers.
10295 * New -macalg option to pkcs12 utility to allow setting of an alternative
10300 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10301 Reorganize PBE internals to lookup from a static table using NIDs,
10302 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10303 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10304 PRF which will be automatically used with PBES2.
10308 * Replace the algorithm specific calls to generate keys in "req" with the
10313 * Update PKCS#7 enveloped data routines to use new API. This is now
10314 supported by any public key method supporting the encrypt operation. A
10315 ctrl is added to allow the public key algorithm to examine or modify
10316 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10321 * Add a ctrl to asn1 method to allow a public key algorithm to express
10322 a default digest type to use. In most cases this will be SHA1 but some
10323 algorithms (such as GOST) need to specify an alternative digest. The
10324 return value indicates how strong the preference is 1 means optional and
10325 2 is mandatory (that is it is the only supported type). Modify
10326 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10327 use the default md. Update openssl utilities to use the default digest
10328 type for signing if it is not explicitly indicated.
10332 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10333 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10334 signing method from the key type. This effectively removes the link
10335 between digests and public key types.
10339 * Add an OID cross reference table and utility functions. Its purpose is to
10340 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10341 rsaEncryption. This will allow some of the algorithm specific hackery
10342 needed to use the correct OID to be removed.
10346 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10347 structures for PKCS7_sign(). They are now set up by the relevant public
10352 * Add provisional EC pkey method with support for ECDSA and ECDH.
10356 * Add support for key derivation (agreement) in the API, DH method and
10361 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10362 public and private key formats. As a side effect these add additional
10363 command line functionality not previously available: DSA signatures can be
10364 generated and verified using pkeyutl and DH key support and generation in
10371 *Oliver Tappe <zooey@hirschkaefer.de>*
10373 * New make target "install_html_docs" installs HTML renditions of the
10376 *Oliver Tappe <zooey@hirschkaefer.de>*
10378 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10379 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10380 support key and parameter generation and add initial key generation
10381 functionality for RSA.
10385 * Add functions for main EVP_PKEY_method operations. The undocumented
10386 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10387 `EVP_PKEY_{encrypt,decrypt}_old`.
10391 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10392 key API, doesn't do much yet.
10396 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10397 public key algorithms. New option to openssl utility:
10398 "list-public-key-algorithms" to print out info.
10402 * Implement the Supported Elliptic Curves Extension for
10403 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10407 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10408 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10412 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10413 utilities such as rsa, dsa, dsaparam etc except they process any key
10418 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10419 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10420 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10425 * Initial support for pluggable public key ASN1.
10426 De-spaghettify the public key ASN1 handling. Move public and private
10427 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10428 algorithm specific handling to a single module within the relevant
10429 algorithm directory. Add functions to allow (near) opaque processing
10430 of public and private key structures.
10434 * Implement the Supported Point Formats Extension for
10435 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10439 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10440 for the psk identity [hint] and the psk callback functions to the
10441 SSL_SESSION, SSL and SSL_CTX structure.
10444 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10448 SSL_CTX_use_psk_identity_hint
10449 SSL_get_psk_identity_hint
10450 SSL_get_psk_identity
10451 SSL_use_psk_identity_hint
10453 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10455 * Add RFC 3161 compliant time stamp request creation, response generation
10456 and response verification functionality.
10458 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10460 * Add initial support for TLS extensions, specifically for the server_name
10461 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10462 have new members for a hostname. The SSL data structure has an
10463 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10464 stored in that context to allow for session resumption, even after the
10465 SSL has been switched to a new SSL_CTX in reaction to a client's
10466 server_name extension.
10468 New functions (subject to change):
10470 SSL_get_servername()
10471 SSL_get_servername_type()
10474 New CTRL codes and macros (subject to change):
10476 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10477 - SSL_CTX_set_tlsext_servername_callback()
10478 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10479 - SSL_CTX_set_tlsext_servername_arg()
10480 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10482 openssl s_client has a new '-servername ...' option.
10484 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10485 '-key2 ...', '-servername_fatal' (subject to change). This allows
10486 testing the HostName extension for a specific single hostname ('-cert'
10487 and '-key' remain fallbacks for handshakes without HostName
10488 negotiation). If the unrecognized_name alert has to be sent, this by
10489 default is a warning; it becomes fatal with the '-servername_fatal'
10492 *Peter Sylvester, Remy Allais, Christophe Renou*
10494 * Whirlpool hash implementation is added.
10498 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10499 bn(64,32). Because of instruction set limitations it doesn't have
10500 any negative impact on performance. This was done mostly in order
10501 to make it possible to share assembler modules, such as bn_mul_mont
10502 implementations, between 32- and 64-bit builds without hassle.
10506 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10507 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10512 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10513 dedicated Montgomery multiplication procedure, is introduced.
10514 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10515 "64-bit" performance on certain 32-bit targets.
10519 * New option SSL_OP_NO_COMP to disable use of compression selectively
10520 in SSL structures. New SSL ctrl to set maximum send fragment size.
10521 Save memory by setting the I/O buffer sizes dynamically instead of
10522 using the maximum available value.
10526 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10527 in addition to the text details.
10531 * Very, very preliminary EXPERIMENTAL support for printing of general
10532 ASN1 structures. This currently produces rather ugly output and doesn't
10533 handle several customised structures at all.
10537 * Integrated support for PVK file format and some related formats such
10538 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10539 these in the 'rsa' and 'dsa' utilities.
10543 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10547 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10548 place for the (very old) "NETSCAPE" format certificates which are now
10549 handled using new ASN1 code equivalents.
10553 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10554 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10555 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10559 * Modify CRL distribution points extension code to print out previously
10560 unsupported fields. Enhance extension setting code to allow setting of
10565 * Add print and set support for Issuing Distribution Point CRL extension.
10569 * Change 'Configure' script to enable Camellia by default.
10576 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10578 * When rejecting SSL/TLS records due to an incorrect version number, never
10579 update s->server with a new major version number. As of
10580 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10581 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10582 the previous behavior could result in a read attempt at NULL when
10583 receiving specific incorrect SSL/TLS records once record payload
10584 protection is active. ([CVE-2010-0740])
10586 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10588 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10589 could be crashed if the relevant tables were not present (e.g. chrooted).
10591 *Tomas Hoger <thoger@redhat.com>*
10593 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10595 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10597 *Martin Olsson, Neel Mehta*
10599 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10600 accommodate for stack sorting, always a write lock!).
10604 * On some versions of WIN32 Heap32Next is very slow. This can cause
10605 excessive delays in the RAND_poll(): over a minute. As a workaround
10606 include a time check in the inner Heap32Next loop too.
10610 * The code that handled flushing of data in SSL/TLS originally used the
10611 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10612 the problem outlined in PR#1949. The fix suggested there however can
10613 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10614 of Apache). So instead simplify the code to flush unconditionally.
10615 This should be fine since flushing with no data to flush is a no op.
10619 * Handle TLS versions 2.0 and later properly and correctly use the
10620 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10621 off ancient servers have a habit of sticking around for a while...
10625 * Modify compression code so it frees up structures without using the
10626 ex_data callbacks. This works around a problem where some applications
10627 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10628 restarting) then use compression (e.g. SSL with compression) later.
10629 This results in significant per-connection memory leaks and
10630 has caused some security issues including CVE-2008-1678 and
10635 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10636 change when encrypting or decrypting.
10640 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10641 connect and renegotiate with servers which do not support RI.
10642 Until RI is more widely deployed this option is enabled by default.
10646 * Add "missing" ssl ctrls to clear options and mode.
10650 * If client attempts to renegotiate and doesn't support RI respond with
10651 a no_renegotiation alert as required by RFC5746. Some renegotiating
10652 TLS clients will continue a connection gracefully when they receive
10653 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10654 waiting for a server hello which it will never receive. Now we treat a
10655 received no_renegotiation alert as a fatal error. This is because
10656 applications requesting a renegotiation might well expect it to succeed
10657 and would have no code in place to handle the server denying it so the
10658 only safe thing to do is to terminate the connection.
10662 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10663 peer supports secure renegotiation and 0 otherwise. Print out peer
10664 renegotiation support in s_client/s_server.
10668 * Replace the highly broken and deprecated SPKAC certification method with
10669 the updated NID creation version. This should correctly handle UTF8.
10673 * Implement RFC5746. Re-enable renegotiation but require the extension
10674 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10675 turns out to be a bad idea. It has been replaced by
10676 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10677 SSL_CTX_set_options(). This is really not recommended unless you
10678 know what you are doing.
10680 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10682 * Fixes to stateless session resumption handling. Use initial_ctx when
10683 issuing and attempting to decrypt tickets in case it has changed during
10684 servername handling. Use a non-zero length session ID when attempting
10685 stateless session resumption: this makes it possible to determine if
10686 a resumption has occurred immediately after receiving server hello
10687 (several places in OpenSSL subtly assume this) instead of later in
10692 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10693 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10694 fixes for a few places where the return code is not checked
10697 *Julia Lawall <julia@diku.dk>*
10699 * Add --strict-warnings option to Configure script to include devteam
10700 warnings in other configurations.
10704 * Add support for --libdir option and LIBDIR variable in makefiles. This
10705 makes it possible to install openssl libraries in locations which
10706 have names other than "lib", for example "/usr/lib64" which some
10709 *Steve Henson, based on patch from Jeremy Utley*
10711 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10712 X690 8.9.12 and can produce some misleading textual output of OIDs.
10714 *Steve Henson, reported by Dan Kaminsky*
10716 * Delete MD2 from algorithm tables. This follows the recommendation in
10717 several standards that it is not used in new applications due to
10718 several cryptographic weaknesses. For binary compatibility reasons
10719 the MD2 API is still compiled in by default.
10723 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10728 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10729 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10732 *Guenter <lists@gknw.net>*
10734 * Fix the server certificate chain building code to use X509_verify_cert(),
10735 it used to have an ad-hoc builder which was unable to cope with anything
10736 other than a simple chain.
10738 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10740 * Don't check self signed certificate signatures in X509_verify_cert()
10741 by default (a flag can override this): it just wastes time without
10742 adding any security. As a useful side effect self signed root CAs
10743 with non-FIPS digests are now usable in FIPS mode.
10747 * In dtls1_process_out_of_seq_message() the check if the current message
10748 is already buffered was missing. For every new message was memory
10749 allocated, allowing an attacker to perform an denial of service attack
10750 with sending out of seq handshake messages until there is no memory
10751 left. Additionally every future message was buffered, even if the
10752 sequence number made no sense and would be part of another handshake.
10753 So only messages with sequence numbers less than 10 in advance will be
10754 buffered. ([CVE-2009-1378])
10756 *Robin Seggelmann, discovered by Daniel Mentz*
10758 * Records are buffered if they arrive with a future epoch to be
10759 processed after finishing the corresponding handshake. There is
10760 currently no limitation to this buffer allowing an attacker to perform
10761 a DOS attack with sending records with future epochs until there is no
10762 memory left. This patch adds the pqueue_size() function to determine
10763 the size of a buffer and limits the record buffer to 100 entries.
10766 *Robin Seggelmann, discovered by Daniel Mentz*
10768 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10769 parent structure is freed. ([CVE-2009-1379])
10773 * Handle non-blocking I/O properly in SSL_shutdown() call.
10775 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10777 * Add `2.5.4.*` OIDs
10779 *Ilya O. <vrghost@gmail.com>*
10781 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10783 * Disable renegotiation completely - this fixes a severe security
10784 problem ([CVE-2009-3555]) at the cost of breaking all
10785 renegotiation. Renegotiation can be re-enabled by setting
10786 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10787 run-time. This is really not recommended unless you know what
10792 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10794 * Don't set val to NULL when freeing up structures, it is freed up by
10795 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10796 zeroing past the valid field. ([CVE-2009-0789])
10798 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10800 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10801 checked correctly. This would allow some invalid signed attributes to
10802 appear to verify correctly. ([CVE-2009-0591])
10804 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10806 * Reject UniversalString and BMPString types with invalid lengths. This
10807 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10808 a legal length. ([CVE-2009-0590])
10812 * Set S/MIME signing as the default purpose rather than setting it
10813 unconditionally. This allows applications to override it at the store
10818 * Permit restricted recursion of ASN1 strings. This is needed in practice
10819 to handle some structures.
10823 * Improve efficiency of mem_gets: don't search whole buffer each time
10826 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10828 * New -hex option for openssl rand.
10832 * Print out UTF8String and NumericString when parsing ASN1.
10836 * Support NumericString type for name components.
10840 * Allow CC in the environment to override the automatically chosen
10841 compiler. Note that nothing is done to ensure flags work with the
10846 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10848 * Properly check EVP_VerifyFinal() and similar return values
10851 *Ben Laurie, Bodo Moeller, Google Security Team*
10853 * Enable TLS extensions by default.
10857 * Allow the CHIL engine to be loaded, whether the application is
10858 multithreaded or not. (This does not release the developer from the
10859 obligation to set up the dynamic locking callbacks.)
10861 *Sander Temme <sander@temme.net>*
10863 * Use correct exit code if there is an error in dgst command.
10865 *Steve Henson; problem pointed out by Roland Dirlewanger*
10867 * Tweak Configure so that you need to say "experimental-jpake" to enable
10868 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10872 * Add experimental JPAKE support, including demo authentication in
10873 s_client and s_server.
10877 * Set the comparison function in v3_addr_canonize().
10879 *Rob Austein <sra@hactrn.net>*
10881 * Add support for XMPP STARTTLS in s_client.
10883 *Philip Paeps <philip@freebsd.org>*
10885 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10886 to ensure that even with this option, only ciphersuites in the
10887 server's preference list will be accepted. (Note that the option
10888 applies only when resuming a session, so the earlier behavior was
10889 just about the algorithm choice for symmetric cryptography.)
10893 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10895 * Fix NULL pointer dereference if a DTLS server received
10896 ChangeCipherSpec as first record ([CVE-2009-1386]).
10900 * Fix a state transition in s3_srvr.c and d1_srvr.c
10901 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10903 *Nagendra Modadugu*
10905 * The fix in 0.9.8c that supposedly got rid of unsafe
10906 double-checked locking was incomplete for RSA blinding,
10907 addressing just one layer of what turns out to have been
10908 doubly unsafe triple-checked locking.
10910 So now fix this for real by retiring the MONT_HELPER macro
10911 in crypto/rsa/rsa_eay.c.
10913 *Bodo Moeller; problem pointed out by Marius Schilder*
10915 * Various precautionary measures:
10917 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10919 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10920 (NB: This would require knowledge of the secret session ticket key
10921 to exploit, in which case you'd be SOL either way.)
10923 - Change bn_nist.c so that it will properly handle input BIGNUMs
10924 outside the expected range.
10926 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10929 *Neel Mehta, Bodo Moeller*
10931 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10932 the load fails. Useful for distros.
10934 *Ben Laurie and the FreeBSD team*
10936 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10940 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10944 * Expand ENGINE to support engine supplied SSL client certificate functions.
10946 This work was sponsored by Logica.
10950 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10951 keystores. Support for SSL/TLS client authentication too.
10952 Not compiled unless enable-capieng specified to Configure.
10954 This work was sponsored by Logica.
10958 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10959 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10960 attribute creation routines such as certificate requests and PKCS#12
10965 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10967 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10968 handshake which could lead to a client crash as found using the
10969 Codenomicon TLS test suite ([CVE-2008-1672])
10971 *Steve Henson, Mark Cox*
10973 * Fix double free in TLS server name extensions which could lead to
10974 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10978 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10980 Clear the error queue to ensure that error entries left from
10981 older function calls do not interfere with the correct operation.
10983 *Lutz Jaenicke, Erik de Castro Lopo*
10985 * Remove root CA certificates of commercial CAs:
10987 The OpenSSL project does not recommend any specific CA and does not
10988 have any policy with respect to including or excluding any CA.
10989 Therefore, it does not make any sense to ship an arbitrary selection
10990 of root CA certificates with the OpenSSL software.
10994 * RSA OAEP patches to fix two separate invalid memory reads.
10995 The first one involves inputs when 'lzero' is greater than
10996 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10997 before the beginning of from). The second one involves inputs where
10998 the 'db' section contains nothing but zeroes (there is a one-byte
10999 invalid read after the end of 'db').
11001 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11003 * Partial backport from 0.9.9-dev:
11005 Introduce bn_mul_mont (dedicated Montgomery multiplication
11006 procedure) as a candidate for BIGNUM assembler implementation.
11007 While 0.9.9-dev uses assembler for various architectures, only
11008 x86_64 is available by default here in the 0.9.8 branch, and
11009 32-bit x86 is available through a compile-time setting.
11011 To try the 32-bit x86 assembler implementation, use Configure
11012 option "enable-montasm" (which exists only for this backport).
11014 As "enable-montasm" for 32-bit x86 disclaims code stability
11015 anyway, in this constellation we activate additional code
11016 backported from 0.9.9-dev for further performance improvements,
11017 namely BN_from_montgomery_word. (To enable this otherwise,
11018 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11020 *Andy Polyakov (backport partially by Bodo Moeller)*
11022 * Add TLS session ticket callback. This allows an application to set
11023 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11024 values. This is useful for key rollover for example where several key
11025 sets may exist with different names.
11029 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11030 This was broken until now in 0.9.8 releases, such that the only way
11031 a registered ENGINE could be used (assuming it initialises
11032 successfully on the host) was to explicitly set it as the default
11033 for the relevant algorithms. This is in contradiction with 0.9.7
11034 behaviour and the documentation. With this fix, when an ENGINE is
11035 registered into a given algorithm's table of implementations, the
11036 'uptodate' flag is reset so that auto-discovery will be used next
11037 time a new context for that algorithm attempts to select an
11040 *Ian Lister (tweaked by Geoff Thorpe)*
11042 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11043 implementation in the following ways:
11045 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11048 Lack of BER streaming support means one pass streaming processing is
11049 only supported if data is detached: setting the streaming flag is
11050 ignored for embedded content.
11052 CMS support is disabled by default and must be explicitly enabled
11053 with the enable-cms configuration option.
11057 * Update the GMP engine glue to do direct copies between BIGNUM and
11058 mpz_t when openssl and GMP use the same limb size. Otherwise the
11059 existing "conversion via a text string export" trick is still used.
11061 *Paul Sheer <paulsheer@gmail.com>*
11063 * Zlib compression BIO. This is a filter BIO which compressed and
11064 uncompresses any data passed through it.
11068 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11069 RFC3394 compatible AES key wrapping.
11073 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11074 sets string data without copying. X509_ALGOR_set0() and
11075 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11076 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11077 from an X509_ATTRIBUTE structure optionally checking it occurs only
11078 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11083 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11084 to get the expected BN_FLG_CONSTTIME behavior.
11086 *Bodo Moeller (Google)*
11090 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11091 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11092 - added some more tests to do_tests.pl
11093 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11094 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11095 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11096 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11097 - various changes to netware.pl to enable gcc-cross builds on Win32
11099 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11100 - various changes to fix missing prototype warnings
11101 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11102 - added AES, WHIRLPOOL and CPUID assembler code to build files
11103 - added missing AES assembler make rules to mk1mf.pl
11104 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11106 *Guenter Knauf <eflash@gmx.net>*
11108 * Implement certificate status request TLS extension defined in RFC3546.
11109 A client can set the appropriate parameters and receive the encoded
11110 OCSP response via a callback. A server can query the supplied parameters
11111 and set the encoded OCSP response in the callback. Add simplified examples
11112 to s_client and s_server.
11116 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11118 * Fix various bugs:
11119 + Binary incompatibility of ssl_ctx_st structure
11120 + DTLS interoperation with non-compliant servers
11121 + Don't call get_session_cb() without proposed session
11122 + Fix ia64 assembler code
11124 *Andy Polyakov, Steve Henson*
11126 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11128 * DTLS Handshake overhaul. There were longstanding issues with
11129 OpenSSL DTLS implementation, which were making it impossible for
11130 RFC 4347 compliant client to communicate with OpenSSL server.
11131 Unfortunately just fixing these incompatibilities would "cut off"
11132 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11133 server keeps tolerating non RFC compliant syntax. The opposite is
11134 not true, 0.9.8f client can not communicate with earlier server.
11135 This update even addresses CVE-2007-4995.
11139 * Changes to avoid need for function casts in OpenSSL: some compilers
11140 (gcc 4.2 and later) reject their use.
11141 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11144 * Add RFC4507 support to OpenSSL. This includes the corrections in
11145 RFC4507bis. The encrypted ticket format is an encrypted encoded
11146 SSL_SESSION structure, that way new session features are automatically
11149 If a client application caches session in an SSL_SESSION structure
11150 support is transparent because tickets are now stored in the encoded
11153 The SSL_CTX structure automatically generates keys for ticket
11154 protection in servers so again support should be possible
11155 with no application modification.
11157 If a client or server wishes to disable RFC4507 support then the option
11158 SSL_OP_NO_TICKET can be set.
11160 Add a TLS extension debugging callback to allow the contents of any client
11161 or server extensions to be examined.
11163 This work was sponsored by Google.
11167 * Add initial support for TLS extensions, specifically for the server_name
11168 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11169 have new members for a hostname. The SSL data structure has an
11170 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11171 stored in that context to allow for session resumption, even after the
11172 SSL has been switched to a new SSL_CTX in reaction to a client's
11173 server_name extension.
11175 New functions (subject to change):
11177 SSL_get_servername()
11178 SSL_get_servername_type()
11181 New CTRL codes and macros (subject to change):
11183 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11184 - SSL_CTX_set_tlsext_servername_callback()
11185 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11186 - SSL_CTX_set_tlsext_servername_arg()
11187 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11189 openssl s_client has a new '-servername ...' option.
11191 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11192 '-key2 ...', '-servername_fatal' (subject to change). This allows
11193 testing the HostName extension for a specific single hostname ('-cert'
11194 and '-key' remain fallbacks for handshakes without HostName
11195 negotiation). If the unrecognized_name alert has to be sent, this by
11196 default is a warning; it becomes fatal with the '-servername_fatal'
11199 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11201 * Add AES and SSE2 assembly language support to VC++ build.
11205 * Mitigate attack on final subtraction in Montgomery reduction.
11209 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11210 (which previously caused an internal error).
11214 * Squeeze another 10% out of IGE mode when in != out.
11218 * AES IGE mode speedup.
11220 *Dean Gaudet (Google)*
11222 * Add the Korean symmetric 128-bit cipher SEED (see
11223 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11224 add SEED ciphersuites from RFC 4162:
11226 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11227 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11228 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11229 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11231 To minimize changes between patchlevels in the OpenSSL 0.9.8
11232 series, SEED remains excluded from compilation unless OpenSSL
11233 is configured with 'enable-seed'.
11235 *KISA, Bodo Moeller*
11237 * Mitigate branch prediction attacks, which can be practical if a
11238 single processor is shared, allowing a spy process to extract
11239 information. For detailed background information, see
11240 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11241 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11242 and Necessary Software Countermeasures"). The core of the change
11243 are new versions BN_div_no_branch() and
11244 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11245 respectively, which are slower, but avoid the security-relevant
11246 conditional branches. These are automatically called by BN_div()
11247 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11248 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11249 remove a conditional branch.
11251 BN_FLG_CONSTTIME is the new name for the previous
11252 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11253 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11254 in the exponent causes BN_mod_exp_mont() to use the alternative
11255 implementation in BN_mod_exp_mont_consttime().) The old name
11256 remains as a deprecated alias.
11258 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11259 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11260 constant-time implementations for more than just exponentiation.
11261 Here too the old name is kept as a deprecated alias.
11263 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11264 the BN_BLINDING structure gets an independent copy of the
11265 modulus. This means that the previous `BIGNUM *m` argument to
11266 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11267 essentially becomes `const BIGNUM *m`, although we can't actually
11268 change this in the header file before 0.9.9. It allows
11269 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11270 enable BN_FLG_CONSTTIME.
11272 *Matthew D Wood (Intel Corp)*
11274 * In the SSL/TLS server implementation, be strict about session ID
11275 context matching (which matters if an application uses a single
11276 external cache for different purposes). Previously,
11277 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11278 set. This did ensure strict client verification, but meant that,
11279 with applications using a single external cache for quite
11280 different requirements, clients could circumvent ciphersuite
11281 restrictions for a given session ID context by starting a session
11282 in a different context.
11286 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11287 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11288 authentication-only ciphersuites.
11292 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11293 not complete and could lead to a possible single byte overflow
11294 ([CVE-2007-5135]) [Ben Laurie]
11296 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11298 * Since AES128 and AES256 (and similarly Camellia128 and
11299 Camellia256) share a single mask bit in the logic of
11300 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11301 kludge to work properly if AES128 is available and AES256 isn't
11302 (or if Camellia128 is available and Camellia256 isn't).
11306 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11307 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11308 When a point or a seed is encoded in a BIT STRING, we need to
11309 prevent the removal of trailing zero bits to get the proper DER
11310 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11311 of a NamedBitList, for which trailing 0 bits need to be removed.)
11315 * Have SSL/TLS server implementation tolerate "mismatched" record
11316 protocol version while receiving ClientHello even if the
11317 ClientHello is fragmented. (The server can't insist on the
11318 particular protocol version it has chosen before the ServerHello
11319 message has informed the client about his choice.)
11323 * Add RFC 3779 support.
11325 *Rob Austein for ARIN, Ben Laurie*
11327 * Load error codes if they are not already present instead of using a
11328 static variable. This allows them to be cleanly unloaded and reloaded.
11329 Improve header file function name parsing.
11333 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11334 or CAPABILITY handshake as required by RFCs.
11336 *Goetz Babin-Ebell*
11338 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11340 * Introduce limits to prevent malicious keys being able to
11341 cause a denial of service. ([CVE-2006-2940])
11343 *Steve Henson, Bodo Moeller*
11345 * Fix ASN.1 parsing of certain invalid structures that can result
11346 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11348 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11349 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11351 * Fix SSL client code which could crash if connecting to a
11352 malicious SSLv2 server. ([CVE-2006-4343])
11354 *Tavis Ormandy and Will Drewry, Google Security Team*
11356 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11357 match only those. Before that, "AES256-SHA" would be interpreted
11358 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11359 the same strength classification in 0.9.7h) as we currently only
11360 have a single AES bit in the ciphersuite description bitmap.
11361 That change, however, also applied to ciphersuite strings such as
11362 "RC4-MD5" that intentionally matched multiple ciphersuites --
11363 namely, SSL 2.0 ciphersuites in addition to the more common ones
11364 from SSL 3.0/TLS 1.0.
11366 So we change the selection algorithm again: Naming an explicit
11367 ciphersuite selects this one ciphersuite, and any other similar
11368 ciphersuite (same bitmap) from *other* protocol versions.
11369 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11370 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11372 Since SSL 2.0 does not have any ciphersuites for which the
11373 128/256 bit distinction would be relevant, this works for now.
11374 The proper fix will be to use different bits for AES128 and
11375 AES256, which would have avoided the problems from the beginning;
11376 however, bits are scarce, so we can only do this in a new release
11377 (not just a patchlevel) when we can change the SSL_CIPHER
11378 definition to split the single 'unsigned long mask' bitmap into
11379 multiple values to extend the available space.
11383 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11385 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11386 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11388 * Add AES IGE and biIGE modes.
11392 * Change the Unix randomness entropy gathering to use poll() when
11393 possible instead of select(), since the latter has some
11394 undesirable limitations.
11396 *Darryl Miles via Richard Levitte and Bodo Moeller*
11398 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11399 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11400 cannot be implicitly activated as part of, e.g., the "AES" alias.
11401 However, please upgrade to OpenSSL 0.9.9[-dev] for
11402 non-experimental use of the ECC ciphersuites to get TLS extension
11403 support, which is required for curve and point format negotiation
11404 to avoid potential handshake problems.
11408 * Disable rogue ciphersuites:
11410 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11411 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11412 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11414 The latter two were purportedly from
11415 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11418 Also deactivate the remaining ciphersuites from
11419 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11420 unofficial, and the ID has long expired.
11424 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11425 dual-core machines) and other potential thread-safety issues.
11429 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11430 versions), which is now available for royalty-free use
11431 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11432 Also, add Camellia TLS ciphersuites from RFC 4132.
11434 To minimize changes between patchlevels in the OpenSSL 0.9.8
11435 series, Camellia remains excluded from compilation unless OpenSSL
11436 is configured with 'enable-camellia'.
11440 * Disable the padding bug check when compression is in use. The padding
11441 bug check assumes the first packet is of even length, this is not
11442 necessarily true if compression is enabled and can result in false
11443 positives causing handshake failure. The actual bug test is ancient
11444 code so it is hoped that implementations will either have fixed it by
11445 now or any which still have the bug do not support compression.
11449 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11451 * When applying a cipher rule check to see if string match is an explicit
11452 cipher suite and only match that one cipher suite if it is.
11456 * Link in manifests for VC++ if needed.
11458 *Austin Ziegler <halostatue@gmail.com>*
11460 * Update support for ECC-based TLS ciphersuites according to
11461 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11462 TLS extensions, which are supported starting with the 0.9.9
11463 branch, not in the OpenSSL 0.9.8 branch).
11467 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11468 opaque EVP_CIPHER_CTX handling.
11472 * Fixes and enhancements to zlib compression code. We now only use
11473 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11474 to conform with the standards mentioned here:
11475 <http://www.zlib.net/DLL_FAQ.txt>
11476 Static zlib linking now works on Windows and the new --with-zlib-include
11477 --with-zlib-lib options to Configure can be used to supply the location
11478 of the headers and library. Gracefully handle case where zlib library
11483 * Several fixes and enhancements to the OID generation code. The old code
11484 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11485 handle numbers larger than ULONG_MAX, truncated printing and had a
11486 non standard OBJ_obj2txt() behaviour.
11490 * Add support for building of engines under engine/ as shared libraries
11491 under VC++ build system.
11495 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11496 Hopefully, we will not see any false combination of paths any more.
11500 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11502 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11503 (part of SSL_OP_ALL). This option used to disable the
11504 countermeasure against man-in-the-middle protocol-version
11505 rollback in the SSL 2.0 server implementation, which is a bad
11506 idea. ([CVE-2005-2969])
11508 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11509 for Information Security, National Institute of Advanced Industrial
11510 Science and Technology [AIST], Japan)*
11512 * Add two function to clear and return the verify parameter flags.
11516 * Keep cipherlists sorted in the source instead of sorting them at
11517 runtime, thus removing the need for a lock.
11521 * Avoid some small subgroup attacks in Diffie-Hellman.
11523 *Nick Mathewson and Ben Laurie*
11525 * Add functions for well-known primes.
11529 * Extended Windows CE support.
11531 *Satoshi Nakamura and Andy Polyakov*
11533 * Initialize SSL_METHOD structures at compile time instead of during
11534 runtime, thus removing the need for a lock.
11538 * Make PKCS7_decrypt() work even if no certificate is supplied by
11539 attempting to decrypt each encrypted key in turn. Add support to
11544 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11546 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11549 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11553 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11554 key into the same file any more.
11558 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11562 * Add -utf8 command line and config file option to 'ca'.
11564 *Stefan <stf@udoma.org*
11566 * Removed the macro des_crypt(), as it seems to conflict with some
11567 libraries. Use DES_crypt().
11571 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11572 involves renaming the source and generated shared-libs for
11573 both. The engines will accept the corrected or legacy ids
11574 ('ncipher' and '4758_cca' respectively) when binding. NB,
11575 this only applies when building 'shared'.
11577 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11579 * Add attribute functions to EVP_PKEY structure. Modify
11580 PKCS12_create() to recognize a CSP name attribute and
11581 use it. Make -CSP option work again in pkcs12 utility.
11585 * Add new functionality to the bn blinding code:
11586 - automatic re-creation of the BN_BLINDING parameters after
11587 a fixed number of uses (currently 32)
11588 - add new function for parameter creation
11589 - introduce flags to control the update behaviour of the
11590 BN_BLINDING parameters
11591 - hide BN_BLINDING structure
11592 Add a second BN_BLINDING slot to the RSA structure to improve
11593 performance when a single RSA object is shared among several
11598 * Add support for DTLS.
11600 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11602 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11603 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11607 * Remove buggy and incomplete DH cert support from
11608 ssl/ssl_rsa.c and ssl/s3_both.c
11612 * Use SHA-1 instead of MD5 as the default digest algorithm for
11613 the `apps/openssl` commands.
11617 * Compile clean with "-Wall -Wmissing-prototypes
11618 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11619 DEBUG_SAFESTACK must also be set.
11623 * Change ./Configure so that certain algorithms can be disabled by default.
11624 The new counterpiece to "no-xxx" is "enable-xxx".
11626 The patented RC5 and MDC2 algorithms will now be disabled unless
11627 "enable-rc5" and "enable-mdc2", respectively, are specified.
11629 (IDEA remains enabled despite being patented. This is because IDEA
11630 is frequently required for interoperability, and there is no license
11631 fee for non-commercial use. As before, "no-idea" can be used to
11632 avoid this algorithm.)
11636 * Add processing of proxy certificates (see RFC 3820). This work was
11637 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11638 EGEE (Enabling Grids for E-science in Europe).
11642 * RC4 performance overhaul on modern architectures/implementations, such
11643 as Intel P4, IA-64 and AMD64.
11647 * New utility extract-section.pl. This can be used specify an alternative
11648 section number in a pod file instead of having to treat each file as
11649 a separate case in Makefile. This can be done by adding two lines to the
11652 =for comment openssl_section:XXX
11654 The blank line is mandatory.
11658 * New arguments -certform, -keyform and -pass for s_client and s_server
11659 to allow alternative format key and certificate files and passphrase
11664 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11665 update associated structures and add various utility functions.
11667 Add new policy related verify parameters, include policy checking in
11668 standard verify code. Enhance 'smime' application with extra parameters
11669 to support policy checking and print out.
11673 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11674 Nehemiah processors. These extensions support AES encryption in hardware
11675 as well as RNG (though RNG support is currently disabled).
11677 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11679 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11683 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11685 *Andy Polyakov and a number of other people*
11687 * Improved PowerPC platform support. Most notably BIGNUM assembler
11688 implementation contributed by IBM.
11690 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11692 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11693 exponent rather than 'unsigned long'. There is a corresponding change to
11694 the new 'rsa_keygen' element of the RSA_METHOD structure.
11696 *Jelte Jansen, Geoff Thorpe*
11698 * Functionality for creating the initial serial number file is now
11699 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11701 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11702 number file to 1, which is bound to cause problems. To avoid
11703 the problems while respecting compatibility between different 0.9.7
11704 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11705 CA.pl for serial number initialization. With the new release 0.9.8,
11706 we can fix the problem directly in the 'ca' utility.)
11710 * Reduced header interdependencies by declaring more opaque objects in
11711 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11712 give fewer recursive includes, which could break lazy source code - so
11713 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11714 developers should define this symbol when building and using openssl to
11715 ensure they track the recommended behaviour, interfaces, [etc], but
11716 backwards-compatible behaviour prevails when this isn't defined.
11720 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11724 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11725 This will generate a random key of the appropriate length based on the
11726 cipher context. The EVP_CIPHER can provide its own random key generation
11727 routine to support keys of a specific form. This is used in the des and
11728 3des routines to generate a key of the correct parity. Update S/MIME
11729 code to use new functions and hence generate correct parity DES keys.
11730 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11731 valid (weak or incorrect parity).
11735 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11736 as looking them up. This is useful when the verified structure may contain
11737 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11738 present unless the new PKCS7_NO_CRL flag is asserted.
11742 * Extend ASN1 oid configuration module. It now additionally accepts the
11745 shortName = some long name, 1.2.3.4
11749 * Reimplemented the BN_CTX implementation. There is now no more static
11750 limitation on the number of variables it can handle nor the depth of the
11751 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11752 information can now expand as required, and rather than having a single
11753 static array of bignums, BN_CTX now uses a linked-list of such arrays
11754 allowing it to expand on demand whilst maintaining the usefulness of
11755 BN_CTX's "bundling".
11759 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11760 to allow all RSA operations to function using a single BN_CTX.
11764 * Preliminary support for certificate policy evaluation and checking. This
11765 is initially intended to pass the tests outlined in "Conformance Testing
11766 of Relying Party Client Certificate Path Processing Logic" v1.07.
11770 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11771 remained unused and not that useful. A variety of other little bignum
11772 tweaks and fixes have also been made continuing on from the audit (see
11777 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11778 associated ASN1, EVP and SSL functions and old ASN1 macros.
11782 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11783 and this should never fail. So the return value from the use of
11784 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11785 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11789 * BN_CTX_get() should return zero-valued bignums, providing the same
11790 initialised value as BN_new().
11792 *Geoff Thorpe, suggested by Ulf Möller*
11794 * Support for inhibitAnyPolicy certificate extension.
11798 * An audit of the BIGNUM code is underway, for which debugging code is
11799 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11800 is considered valid when processing BIGNUMs, and causes execution to
11801 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11802 further steps are taken to deliberately pollute unused data in BIGNUM
11803 structures to try and expose faulty code further on. For now, openssl will
11804 (in its default mode of operation) continue to tolerate the inconsistent
11805 forms that it has tolerated in the past, but authors and packagers should
11806 consider trying openssl and their own applications when compiled with
11807 these debugging symbols defined. It will help highlight potential bugs in
11808 their own code, and will improve the test coverage for OpenSSL itself. At
11809 some point, these tighter rules will become openssl's default to improve
11810 maintainability, though the assert()s and other overheads will remain only
11811 in debugging configurations. See bn.h for more details.
11813 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11815 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11816 that can only be obtained through BN_CTX_new() (which implicitly
11817 initialises it). The presence of this function only made it possible
11818 to overwrite an existing structure (and cause memory leaks).
11822 * Because of the callback-based approach for implementing LHASH as a
11823 template type, lh_insert() adds opaque objects to hash-tables and
11824 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11825 to clean up those corresponding objects before destroying the hash table
11826 (and losing the object pointers). So some over-zealous constifications in
11827 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11828 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11829 prototyped to have "const" restrictions on the object pointers they are
11830 given (and so aren't required to cast them away any more).
11834 * The tmdiff.h API was so ugly and minimal that our own timing utility
11835 (speed) prefers to use its own implementation. The two implementations
11836 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11837 its object type properly exposed (MS_TM) instead of casting to/from
11838 `char *`. This may still change yet if someone realises MS_TM and
11840 aren't necessarily the greatest nomenclatures - but this is what was used
11841 internally to the implementation so I've used that for now.
11845 * Ensure that deprecated functions do not get compiled when
11846 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11847 the self-tests were still using deprecated key-generation functions so
11848 these have been updated also.
11852 * Reorganise PKCS#7 code to separate the digest location functionality
11853 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11854 New function PKCS7_set_digest() to set the digest type for PKCS#7
11855 digestedData type. Add additional code to correctly generate the
11856 digestedData type and add support for this type in PKCS7 initialization
11861 * New function PKCS7_set0_type_other() this initializes a PKCS7
11862 structure of type "other".
11866 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11867 sure the loop does correctly stop and breaking ("division by zero")
11868 modulus operations are not performed. The (pre-generated) prime
11869 table crypto/bn/bn_prime.h was already correct, but it could not be
11870 re-generated on some platforms because of the "division by zero"
11871 situation in the script.
11873 *Ralf S. Engelschall*
11875 * Update support for ECC-based TLS ciphersuites according to
11876 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11877 SHA-1 now is only used for "small" curves (where the
11878 representation of a field element takes up to 24 bytes); for
11879 larger curves, the field element resulting from ECDH is directly
11880 used as premaster secret.
11882 *Douglas Stebila (Sun Microsystems Laboratories)*
11884 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11885 curve secp160r1 to the tests.
11887 *Douglas Stebila (Sun Microsystems Laboratories)*
11889 * Add the possibility to load symbols globally with DSO.
11891 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11893 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11894 control of the error stack.
11898 * Add support for STORE in ENGINE.
11902 * Add the STORE type. The intention is to provide a common interface
11903 to certificate and key stores, be they simple file-based stores, or
11904 HSM-type store, or LDAP stores, or...
11905 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11909 * Add a generic structure called OPENSSL_ITEM. This can be used to
11910 pass a list of arguments to any function as well as provide a way
11911 for a function to pass data back to the caller.
11915 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11916 works like BUF_strdup() but can be used to duplicate a portion of
11917 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11922 * Add the function sk_find_ex() which works like sk_find(), but will
11923 return an index to an element even if an exact match couldn't be
11924 found. The index is guaranteed to point at the element where the
11925 searched-for key would be inserted to preserve sorting order.
11929 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11930 takes an extra flags argument for optional functionality. Currently,
11931 the following flags are defined:
11933 OBJ_BSEARCH_VALUE_ON_NOMATCH
11934 This one gets OBJ_bsearch_ex() to return a pointer to the first
11935 element where the comparing function returns a negative or zero
11938 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11939 This one gets OBJ_bsearch_ex() to return a pointer to the first
11940 element where the comparing function returns zero. This is useful
11941 if there are more than one element where the comparing function
11946 * Make it possible to create self-signed certificates with 'openssl ca'
11947 in such a way that the self-signed certificate becomes part of the
11948 CA database and uses the same mechanisms for serial number generation
11949 as all other certificate signing. The new flag '-selfsign' enables
11950 this functionality. Adapt CA.sh and CA.pl.in.
11954 * Add functionality to check the public key of a certificate request
11955 against a given private. This is useful to check that a certificate
11956 request can be signed by that key (self-signing).
11960 * Make it possible to have multiple active certificates with the same
11961 subject in the CA index file. This is done only if the keyword
11962 'unique_subject' is set to 'no' in the main CA section (default
11963 if 'CA_default') of the configuration file. The value is saved
11964 with the database itself in a separate index attribute file,
11965 named like the index file with '.attr' appended to the name.
11969 * Generate multi-valued AVAs using '+' notation in config files for
11974 * Support for nameConstraints certificate extension.
11978 * Support for policyConstraints certificate extension.
11982 * Support for policyMappings certificate extension.
11986 * Make sure the default DSA_METHOD implementation only uses its
11987 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11988 and change its own handlers to be NULL so as to remove unnecessary
11989 indirection. This lets alternative implementations fallback to the
11990 default implementation more easily.
11994 * Support for directoryName in GeneralName related extensions
11999 * Make it possible to link applications using Makefile.shared.
12000 Make that possible even when linking against static libraries!
12004 * Support for single pass processing for S/MIME signing. This now
12005 means that S/MIME signing can be done from a pipe, in addition
12006 cleartext signing (multipart/signed type) is effectively streaming
12007 and the signed data does not need to be all held in memory.
12009 This is done with a new flag PKCS7_STREAM. When this flag is set
12010 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12011 is done after the data is output (and digests calculated) in
12012 SMIME_write_PKCS7().
12016 * Add full support for -rpath/-R, both in shared libraries and
12017 applications, at least on the platforms where it's known how
12022 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12023 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12024 will now compute a table of multiples of the generator that
12025 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12026 faster (notably in the case of a single point multiplication,
12027 scalar * generator).
12029 *Nils Larsch, Bodo Moeller*
12031 * IPv6 support for certificate extensions. The various extensions
12032 which use the IP:a.b.c.d can now take IPv6 addresses using the
12033 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12038 * Added an ENGINE that implements RSA by performing private key
12039 exponentiations with the GMP library. The conversions to and from
12040 GMP's mpz_t format aren't optimised nor are any montgomery forms
12041 cached, and on x86 it appears OpenSSL's own performance has caught up.
12042 However there are likely to be other architectures where GMP could
12043 provide a boost. This ENGINE is not built in by default, but it can be
12044 specified at Configure time and should be accompanied by the necessary
12045 linker additions, eg;
12046 ./config -DOPENSSL_USE_GMP -lgmp
12050 * "openssl engine" will not display ENGINE/DSO load failure errors when
12051 testing availability of engines with "-t" - the old behaviour is
12052 produced by increasing the feature's verbosity with "-tt".
12056 * ECDSA routines: under certain error conditions uninitialized BN objects
12057 could be freed. Solution: make sure initialization is performed early
12058 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12063 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12064 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12065 software implementations. For DSA and DH, parameter generation can
12066 also be overridden by providing the appropriate method callbacks.
12070 * Change the "progress" mechanism used in key-generation and
12071 primality testing to functions that take a new BN_GENCB pointer in
12072 place of callback/argument pairs. The new API functions have `_ex`
12073 postfixes and the older functions are reimplemented as wrappers for
12074 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12075 declarations of the old functions to help (graceful) attempts to
12076 migrate to the new functions. Also, the new key-generation API
12077 functions operate on a caller-supplied key-structure and return
12078 success/failure rather than returning a key or NULL - this is to
12079 help make "keygen" another member function of RSA_METHOD etc.
12081 Example for using the new callback interface:
12083 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12084 void *my_arg = ...;
12087 BN_GENCB_set(&my_cb, my_callback, my_arg);
12089 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12090 /* For the meaning of a, b in calls to my_callback(), see the
12091 * documentation of the function that calls the callback.
12092 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12093 * my_callback should return 1 if it wants BN_is_prime_ex()
12094 * to continue, or 0 to stop.
12099 * Change the ZLIB compression method to be stateful, and make it
12100 available to TLS with the number defined in
12101 draft-ietf-tls-compression-04.txt.
12105 * Add the ASN.1 structures and functions for CertificatePair, which
12106 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12108 CertificatePair ::= SEQUENCE {
12109 forward [0] Certificate OPTIONAL,
12110 reverse [1] Certificate OPTIONAL,
12111 -- at least one of the pair shall be present -- }
12113 Also implement the PEM functions to read and write certificate
12114 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12116 This needed to be defined, mostly for the sake of the LDAP
12117 attribute crossCertificatePair, but may prove useful elsewhere as
12122 * Make it possible to inhibit symlinking of shared libraries in
12123 Makefile.shared, for Cygwin's sake.
12127 * Extend the BIGNUM API by creating a function
12128 void BN_set_negative(BIGNUM *a, int neg);
12129 and a macro that behave like
12130 int BN_is_negative(const BIGNUM *a);
12132 to avoid the need to access 'a->neg' directly in applications.
12136 * Implement fast modular reduction for pseudo-Mersenne primes
12137 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12138 EC_GROUP_new_curve_GFp() will now automatically use this
12141 *Nils Larsch <nla@trustcenter.de>*
12143 * Add new lock type (CRYPTO_LOCK_BN).
12147 * Change the ENGINE framework to automatically load engines
12148 dynamically from specific directories unless they could be
12149 found to already be built in or loaded. Move all the
12150 current engines except for the cryptodev one to a new
12151 directory engines/.
12152 The engines in engines/ are built as shared libraries if
12153 the "shared" options was given to ./Configure or ./config.
12154 Otherwise, they are inserted in libcrypto.a.
12155 /usr/local/ssl/engines is the default directory for dynamic
12156 engines, but that can be overridden at configure time through
12157 the usual use of --prefix and/or --openssldir, and at run
12158 time with the environment variable OPENSSL_ENGINES.
12160 *Geoff Thorpe and Richard Levitte*
12162 * Add Makefile.shared, a helper makefile to build shared
12163 libraries. Adapt Makefile.org.
12167 * Add version info to Win32 DLLs.
12169 *Peter 'Luna' Runestig" <peter@runestig.com>*
12171 * Add new 'medium level' PKCS#12 API. Certificates and keys
12172 can be added using this API to created arbitrary PKCS#12
12173 files while avoiding the low-level API.
12175 New options to PKCS12_create(), key or cert can be NULL and
12176 will then be omitted from the output file. The encryption
12177 algorithm NIDs can be set to -1 for no encryption, the mac
12178 iteration count can be set to 0 to omit the mac.
12180 Enhance pkcs12 utility by making the -nokeys and -nocerts
12181 options work when creating a PKCS#12 file. New option -nomac
12182 to omit the mac, NONE can be set for an encryption algorithm.
12183 New code is modified to use the enhanced PKCS12_create()
12184 instead of the low-level API.
12188 * Extend ASN1 encoder to support indefinite length constructed
12189 encoding. This can output sequences tags and octet strings in
12190 this form. Modify pk7_asn1.c to support indefinite length
12191 encoding. This is experimental and needs additional code to
12192 be useful, such as an ASN1 bio and some enhanced streaming
12195 Extend template encode functionality so that tagging is passed
12196 down to the template encoder.
12200 * Let 'openssl req' fail if an argument to '-newkey' is not
12201 recognized instead of using RSA as a default.
12205 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12206 As these are not official, they are not included in "ALL";
12207 the "ECCdraft" ciphersuite group alias can be used to select them.
12209 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12211 * Add ECDH engine support.
12213 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12215 * Add ECDH in new directory crypto/ecdh/.
12217 *Douglas Stebila (Sun Microsystems Laboratories)*
12219 * Let BN_rand_range() abort with an error after 100 iterations
12220 without success (which indicates a broken PRNG).
12224 * Change BN_mod_sqrt() so that it verifies that the input value
12225 is really the square of the return value. (Previously,
12226 BN_mod_sqrt would show GIGO behaviour.)
12230 * Add named elliptic curves over binary fields from X9.62, SECG,
12231 and WAP/WTLS; add OIDs that were still missing.
12233 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12235 * Extend the EC library for elliptic curves over binary fields
12236 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12239 EC_GF2m_simple_method
12243 EC_GROUP_new_curve_GF2m
12244 EC_GROUP_set_curve_GF2m
12245 EC_GROUP_get_curve_GF2m
12246 EC_POINT_set_affine_coordinates_GF2m
12247 EC_POINT_get_affine_coordinates_GF2m
12248 EC_POINT_set_compressed_coordinates_GF2m
12250 Point compression for binary fields is disabled by default for
12251 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12254 As binary polynomials are represented as BIGNUMs, various members
12255 of the EC_GROUP and EC_POINT data structures can be shared
12256 between the implementations for prime fields and binary fields;
12257 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12258 are essentially identical to their `..._GFp` counterparts.
12259 (For simplicity, the `..._GFp` prefix has been dropped from
12260 various internal method names.)
12262 An internal 'field_div' method (similar to 'field_mul' and
12263 'field_sqr') has been added; this is used only for binary fields.
12265 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12267 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12268 through methods ('mul', 'precompute_mult').
12270 The generic implementations (now internally called 'ec_wNAF_mul'
12271 and 'ec_wNAF_precomputed_mult') remain the default if these
12272 methods are undefined.
12274 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12276 * New function EC_GROUP_get_degree, which is defined through
12277 EC_METHOD. For curves over prime fields, this returns the bit
12278 length of the modulus.
12280 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12282 * New functions EC_GROUP_dup, EC_POINT_dup.
12283 (These simply call ..._new and ..._copy).
12285 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12287 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12288 Polynomials are represented as BIGNUMs (where the sign bit is not
12289 used) in the following functions [macros]:
12292 BN_GF2m_sub [= BN_GF2m_add]
12293 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12294 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12295 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12297 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12298 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12299 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12300 BN_GF2m_cmp [= BN_ucmp]
12302 (Note that only the 'mod' functions are actually for fields GF(2^m).
12303 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12305 For some functions, an the irreducible polynomial defining a
12306 field can be given as an 'unsigned int[]' with strictly
12307 decreasing elements giving the indices of those bits that are set;
12308 i.e., p[] represents the polynomial
12309 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12311 p[0] > p[1] > ... > p[k] = 0.
12312 This applies to the following functions:
12315 BN_GF2m_mod_mul_arr
12316 BN_GF2m_mod_sqr_arr
12317 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12318 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12319 BN_GF2m_mod_exp_arr
12320 BN_GF2m_mod_sqrt_arr
12321 BN_GF2m_mod_solve_quad_arr
12325 Conversion can be performed by the following functions:
12330 bntest.c has additional tests for binary polynomial arithmetic.
12332 Two implementations for BN_GF2m_mod_div() are available.
12333 The default algorithm simply uses BN_GF2m_mod_inv() and
12334 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12335 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12336 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12338 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12340 * Add new error code 'ERR_R_DISABLED' that can be used when some
12341 functionality is disabled at compile-time.
12343 *Douglas Stebila <douglas.stebila@sun.com>*
12345 * Change default behaviour of 'openssl asn1parse' so that more
12346 information is visible when viewing, e.g., a certificate:
12348 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12349 mode the content of non-printable OCTET STRINGs is output in a
12350 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12351 avoid the appearance of a printable string.
12353 *Nils Larsch <nla@trustcenter.de>*
12355 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12357 EC_GROUP_set_asn1_flag()
12358 EC_GROUP_get_asn1_flag()
12359 EC_GROUP_set_point_conversion_form()
12360 EC_GROUP_get_point_conversion_form()
12361 These control ASN1 encoding details:
12362 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12363 has been set to OPENSSL_EC_NAMED_CURVE.
12364 - Points are encoded in uncompressed form by default; options for
12365 asn1_for are as for point2oct, namely
12366 POINT_CONVERSION_COMPRESSED
12367 POINT_CONVERSION_UNCOMPRESSED
12368 POINT_CONVERSION_HYBRID
12370 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12372 EC_GROUP_set_seed()
12373 EC_GROUP_get0_seed()
12374 EC_GROUP_get_seed_len()
12375 This is used only for ASN1 purposes (so far).
12377 *Nils Larsch <nla@trustcenter.de>*
12379 * Add 'field_type' member to EC_METHOD, which holds the NID
12380 of the appropriate field type OID. The new function
12381 EC_METHOD_get_field_type() returns this value.
12383 *Nils Larsch <nla@trustcenter.de>*
12386 EC_POINT_point2bn()
12387 EC_POINT_bn2point()
12388 EC_POINT_point2hex()
12389 EC_POINT_hex2point()
12390 providing useful interfaces to EC_POINT_point2oct() and
12391 EC_POINT_oct2point().
12393 *Nils Larsch <nla@trustcenter.de>*
12395 * Change internals of the EC library so that the functions
12396 EC_GROUP_set_generator()
12397 EC_GROUP_get_generator()
12398 EC_GROUP_get_order()
12399 EC_GROUP_get_cofactor()
12400 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12401 to methods, which would lead to unnecessary code duplication when
12402 adding different types of curves.
12404 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12406 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12407 arithmetic, and such that modified wNAFs are generated
12408 (which avoid length expansion in many cases).
12412 * Add a function EC_GROUP_check_discriminant() (defined via
12413 EC_METHOD) that verifies that the curve discriminant is non-zero.
12415 Add a function EC_GROUP_check() that makes some sanity tests
12416 on a EC_GROUP, its generator and order. This includes
12417 EC_GROUP_check_discriminant().
12419 *Nils Larsch <nla@trustcenter.de>*
12421 * Add ECDSA in new directory crypto/ecdsa/.
12423 Add applications 'openssl ecparam' and 'openssl ecdsa'
12424 (these are based on 'openssl dsaparam' and 'openssl dsa').
12426 ECDSA support is also included in various other files across the
12427 library. Most notably,
12428 - 'openssl req' now has a '-newkey ecdsa:file' option;
12429 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12430 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12431 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12432 them suitable for ECDSA where domain parameters must be
12433 extracted before the specific public key;
12434 - ECDSA engine support has been added.
12436 *Nils Larsch <nla@trustcenter.de>*
12438 * Include some named elliptic curves, and add OIDs from X9.62,
12439 SECG, and WAP/WTLS. Each curve can be obtained from the new
12441 EC_GROUP_new_by_curve_name(),
12442 and the list of available named curves can be obtained with
12443 EC_get_builtin_curves().
12444 Also add a 'curve_name' member to EC_GROUP objects, which can be
12446 EC_GROUP_set_curve_name()
12447 EC_GROUP_get_curve_name()
12449 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12451 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12452 was actually never needed) and in BN_mul(). The removal in BN_mul()
12453 required a small change in bn_mul_part_recursive() and the addition
12454 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12455 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12456 bn_sub_words() and bn_add_words() except they take arrays with
12461 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12463 * Cleanse PEM buffers before freeing them since they may contain
12466 *Benjamin Bennett <ben@psc.edu>*
12468 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12469 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12470 authentication-only ciphersuites.
12474 * Since AES128 and AES256 share a single mask bit in the logic of
12475 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12476 kludge to work properly if AES128 is available and AES256 isn't.
12480 * Expand security boundary to match 1.1.1 module.
12484 * Remove redundant features: hash file source, editing of test vectors
12485 modify fipsld to use external fips_premain.c signature.
12489 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12490 run algorithm test programs.
12494 * Make algorithm test programs more tolerant of whitespace.
12498 * Have SSL/TLS server implementation tolerate "mismatched" record
12499 protocol version while receiving ClientHello even if the
12500 ClientHello is fragmented. (The server can't insist on the
12501 particular protocol version it has chosen before the ServerHello
12502 message has informed the client about his choice.)
12506 * Load error codes if they are not already present instead of using a
12507 static variable. This allows them to be cleanly unloaded and reloaded.
12511 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12513 * Introduce limits to prevent malicious keys being able to
12514 cause a denial of service. ([CVE-2006-2940])
12516 *Steve Henson, Bodo Moeller*
12518 * Fix ASN.1 parsing of certain invalid structures that can result
12519 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12521 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12522 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12524 * Fix SSL client code which could crash if connecting to a
12525 malicious SSLv2 server. ([CVE-2006-4343])
12527 *Tavis Ormandy and Will Drewry, Google Security Team*
12529 * Change ciphersuite string processing so that an explicit
12530 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12531 will no longer include "AES128-SHA"), and any other similar
12532 ciphersuite (same bitmap) from *other* protocol versions (so that
12533 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12534 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12535 changes from 0.9.8b and 0.9.8d.
12539 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12541 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12542 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12544 * Change the Unix randomness entropy gathering to use poll() when
12545 possible instead of select(), since the latter has some
12546 undesirable limitations.
12548 *Darryl Miles via Richard Levitte and Bodo Moeller*
12550 * Disable rogue ciphersuites:
12552 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12553 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12554 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12556 The latter two were purportedly from
12557 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12560 Also deactivate the remaining ciphersuites from
12561 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12562 unofficial, and the ID has long expired.
12566 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12567 dual-core machines) and other potential thread-safety issues.
12571 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12573 * Adapt fipsld and the build system to link against the validated FIPS
12574 module in FIPS mode.
12578 * Fixes for VC++ 2005 build under Windows.
12582 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12583 from a Windows bash shell such as MSYS. It is autodetected from the
12584 "config" script when run from a VC++ environment. Modify standard VC++
12585 build to use fipscanister.o from the GNU make build.
12589 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12591 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12592 The value now differs depending on if you build for FIPS or not.
12593 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12594 safely run with a non-FIPSed libcrypto, as it may crash because of
12595 the difference induced by this change.
12599 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12601 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12602 (part of SSL_OP_ALL). This option used to disable the
12603 countermeasure against man-in-the-middle protocol-version
12604 rollback in the SSL 2.0 server implementation, which is a bad
12605 idea. ([CVE-2005-2969])
12607 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12608 for Information Security, National Institute of Advanced Industrial
12609 Science and Technology [AIST, Japan)]*
12611 * Minimal support for X9.31 signatures and PSS padding modes. This is
12612 mainly for FIPS compliance and not fully integrated at this stage.
12616 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12617 the exponentiation using a fixed-length exponent. (Otherwise,
12618 the information leaked through timing could expose the secret key
12619 after many signatures; cf. Bleichenbacher's attack on DSA with
12624 * Make a new fixed-window mod_exp implementation the default for
12625 RSA, DSA, and DH private-key operations so that the sequence of
12626 squares and multiplies and the memory access pattern are
12627 independent of the particular secret key. This will mitigate
12628 cache-timing and potential related attacks.
12630 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12631 and this is automatically used by BN_mod_exp_mont() if the new flag
12632 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12633 will use this BN flag for private exponents unless the flag
12634 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12635 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12637 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12639 * Change the client implementation for SSLv23_method() and
12640 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12641 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12642 (Previously, the SSL 2.0 backwards compatible Client Hello
12643 message format would be used even with SSL_OP_NO_SSLv2.)
12647 * Add support for smime-type MIME parameter in S/MIME messages which some
12652 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12653 a threadsafe manner. Modify rsa code to use new function and add calls
12654 to dsa and dh code (which had race conditions before).
12658 * Include the fixed error library code in the C error file definitions
12659 instead of fixing them up at runtime. This keeps the error code
12660 structures constant.
12664 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12666 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12669 * Fixes for newer kerberos headers. NB: the casts are needed because
12670 the 'length' field is signed on one version and unsigned on another
12671 with no (?) obvious way to tell the difference, without these VC++
12672 complains. Also the "definition" of FAR (blank) is no longer included
12673 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12674 some needed definitions.
12678 * Undo Cygwin change.
12682 * Added support for proxy certificates according to RFC 3820.
12683 Because they may be a security thread to unaware applications,
12684 they must be explicitly allowed in run-time. See
12685 docs/HOWTO/proxy_certificates.txt for further information.
12689 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12691 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12692 server and client random values. Previously
12693 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12694 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12696 This change has negligible security impact because:
12698 1. Server and client random values still have 24 bytes of pseudo random
12701 2. Server and client random values are sent in the clear in the initial
12704 3. The master secret is derived using the premaster secret (48 bytes in
12705 size for static RSA ciphersuites) as well as client server and random
12708 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12711 *Stephen Henson, reported by UK NISCC*
12713 * Use Windows randomness collection on Cygwin.
12717 * Fix hang in EGD/PRNGD query when communication socket is closed
12718 prematurely by EGD/PRNGD.
12720 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12722 * Prompt for pass phrases when appropriate for PKCS12 input format.
12726 * Back-port of selected performance improvements from development
12727 branch, as well as improved support for PowerPC platforms.
12731 * Add lots of checks for memory allocation failure, error codes to indicate
12732 failure and freeing up memory if a failure occurs.
12734 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12736 * Add new -passin argument to dgst.
12740 * Perform some character comparisons of different types in X509_NAME_cmp:
12741 this is needed for some certificates that re-encode DNs into UTF8Strings
12742 (in violation of RFC3280) and can't or won't issue name rollover
12747 * Make an explicit check during certificate validation to see that
12748 the CA setting in each certificate on the chain is correct. As a
12749 side effect always do the following basic checks on extensions,
12750 not just when there's an associated purpose to the check:
12752 - if there is an unhandled critical extension (unless the user
12753 has chosen to ignore this fault)
12754 - if the path length has been exceeded (if one is set at all)
12755 - that certain extensions fit the associated purpose (if one has
12760 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12762 * Avoid a race condition when CRLs are checked in a multi threaded
12763 environment. This would happen due to the reordering of the revoked
12764 entries during signature checking and serial number lookup. Now the
12765 encoding is cached and the serial number sort performed under a lock.
12766 Add new STACK function sk_is_sorted().
12770 * Add Delta CRL to the extension code.
12774 * Various fixes to s3_pkt.c so alerts are sent properly.
12776 *David Holmes <d.holmes@f5.com>*
12778 * Reduce the chances of duplicate issuer name and serial numbers (in
12779 violation of RFC3280) using the OpenSSL certificate creation utilities.
12780 This is done by creating a random 64 bit value for the initial serial
12781 number when a serial number file is created or when a self signed
12782 certificate is created using 'openssl req -x509'. The initial serial
12783 number file is created using 'openssl x509 -next_serial' in CA.pl
12784 rather than being initialized to 1.
12788 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12790 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12791 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12793 *Joe Orton, Steve Henson*
12795 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12798 *Joe Orton, Steve Henson*
12800 * Make it possible to have multiple active certificates with the same
12801 subject in the CA index file. This is done only if the keyword
12802 'unique_subject' is set to 'no' in the main CA section (default
12803 if 'CA_default') of the configuration file. The value is saved
12804 with the database itself in a separate index attribute file,
12805 named like the index file with '.attr' appended to the name.
12809 * X509 verify fixes. Disable broken certificate workarounds when
12810 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12811 keyUsage extension present. Don't accept CRLs with unhandled critical
12812 extensions: since verify currently doesn't process CRL extensions this
12813 rejects a CRL with *any* critical extensions. Add new verify error codes
12818 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12819 A clarification of RFC2560 will require the use of OCTET STRINGs and
12820 some implementations cannot handle the current raw format. Since OpenSSL
12821 copies and compares OCSP nonces as opaque blobs without any attempt at
12822 parsing them this should not create any compatibility issues.
12826 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12827 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12828 this HMAC (and other) operations are several times slower than OpenSSL
12833 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12835 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12837 * Use the correct content when signing type "other".
12841 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12843 * Fix various bugs revealed by running the NISCC test suite:
12845 Stop out of bounds reads in the ASN1 code when presented with
12846 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12848 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12850 If verify callback ignores invalid public key errors don't try to check
12851 certificate signature with the NULL public key.
12855 * New -ignore_err option in ocsp application to stop the server
12856 exiting on the first error in a request.
12860 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12861 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12866 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12867 extra data after the compression methods not only for TLS 1.0
12868 but also for SSL 3.0 (as required by the specification).
12870 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12872 * Change X509_certificate_type() to mark the key as exported/exportable
12873 when it's 512 *bits* long, not 512 bytes.
12877 * Change AES_cbc_encrypt() so it outputs exact multiple of
12878 blocks during encryption.
12882 * Various fixes to base64 BIO and non blocking I/O. On write
12883 flushes were not handled properly if the BIO retried. On read
12884 data was not being buffered properly and had various logic bugs.
12885 This also affects blocking I/O when the data being decoded is a
12890 * Various S/MIME bugfixes and compatibility changes:
12891 output correct application/pkcs7 MIME type if
12892 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12893 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12894 of files as .eml work). Correctly handle very long lines in MIME
12899 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12901 * Countermeasure against the Klima-Pokorny-Rosa extension of
12902 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12903 a protocol version number mismatch like a decryption error
12904 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12908 * Turn on RSA blinding by default in the default implementation
12909 to avoid a timing attack. Applications that don't want it can call
12910 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12911 They would be ill-advised to do so in most cases.
12913 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12915 * Change RSA blinding code so that it works when the PRNG is not
12916 seeded (in this case, the secret RSA exponent is abused as
12917 an unpredictable seed -- if it is not unpredictable, there
12918 is no point in blinding anyway). Make RSA blinding thread-safe
12919 by remembering the creator's thread ID in rsa->blinding and
12920 having all other threads use local one-time blinding factors
12921 (this requires more computation than sharing rsa->blinding, but
12922 avoids excessive locking; and if an RSA object is not shared
12923 between threads, blinding will still be very fast).
12927 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12928 ENGINE as defaults for all supported algorithms irrespective of
12929 the 'flags' parameter. 'flags' is now honoured, so applications
12930 should make sure they are passing it correctly.
12934 * Target "mingw" now allows native Windows code to be generated in
12935 the Cygwin environment as well as with the MinGW compiler.
12939 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12941 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12942 via timing by performing a MAC computation even if incorrect
12943 block cipher padding has been found. This is a countermeasure
12944 against active attacks where the attacker has to distinguish
12945 between bad padding and a MAC verification error. ([CVE-2003-0078])
12947 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12948 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12949 Martin Vuagnoux (EPFL, Ilion)*
12951 * Make the no-err option work as intended. The intention with no-err
12952 is not to have the whole error stack handling routines removed from
12953 libcrypto, it's only intended to remove all the function name and
12954 reason texts, thereby removing some of the footprint that may not
12955 be interesting if those errors aren't displayed anyway.
12957 NOTE: it's still possible for any application or module to have its
12958 own set of error texts inserted. The routines are there, just not
12959 used by default when no-err is given.
12963 * Add support for FreeBSD on IA64.
12965 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12967 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12968 Kerberos function mit_des_cbc_cksum(). Before this change,
12969 the value returned by DES_cbc_cksum() was like the one from
12970 mit_des_cbc_cksum(), except the bytes were swapped.
12972 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12974 * Allow an application to disable the automatic SSL chain building.
12975 Before this a rather primitive chain build was always performed in
12976 ssl3_output_cert_chain(): an application had no way to send the
12977 correct chain if the automatic operation produced an incorrect result.
12979 Now the chain builder is disabled if either:
12981 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12983 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12985 The reasoning behind this is that an application would not want the
12986 auto chain building to take place if extra chain certificates are
12987 present and it might also want a means of sending no additional
12988 certificates (for example the chain has two certificates and the
12993 * Add the possibility to build without the ENGINE framework.
12995 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12997 * Under Win32 gmtime() can return NULL: check return value in
12998 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13002 * DSA routines: under certain error conditions uninitialized BN objects
13003 could be freed. Solution: make sure initialization is performed early
13004 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13005 Nils Larsch <nla@trustcenter.de> via PR#459)
13009 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13010 checked on reconnect on the client side, therefore session resumption
13011 could still fail with a "ssl session id is different" error. This
13012 behaviour is masked when SSL_OP_ALL is used due to
13013 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13014 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13015 followup to PR #377.
13019 * IA-32 assembler support enhancements: unified ELF targets, support
13020 for SCO/Caldera platforms, fix for Cygwin shared build.
13024 * Add support for FreeBSD on sparc64. As a consequence, support for
13025 FreeBSD on non-x86 processors is separate from x86 processors on
13026 the config script, much like the NetBSD support.
13028 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13030 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13032 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13035 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13036 code (06) was taken as the first octet of the session ID and the last
13037 octet was ignored consequently. As a result SSLv2 client side session
13038 caching could not have worked due to the session ID mismatch between
13040 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13045 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13046 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13051 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13052 seems that in spite of existing for more than a year, many application
13053 author have done nothing to provide the necessary callbacks, which
13054 means that this particular engine will not work properly anywhere.
13055 This is a very unfortunate situation which forces us, in the name
13056 of usability, to give the hw_ncipher.c a static lock, which is part
13058 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13059 appear in 0.9.8 or later. We EXPECT application authors to have
13060 dealt properly with this when 0.9.8 is released (unless we actually
13061 make such changes in the libcrypto locking code that changes will
13062 have to be made anyway).
13066 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13067 octets have been read, EOF or an error occurs. Without this change
13068 some truncated ASN1 structures will not produce an error.
13072 * Disable Heimdal support, since it hasn't been fully implemented.
13073 Still give the possibility to force the use of Heimdal, but with
13074 warnings and a request that patches get sent to openssl-dev.
13078 * Add the VC-CE target, introduce the WINCE sysname, and add
13079 INSTALL.WCE and appropriate conditionals to make it build.
13081 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13083 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13084 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13085 edit numbers of the version.
13087 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13089 * Introduce safe string copy and catenation functions
13090 (BUF_strlcpy() and BUF_strlcat()).
13092 *Ben Laurie (CHATS) and Richard Levitte*
13094 * Avoid using fixed-size buffers for one-line DNs.
13096 *Ben Laurie (CHATS)*
13098 * Add BUF_MEM_grow_clean() to avoid information leakage when
13099 resizing buffers containing secrets, and use where appropriate.
13101 *Ben Laurie (CHATS)*
13103 * Avoid using fixed size buffers for configuration file location.
13105 *Ben Laurie (CHATS)*
13107 * Avoid filename truncation for various CA files.
13109 *Ben Laurie (CHATS)*
13111 * Use sizeof in preference to magic numbers.
13113 *Ben Laurie (CHATS)*
13115 * Avoid filename truncation in cert requests.
13117 *Ben Laurie (CHATS)*
13119 * Add assertions to check for (supposedly impossible) buffer
13122 *Ben Laurie (CHATS)*
13124 * Don't cache truncated DNS entries in the local cache (this could
13125 potentially lead to a spoofing attack).
13127 *Ben Laurie (CHATS)*
13129 * Fix various buffers to be large enough for hex/decimal
13130 representations in a platform independent manner.
13132 *Ben Laurie (CHATS)*
13134 * Add CRYPTO_realloc_clean() to avoid information leakage when
13135 resizing buffers containing secrets, and use where appropriate.
13137 *Ben Laurie (CHATS)*
13139 * Add BIO_indent() to avoid much slightly worrying code to do
13142 *Ben Laurie (CHATS)*
13144 * Convert sprintf()/BIO_puts() to BIO_printf().
13146 *Ben Laurie (CHATS)*
13148 * buffer_gets() could terminate with the buffer only half
13151 *Ben Laurie (CHATS)*
13153 * Add assertions to prevent user-supplied crypto functions from
13154 overflowing internal buffers by having large block sizes, etc.
13156 *Ben Laurie (CHATS)*
13158 * New OPENSSL_assert() macro (similar to assert(), but enabled
13161 *Ben Laurie (CHATS)*
13163 * Eliminate unused copy of key in RC4.
13165 *Ben Laurie (CHATS)*
13167 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13169 *Ben Laurie (CHATS)*
13171 * Fix off-by-one error in EGD path.
13173 *Ben Laurie (CHATS)*
13175 * If RANDFILE path is too long, ignore instead of truncating.
13177 *Ben Laurie (CHATS)*
13179 * Eliminate unused and incorrectly sized X.509 structure
13182 *Ben Laurie (CHATS)*
13184 * Eliminate unused and dangerous function knumber().
13186 *Ben Laurie (CHATS)*
13188 * Eliminate unused and dangerous structure, KSSL_ERR.
13190 *Ben Laurie (CHATS)*
13192 * Protect against overlong session ID context length in an encoded
13193 session object. Since these are local, this does not appear to be
13196 *Ben Laurie (CHATS)*
13198 * Change from security patch (see 0.9.6e below) that did not affect
13199 the 0.9.6 release series:
13201 Remote buffer overflow in SSL3 protocol - an attacker could
13202 supply an oversized master key in Kerberos-enabled versions.
13205 *Ben Laurie (CHATS)*
13207 * Change the SSL kerb5 codes to match RFC 2712.
13211 * Make -nameopt work fully for req and add -reqopt switch.
13213 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13215 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13217 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13219 * Make sure tests can be performed even if the corresponding algorithms
13220 have been removed entirely. This was also the last step to make
13221 OpenSSL compilable with DJGPP under all reasonable conditions.
13223 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13225 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13226 to allow version independent disabling of normally unselected ciphers,
13227 which may be activated as a side-effect of selecting a single cipher.
13229 (E.g., cipher list string "RSA" enables ciphersuites that are left
13230 out of "ALL" because they do not provide symmetric encryption.
13231 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13233 *Lutz Jaenicke, Bodo Moeller*
13235 * Add appropriate support for separate platform-dependent build
13236 directories. The recommended way to make a platform-dependent
13237 build directory is the following (tested on Linux), maybe with
13240 # Place yourself outside of the OpenSSL source tree. In
13241 # this example, the environment variable OPENSSL_SOURCE
13242 # is assumed to contain the absolute OpenSSL source directory.
13243 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13244 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13245 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13246 mkdir -p `dirname $F`
13247 ln -s $OPENSSL_SOURCE/$F $F
13250 To be absolutely sure not to disturb the source tree, a "make clean"
13251 is a good thing. If it isn't successful, don't worry about it,
13252 it probably means the source directory is very clean.
13256 * Make sure any ENGINE control commands make local copies of string
13257 pointers passed to them whenever necessary. Otherwise it is possible
13258 the caller may have overwritten (or deallocated) the original string
13259 data when a later ENGINE operation tries to use the stored values.
13261 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13263 * Improve diagnostics in file reading and command-line digests.
13265 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13267 * Add AES modes CFB and OFB to the object database. Correct an
13268 error in AES-CFB decryption.
13272 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13273 allows existing EVP_CIPHER_CTX structures to be reused after
13274 calling `EVP_*Final()`. This behaviour is used by encryption
13275 BIOs and some applications. This has the side effect that
13276 applications must explicitly clean up cipher contexts with
13277 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13281 * Check the values of dna and dnb in bn_mul_recursive before calling
13282 bn_mul_comba (a non zero value means the a or b arrays do not contain
13283 n2 elements) and fallback to bn_mul_normal if either is not zero.
13287 * Fix escaping of non-ASCII characters when using the -subj option
13288 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13292 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13293 form for "surname", serialNumber has no short form.
13294 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13295 therefore remove "mail" short name for "internet 7".
13296 The OID for unique identifiers in X509 certificates is
13297 x500UniqueIdentifier, not uniqueIdentifier.
13298 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13302 * Add an "init" command to the ENGINE config module and auto initialize
13303 ENGINEs. Without any "init" command the ENGINE will be initialized
13304 after all ctrl commands have been executed on it. If init=1 the
13305 ENGINE is initialized at that point (ctrls before that point are run
13306 on the uninitialized ENGINE and after on the initialized one). If
13307 init=0 then the ENGINE will not be initialized at all.
13311 * Fix the 'app_verify_callback' interface so that the user-defined
13312 argument is actually passed to the callback: In the
13313 SSL_CTX_set_cert_verify_callback() prototype, the callback
13314 declaration has been changed from
13317 int (*cb)(X509_STORE_CTX *,void *);
13318 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13319 i=s->ctx->app_verify_callback(&ctx)
13320 has been changed into
13321 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13323 To update applications using SSL_CTX_set_cert_verify_callback(),
13324 a dummy argument can be added to their callback functions.
13326 *D. K. Smetters <smetters@parc.xerox.com>*
13328 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13330 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13332 * Add and OPENSSL_LOAD_CONF define which will cause
13333 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13334 This allows older applications to transparently support certain
13335 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13336 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13337 load the config file and OPENSSL_add_all_algorithms_conf() which will
13338 always load it have also been added.
13342 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13343 Adjust NIDs and EVP layer.
13345 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13347 * Config modules support in openssl utility.
13349 Most commands now load modules from the config file,
13350 though in a few (such as version) this isn't done
13351 because it couldn't be used for anything.
13353 In the case of ca and req the config file used is
13354 the same as the utility itself: that is the -config
13355 command line option can be used to specify an
13360 * Move default behaviour from OPENSSL_config(). If appname is NULL
13361 use "openssl_conf" if filename is NULL use default openssl config file.
13365 * Add an argument to OPENSSL_config() to allow the use of an alternative
13366 config section name. Add a new flag to tolerate a missing config file
13367 and move code to CONF_modules_load_file().
13371 * Support for crypto accelerator cards from Accelerated Encryption
13372 Processing, www.aep.ie. (Use engine 'aep')
13373 The support was copied from 0.9.6c [engine] and adapted/corrected
13374 to work with the new engine framework.
13376 *AEP Inc. and Richard Levitte*
13378 * Support for SureWare crypto accelerator cards from Baltimore
13379 Technologies. (Use engine 'sureware')
13380 The support was copied from 0.9.6c [engine] and adapted
13381 to work with the new engine framework.
13385 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13386 make the newer ENGINE framework commands for the CHIL engine work.
13388 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13390 * Make it possible to produce shared libraries on ReliantUNIX.
13392 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13394 * Add the configuration target debug-linux-ppro.
13395 Make 'openssl rsa' use the general key loading routines
13396 implemented in `apps.c`, and make those routines able to
13397 handle the key format FORMAT_NETSCAPE and the variant
13400 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13402 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13404 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13406 * Add -keyform to rsautl, and document -engine.
13408 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13410 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13411 BIO_R_NO_SUCH_FILE error code rather than the generic
13412 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13416 * Add new functions
13417 ERR_peek_last_error
13418 ERR_peek_last_error_line
13419 ERR_peek_last_error_line_data.
13420 These are similar to
13422 ERR_peek_error_line
13423 ERR_peek_error_line_data,
13424 but report on the latest error recorded rather than the first one
13425 still in the error queue.
13427 *Ben Laurie, Bodo Moeller*
13429 * default_algorithms option in ENGINE config module. This allows things
13431 default_algorithms = ALL
13432 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13436 * Preliminary ENGINE config module.
13440 * New experimental application configuration code.
13444 * Change the AES code to follow the same name structure as all other
13445 symmetric ciphers, and behave the same way. Move everything to
13446 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13448 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13450 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13452 *Ben Laurie and Theo de Raadt*
13454 * Add option to output public keys in req command.
13456 *Massimiliano Pala madwolf@openca.org*
13458 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13459 (up to about 10% better than before for P-192 and P-224).
13463 * New functions/macros
13465 SSL_CTX_set_msg_callback(ctx, cb)
13466 SSL_CTX_set_msg_callback_arg(ctx, arg)
13467 SSL_set_msg_callback(ssl, cb)
13468 SSL_set_msg_callback_arg(ssl, arg)
13470 to request calling a callback function
13472 void cb(int write_p, int version, int content_type,
13473 const void *buf, size_t len, SSL *ssl, void *arg)
13475 whenever a protocol message has been completely received
13476 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13477 protocol version according to which the SSL library interprets
13478 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13479 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13480 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13481 specification (change_cipher_spec(20), alert(21), handshake(22)).
13482 'buf' and 'len' point to the actual message, 'ssl' to the
13483 SSL object, and 'arg' is the application-defined value set by
13484 SSL[_CTX]_set_msg_callback_arg().
13486 'openssl s_client' and 'openssl s_server' have new '-msg' options
13487 to enable a callback that displays all protocol messages.
13491 * Change the shared library support so shared libraries are built as
13492 soon as the corresponding static library is finished, and thereby get
13493 openssl and the test programs linked against the shared library.
13494 This still only happens when the keyword "shard" has been given to
13495 the configuration scripts.
13497 NOTE: shared library support is still an experimental thing, and
13498 backward binary compatibility is still not guaranteed.
13500 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13502 * Add support for Subject Information Access extension.
13504 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13506 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13507 additional bytes when new memory had to be allocated, not just
13508 when reusing an existing buffer.
13512 * New command line and configuration option 'utf8' for the req command.
13513 This allows field values to be specified as UTF8 strings.
13517 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13518 runs for the former and machine-readable output for the latter.
13522 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13523 of the e-mail address in the DN (i.e., it will go into a certificate
13524 extension only). The new configuration file option 'email_in_dn = no'
13525 has the same effect.
13527 *Massimiliano Pala madwolf@openca.org*
13529 * Change all functions with names starting with `des_` to be starting
13530 with `DES_` instead. Add wrappers that are compatible with libdes,
13531 but are named `_ossl_old_des_*`. Finally, add macros that map the
13532 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13533 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13534 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13537 Since we provide two compatibility mappings, the user needs to
13538 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13539 compatibility is desired. The default (i.e., when that macro
13540 isn't defined) is OpenSSL 0.9.6c compatibility.
13542 There are also macros that enable and disable the support of old
13543 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13544 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13545 are defined, the default will apply: to support the old des routines.
13547 In either case, one must include openssl/des.h to get the correct
13548 definitions. Do not try to just include openssl/des_old.h, that
13551 NOTE: This is a major break of an old API into a new one. Software
13552 authors are encouraged to switch to the `DES_` style functions. Some
13553 time in the future, des_old.h and the libdes compatibility functions
13554 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13555 default), and then completely removed.
13559 * Test for certificates which contain unsupported critical extensions.
13560 If such a certificate is found during a verify operation it is
13561 rejected by default: this behaviour can be overridden by either
13562 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13563 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13564 X509_supported_extension() has also been added which returns 1 if a
13565 particular extension is supported.
13569 * Modify the behaviour of EVP cipher functions in similar way to digests
13570 to retain compatibility with existing code.
13574 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13575 compatibility with existing code. In particular the 'ctx' parameter does
13576 not have to be to be initialized before the call to EVP_DigestInit() and
13577 it is tidied up after a call to EVP_DigestFinal(). New function
13578 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13579 EVP_MD_CTX_copy() changed to not require the destination to be
13580 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13581 requires the destination to be valid.
13583 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13584 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13588 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13589 so that complete 'Handshake' protocol structures are kept in memory
13590 instead of overwriting 'msg_type' and 'length' with 'body' data.
13594 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13596 *Massimo Santin via Richard Levitte*
13598 * Major restructuring to the underlying ENGINE code. This includes
13599 reduction of linker bloat, separation of pure "ENGINE" manipulation
13600 (initialisation, etc) from functionality dealing with implementations
13601 of specific crypto interfaces. This change also introduces integrated
13602 support for symmetric ciphers and digest implementations - so ENGINEs
13603 can now accelerate these by providing EVP_CIPHER and EVP_MD
13604 implementations of their own. This is detailed in
13605 [crypto/engine/README.md](crypto/engine/README.md)
13606 as it couldn't be adequately described here. However, there are a few
13607 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13608 were changed in the original introduction of ENGINE code have now
13609 reverted back - the hooking from this code to ENGINE is now a good
13610 deal more passive and at run-time, operations deal directly with
13611 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13612 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13613 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13614 they were not being used by the framework as there is no concept of a
13615 BIGNUM_METHOD and they could not be generalised to the new
13616 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13617 ENGINE_cpy() has been removed as it cannot be consistently defined in
13622 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13626 * Change mkdef.pl to sort symbols that get the same entry number,
13627 and make sure the automatically generated functions `ERR_load_*`
13628 become part of libeay.num as well.
13632 * New function SSL_renegotiate_pending(). This returns true once
13633 renegotiation has been requested (either SSL_renegotiate() call
13634 or HelloRequest/ClientHello received from the peer) and becomes
13635 false once a handshake has been completed.
13636 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13637 sends a HelloRequest, but does not ensure that a handshake takes
13638 place. SSL_renegotiate_pending() is useful for checking if the
13639 client has followed the request.)
13643 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13644 By default, clients may request session resumption even during
13645 renegotiation (if session ID contexts permit); with this option,
13646 session resumption is possible only in the first handshake.
13648 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13649 more bits available for options that should not be part of
13650 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13654 * Add some demos for certificate and certificate request creation.
13658 * Make maximum certificate chain size accepted from the peer application
13659 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13660 "Douglas E. Engert" <deengert@anl.gov>.
13664 * Add support for shared libraries for Unixware-7
13665 (Boyd Lynn Gerber <gerberb@zenez.com>).
13669 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13670 be done prior to destruction. Use this to unload error strings from
13671 ENGINEs that load their own error strings. NB: This adds two new API
13672 functions to "get" and "set" this destroy handler in an ENGINE.
13676 * Alter all existing ENGINE implementations (except "openssl" and
13677 "openbsd") to dynamically instantiate their own error strings. This
13678 makes them more flexible to be built both as statically-linked ENGINEs
13679 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13680 Also, add stub code to each that makes building them as self-contained
13681 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13685 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13686 implementations into applications that are completely implemented in
13687 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13688 commands that can be used to configure what shared-library to load and
13689 to control aspects of the way it is handled. Also, made an update to
13690 the [README-Engine.md](README-Engine.md) file
13691 that brings its information up-to-date and
13692 provides some information and instructions on the "dynamic" ENGINE
13693 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13697 * Make it possible to unload ranges of ERR strings with a new
13698 "ERR_unload_strings" function.
13702 * Add a copy() function to EVP_MD.
13706 * Make EVP_MD routines take a context pointer instead of just the
13707 md_data void pointer.
13711 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13712 that the digest can only process a single chunk of data
13713 (typically because it is provided by a piece of
13714 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13715 is only going to provide a single chunk of data, and hence the
13716 framework needn't accumulate the data for oneshot drivers.
13720 * As with "ERR", make it possible to replace the underlying "ex_data"
13721 functions. This change also alters the storage and management of global
13722 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13723 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13724 index counters. The API functions that use this state have been changed
13725 to take a "class_index" rather than pointers to the class's local STACK
13726 and counter, and there is now an API function to dynamically create new
13727 classes. This centralisation allows us to (a) plug a lot of the
13728 thread-safety problems that existed, and (b) makes it possible to clean
13729 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13730 such data would previously have always leaked in application code and
13731 workarounds were in place to make the memory debugging turn a blind eye
13732 to it. Application code that doesn't use this new function will still
13733 leak as before, but their memory debugging output will announce it now
13734 rather than letting it slide.
13736 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13737 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13738 has a return value to indicate success or failure.
13742 * Make it possible to replace the underlying "ERR" functions such that the
13743 global state (2 LHASH tables and 2 locks) is only used by the "default"
13744 implementation. This change also adds two functions to "get" and "set"
13745 the implementation prior to it being automatically set the first time
13746 any other ERR function takes place. Ie. an application can call "get",
13747 pass the return value to a module it has just loaded, and that module
13748 can call its own "set" function using that value. This means the
13749 module's "ERR" operations will use (and modify) the error state in the
13750 application and not in its own statically linked copy of OpenSSL code.
13754 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13755 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13756 the operation, and provides a more encapsulated way for external code
13757 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13758 to use these functions rather than manually incrementing the counts.
13760 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13764 * Add EVP test program.
13768 * Add symmetric cipher support to ENGINE. Expect the API to change!
13772 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13773 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13774 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13775 These allow a CRL to be built without having to access X509_CRL fields
13776 directly. Modify 'ca' application to use new functions.
13780 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13781 bug workarounds. Rollback attack detection is a security feature.
13782 The problem will only arise on OpenSSL servers when TLSv1 is not
13783 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13784 Software authors not wanting to support TLSv1 will have special reasons
13785 for their choice and can explicitly enable this option.
13787 *Bodo Moeller, Lutz Jaenicke*
13789 * Rationalise EVP so it can be extended: don't include a union of
13790 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13791 (similar to those existing for EVP_CIPHER_CTX).
13796 EVP_MD_CTX_init(&md); /* new function call */
13797 EVP_DigestInit(&md, EVP_sha1());
13798 EVP_DigestUpdate(&md, in, len);
13799 EVP_DigestFinal(&md, out, NULL);
13800 EVP_MD_CTX_cleanup(&md); /* new function call */
13804 * Make DES key schedule conform to the usual scheme, as well as
13805 correcting its structure. This means that calls to DES functions
13806 now have to pass a pointer to a des_key_schedule instead of a
13807 plain des_key_schedule (which was actually always a pointer
13810 des_key_schedule ks;
13812 des_set_key_checked(..., &ks);
13813 des_ncbc_encrypt(..., &ks, ...);
13815 (Note that a later change renames 'des_...' into 'DES_...'.)
13819 * Initial reduction of linker bloat: the use of some functions, such as
13820 PEM causes large amounts of unused functions to be linked in due to
13821 poor organisation. For example pem_all.c contains every PEM function
13822 which has a knock on effect of linking in large amounts of (unused)
13823 ASN1 code. Grouping together similar functions and splitting unrelated
13824 functions prevents this.
13828 * Cleanup of EVP macros.
13832 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13833 correct `_ecb suffix`.
13837 * Add initial OCSP responder support to ocsp application. The
13838 revocation information is handled using the text based index
13839 use by the ca application. The responder can either handle
13840 requests generated internally, supplied in files (for example
13841 via a CGI script) or using an internal minimal server.
13845 * Add configuration choices to get zlib compression for TLS.
13849 * Changes to Kerberos SSL for RFC 2712 compliance:
13850 1. Implemented real KerberosWrapper, instead of just using
13851 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13852 2. Implemented optional authenticator field of KerberosWrapper.
13854 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13855 and authenticator structs; see crypto/krb5/.
13857 Generalized Kerberos calls to support multiple Kerberos libraries.
13858 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13859 via Richard Levitte*
13861 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13862 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13863 values for each of the key sizes rather than having just
13864 parameters (and 'speed' generating keys each time).
13868 * Speed up EVP routines.
13871 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13872 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13873 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13874 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13876 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13877 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13878 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13881 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13883 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13887 * Added the OS2-EMX target.
13889 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13891 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13892 New functions to support `NCONF` routines in extension code.
13893 New function `CONF_set_nconf()`
13894 to allow functions which take an `NCONF` to also handle the old `LHASH`
13895 structure: this means that the old `CONF` compatible routines can be
13896 retained (in particular w.rt. extensions) without having to duplicate the
13897 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13901 * Enhance the general user interface with mechanisms for inner control
13902 and with possibilities to have yes/no kind of prompts.
13906 * Change all calls to low-level digest routines in the library and
13907 applications to use EVP. Add missing calls to HMAC_cleanup() and
13908 don't assume HMAC_CTX can be copied using memcpy().
13910 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13912 * Add the possibility to control engines through control names but with
13913 arbitrary arguments instead of just a string.
13914 Change the key loaders to take a UI_METHOD instead of a callback
13915 function pointer. NOTE: this breaks binary compatibility with earlier
13916 versions of OpenSSL [engine].
13917 Adapt the nCipher code for these new conditions and add a card insertion
13922 * Enhance the general user interface with mechanisms to better support
13923 dialog box interfaces, application-defined prompts, the possibility
13924 to use defaults (for example default passwords from somewhere else)
13925 and interrupts/cancellations.
13929 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13930 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13934 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13935 tidy up some unnecessarily weird code in 'sk_new()').
13937 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13939 * Change the key loading routines for ENGINEs to use the same kind
13940 callback (pem_password_cb) as all other routines that need this
13945 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13946 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13947 than this minimum value is recommended.
13951 * New random seeder for OpenVMS, using the system process statistics
13952 that are easily reachable.
13956 * Windows apparently can't transparently handle global
13957 variables defined in DLLs. Initialisations such as:
13959 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13961 won't compile. This is used by the any applications that need to
13962 declare their own ASN1 modules. This was fixed by adding the option
13963 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13964 needed for static libraries under Win32.
13968 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13969 setting of purpose and trust fields. New X509_STORE trust and
13970 purpose functions and tidy up setting in other SSL functions.
13974 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13975 structure. These are inherited by X509_STORE_CTX when it is
13976 initialised. This allows various defaults to be set in the
13977 X509_STORE structure (such as flags for CRL checking and custom
13978 purpose or trust settings) for functions which only use X509_STORE_CTX
13979 internally such as S/MIME.
13981 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13982 trust settings if they are not set in X509_STORE. This allows X509_STORE
13983 purposes and trust (in S/MIME for example) to override any set by default.
13985 Add command line options for CRL checking to smime, s_client and s_server
13990 * Initial CRL based revocation checking. If the CRL checking flag(s)
13991 are set then the CRL is looked up in the X509_STORE structure and
13992 its validity and signature checked, then if the certificate is found
13993 in the CRL the verify fails with a revoked error.
13995 Various new CRL related callbacks added to X509_STORE_CTX structure.
13997 Command line options added to 'verify' application to support this.
13999 This needs some additional work, such as being able to handle multiple
14000 CRLs with different times, extension based lookup (rather than just
14001 by subject name) and ultimately more complete V2 CRL extension
14006 * Add a general user interface API (crypto/ui/). This is designed
14007 to replace things like des_read_password and friends (backward
14008 compatibility functions using this new API are provided).
14009 The purpose is to remove prompting functions from the DES code
14010 section as well as provide for prompting through dialog boxes in
14011 a window system and the like.
14015 * Add "ex_data" support to ENGINE so implementations can add state at a
14016 per-structure level rather than having to store it globally.
14020 * Make it possible for ENGINE structures to be copied when retrieved by
14021 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14022 This causes the "original" ENGINE structure to act like a template,
14023 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14024 operational state can be localised to each ENGINE structure, despite the
14025 fact they all share the same "methods". New ENGINE structures returned in
14026 this case have no functional references and the return value is the single
14027 structural reference. This matches the single structural reference returned
14028 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14033 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14034 needs to match any other type at all we need to manually clear the
14039 * Changes to the "openssl engine" utility to include;
14040 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14041 about an ENGINE's available control commands.
14042 - executing control commands from command line arguments using the
14043 '-pre' and '-post' switches. '-post' is only used if '-t' is
14044 specified and the ENGINE is successfully initialised. The syntax for
14045 the individual commands are colon-separated, for example;
14046 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14050 * New dynamic control command support for ENGINEs. ENGINEs can now
14051 declare their own commands (numbers), names (strings), descriptions,
14052 and input types for run-time discovery by calling applications. A
14053 subset of these commands are implicitly classed as "executable"
14054 depending on their input type, and only these can be invoked through
14055 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14056 can be based on user input, config files, etc). The distinction is
14057 that "executable" commands cannot return anything other than a boolean
14058 result and can only support numeric or string input, whereas some
14059 discoverable commands may only be for direct use through
14060 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14061 pointers, or other custom uses. The "executable" commands are to
14062 support parameterisations of ENGINE behaviour that can be
14063 unambiguously defined by ENGINEs and used consistently across any
14064 OpenSSL-based application. Commands have been added to all the
14065 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14066 control over shared-library paths without source code alterations.
14070 * Changed all ENGINE implementations to dynamically allocate their
14071 ENGINEs rather than declaring them statically. Apart from this being
14072 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14073 this also allows the implementations to compile without using the
14074 internal engine_int.h header.
14078 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14079 'const' value. Any code that should be able to modify a RAND_METHOD
14080 should already have non-const pointers to it (ie. they should only
14081 modify their own ones).
14085 * Made a variety of little tweaks to the ENGINE code.
14086 - "atalla" and "ubsec" string definitions were moved from header files
14087 to C code. "nuron" string definitions were placed in variables
14088 rather than hard-coded - allowing parameterisation of these values
14089 later on via ctrl() commands.
14090 - Removed unused "#if 0"'d code.
14091 - Fixed engine list iteration code so it uses ENGINE_free() to release
14092 structural references.
14093 - Constified the RAND_METHOD element of ENGINE structures.
14094 - Constified various get/set functions as appropriate and added
14095 missing functions (including a catch-all ENGINE_cpy that duplicates
14096 all ENGINE values onto a new ENGINE except reference counts/state).
14097 - Removed NULL parameter checks in get/set functions. Setting a method
14098 or function to NULL is a way of cancelling out a previously set
14099 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14100 and doesn't justify the extra error symbols and code.
14101 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14102 flags from engine_int.h to engine.h.
14103 - Changed prototypes for ENGINE handler functions (init(), finish(),
14104 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14108 * Implement binary inversion algorithm for BN_mod_inverse in addition
14109 to the algorithm using long division. The binary algorithm can be
14110 used only if the modulus is odd. On 32-bit systems, it is faster
14111 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14112 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14113 up to 450 bits. In 64-bit environments, the binary algorithm
14114 appears to be advantageous for much longer moduli; here we use it
14115 for moduli up to 2048 bits.
14119 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14120 could not support the combine flag in choice fields.
14124 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14125 extensions from a certificate request to the certificate.
14129 * Allow multiple 'certopt' and 'nameopt' options to be separated
14130 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14131 file: this allows the display of the certificate about to be
14132 signed to be customised, to allow certain fields to be included
14133 or excluded and extension details. The old system didn't display
14134 multicharacter strings properly, omitted fields not in the policy
14135 and couldn't display additional details such as extensions.
14139 * Function EC_POINTs_mul for multiple scalar multiplication
14140 of an arbitrary number of elliptic curve points
14141 \sum scalars[i]*points[i],
14142 optionally including the generator defined for the EC_GROUP:
14143 scalar*generator + \sum scalars[i]*points[i].
14145 EC_POINT_mul is a simple wrapper function for the typical case
14146 that the point list has just one item (besides the optional
14151 * First EC_METHODs for curves over GF(p):
14153 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14154 operations and provides various method functions that can also
14155 operate with faster implementations of modular arithmetic.
14157 EC_GFp_mont_method() reuses most functions that are part of
14158 EC_GFp_simple_method, but uses Montgomery arithmetic.
14160 *Bodo Moeller; point addition and point doubling
14161 implementation directly derived from source code provided by
14162 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14164 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14165 crypto/ec/ec_lib.c):
14167 Curves are EC_GROUP objects (with an optional group generator)
14168 based on EC_METHODs that are built into the library.
14170 Points are EC_POINT objects based on EC_GROUP objects.
14172 Most of the framework would be able to handle curves over arbitrary
14173 finite fields, but as there are no obvious types for fields other
14174 than GF(p), some functions are limited to that for now.
14178 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14179 that the file contains a complete HTTP response.
14183 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14184 change the def and num file printf format specifier from "%-40sXXX"
14185 to "%-39s XXX". The latter will always guarantee a space after the
14186 field while the former will cause them to run together if the field
14187 is 40 of more characters long.
14191 * Constify the cipher and digest 'method' functions and structures
14192 and modify related functions to take constant EVP_MD and EVP_CIPHER
14197 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14198 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14202 * Modify `EVP_Digest*()` routines so they now return values. Although the
14203 internal software routines can never fail additional hardware versions
14208 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14210 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14211 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14214 ERR_R_NESTED_ASN1_ERROR
14216 ERR_R_MISSING_ASN1_EOS
14217 were 4 .. 9, conflicting with
14218 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14220 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14221 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14223 Add new error code 'ERR_R_INTERNAL_ERROR'.
14227 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14232 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14233 sets the subject name for a new request or supersedes the
14234 subject name in a given request. Formats that can be parsed are
14235 'CN=Some Name, OU=myOU, C=IT'
14237 'CN=Some Name/OU=myOU/C=IT'.
14239 Add options '-batch' and '-verbose' to 'openssl req'.
14241 *Massimiliano Pala <madwolf@hackmasters.net>*
14243 * Introduce the possibility to access global variables through
14244 functions on platform were that's the best way to handle exporting
14245 global variables in shared libraries. To enable this functionality,
14246 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14247 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14248 is normally done by Configure or something similar).
14250 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14251 in the source file (foo.c) like this:
14253 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14254 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14256 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14257 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14259 OPENSSL_DECLARE_GLOBAL(int,foo);
14260 #define foo OPENSSL_GLOBAL_REF(foo)
14261 OPENSSL_DECLARE_GLOBAL(double,bar);
14262 #define bar OPENSSL_GLOBAL_REF(bar)
14264 The #defines are very important, and therefore so is including the
14265 header file everywhere where the defined globals are used.
14267 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14268 of ASN.1 items, but that structure is a bit different.
14270 The largest change is in util/mkdef.pl which has been enhanced with
14271 better and easier to understand logic to choose which symbols should
14272 go into the Windows .def files as well as a number of fixes and code
14273 cleanup (among others, algorithm keywords are now sorted
14274 lexicographically to avoid constant rewrites).
14278 * In BN_div() keep a copy of the sign of 'num' before writing the
14279 result to 'rm' because if rm==num the value will be overwritten
14280 and produce the wrong result if 'num' is negative: this caused
14281 problems with BN_mod() and BN_nnmod().
14285 * Function OCSP_request_verify(). This checks the signature on an
14286 OCSP request and verifies the signer certificate. The signer
14287 certificate is just checked for a generic purpose and OCSP request
14292 * Add OCSP_check_validity() function to check the validity of OCSP
14293 responses. OCSP responses are prepared in real time and may only
14294 be a few seconds old. Simply checking that the current time lies
14295 between thisUpdate and nextUpdate max reject otherwise valid responses
14296 caused by either OCSP responder or client clock inaccuracy. Instead
14297 we allow thisUpdate and nextUpdate to fall within a certain period of
14298 the current time. The age of the response can also optionally be
14299 checked. Two new options -validity_period and -status_age added to
14304 * If signature or public key algorithm is unrecognized print out its
14305 OID rather that just UNKNOWN.
14309 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14310 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14311 ID to be generated from the issuer certificate alone which can then be
14312 passed to OCSP_id_issuer_cmp().
14316 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14317 ASN1 modules to export functions returning ASN1_ITEM pointers
14318 instead of the ASN1_ITEM structures themselves. This adds several
14319 new macros which allow the underlying ASN1 function/structure to
14320 be accessed transparently. As a result code should not use ASN1_ITEM
14321 references directly (such as &X509_it) but instead use the relevant
14322 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14323 use of the new ASN1 code on platforms where exporting structures
14324 is problematical (for example in shared libraries) but exporting
14325 functions returning pointers to structures is not.
14329 * Add support for overriding the generation of SSL/TLS session IDs.
14330 These callbacks can be registered either in an SSL_CTX or per SSL.
14331 The purpose of this is to allow applications to control, if they wish,
14332 the arbitrary values chosen for use as session IDs, particularly as it
14333 can be useful for session caching in multiple-server environments. A
14334 command-line switch for testing this (and any client code that wishes
14335 to use such a feature) has been added to "s_server".
14337 *Geoff Thorpe, Lutz Jaenicke*
14339 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14340 of the form `#if defined(...) || defined(...) || ...` and
14341 `#if !defined(...) && !defined(...) && ...`. This also avoids
14342 the growing number of special cases it was previously handling.
14346 * Make all configuration macros available for application by making
14347 sure they are available in opensslconf.h, by giving them names starting
14348 with `OPENSSL_` to avoid conflicts with other packages and by making
14349 sure e_os2.h will cover all platform-specific cases together with
14351 Additionally, it is now possible to define configuration/platform-
14352 specific names (called "system identities"). In the C code, these
14353 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14354 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14355 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14360 * New option -set_serial to 'req' and 'x509' this allows the serial
14361 number to use to be specified on the command line. Previously self
14362 signed certificates were hard coded with serial number 0 and the
14363 CA options of 'x509' had to use a serial number in a file which was
14368 * New options to 'ca' utility to support V2 CRL entry extensions.
14369 Currently CRL reason, invalidity date and hold instruction are
14370 supported. Add new CRL extensions to V3 code and some new objects.
14374 * New function EVP_CIPHER_CTX_set_padding() this is used to
14375 disable standard block padding (aka PKCS#5 padding) in the EVP
14376 API, which was previously mandatory. This means that the data is
14377 not padded in any way and so the total length much be a multiple
14378 of the block size, otherwise an error occurs.
14382 * Initial (incomplete) OCSP SSL support.
14386 * New function OCSP_parse_url(). This splits up a URL into its host,
14387 port and path components: primarily to parse OCSP URLs. New -url
14388 option to ocsp utility.
14392 * New nonce behavior. The return value of OCSP_check_nonce() now
14393 reflects the various checks performed. Applications can decide
14394 whether to tolerate certain situations such as an absent nonce
14395 in a response when one was present in a request: the ocsp application
14396 just prints out a warning. New function OCSP_add1_basic_nonce()
14397 this is to allow responders to include a nonce in a response even if
14398 the request is nonce-less.
14402 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14403 skipped when using openssl x509 multiple times on a single input file,
14404 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14408 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14409 set string type: to handle setting ASN1_TIME structures. Fix ca
14410 utility to correctly initialize revocation date of CRLs.
14414 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14415 the clients preferred ciphersuites and rather use its own preferences.
14416 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14417 Internet Explorer by ensuring unchanged hash method during stepup.
14418 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14422 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14423 to aes and add a new 'exist' option to print out symbols that don't
14428 * Additional options to ocsp utility to allow flags to be set and
14429 additional certificates supplied.
14433 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14434 OCSP client a number of certificate to only verify the response
14439 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14440 handle the new API. Currently only ECB, CBC modes supported. Add new
14443 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14444 Encryption Standard (AES) Ciphersuites for Transport Layer
14445 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14446 not enabled by default and were not part of the "ALL" ciphersuite
14447 alias because they were not yet official; they could be
14448 explicitly requested by specifying the "AESdraft" ciphersuite
14449 group alias. In the final release of OpenSSL 0.9.7, the group
14450 alias is called "AES" and is part of "ALL".)
14452 *Ben Laurie, Steve Henson, Bodo Moeller*
14454 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14455 request to response.
14459 * Functions for OCSP responders. OCSP_request_onereq_count(),
14460 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14461 extract information from a certificate request. OCSP_response_create()
14462 creates a response and optionally adds a basic response structure.
14463 OCSP_basic_add1_status() adds a complete single response to a basic
14464 response and returns the OCSP_SINGLERESP structure just added (to allow
14465 extensions to be included for example). OCSP_basic_add1_cert() adds a
14466 certificate to a basic response and OCSP_basic_sign() signs a basic
14467 response with various flags. New helper functions ASN1_TIME_check()
14468 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14469 (converts ASN1_TIME to GeneralizedTime).
14473 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14474 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14475 structure from a certificate. X509_pubkey_digest() digests the public_key
14476 contents: this is used in various key identifiers.
14480 * Make sk_sort() tolerate a NULL argument.
14482 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14484 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14485 passed by the function are trusted implicitly. If any of them signed the
14486 response then it is assumed to be valid and is not verified.
14490 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14491 to data. This was previously part of the PKCS7 ASN1 code. This
14492 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14493 *Steve Henson, reported by Kenneth R. Robinette
14494 <support@securenetterm.com>*
14496 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14497 routines: without these tracing memory leaks is very painful.
14498 Fix leaks in PKCS12 and PKCS7 routines.
14502 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14503 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14504 effectively meant GeneralizedTime would never be used. Now it
14505 is initialised to -1 but X509_time_adj() now has to check the value
14506 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14507 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14508 *Steve Henson, reported by Kenneth R. Robinette
14509 <support@securenetterm.com>*
14511 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14512 result in a zero length in the ASN1_INTEGER structure which was
14513 not consistent with the structure when d2i_ASN1_INTEGER() was used
14514 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14515 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14516 where it did not print out a minus for negative ASN1_INTEGER.
14520 * Add summary printout to ocsp utility. The various functions which
14521 convert status values to strings have been renamed to:
14522 OCSP_response_status_str(), OCSP_cert_status_str() and
14523 OCSP_crl_reason_str() and are no longer static. New options
14524 to verify nonce values and to disable verification. OCSP response
14525 printout format cleaned up.
14529 * Add additional OCSP certificate checks. These are those specified
14530 in RFC2560. This consists of two separate checks: the CA of the
14531 certificate being checked must either be the OCSP signer certificate
14532 or the issuer of the OCSP signer certificate. In the latter case the
14533 OCSP signer certificate must contain the OCSP signing extended key
14534 usage. This check is performed by attempting to match the OCSP
14535 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14536 in the OCSP_CERTID structures of the response.
14540 * Initial OCSP certificate verification added to OCSP_basic_verify()
14541 and related routines. This uses the standard OpenSSL certificate
14542 verify routines to perform initial checks (just CA validity) and
14543 to obtain the certificate chain. Then additional checks will be
14544 performed on the chain. Currently the root CA is checked to see
14545 if it is explicitly trusted for OCSP signing. This is used to set
14546 a root CA as a global signing root: that is any certificate that
14547 chains to that CA is an acceptable OCSP signing certificate.
14551 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14552 extensions from a separate configuration file.
14553 As when reading extensions from the main configuration file,
14554 the '-extensions ...' option may be used for specifying the
14557 *Massimiliano Pala <madwolf@comune.modena.it>*
14559 * New OCSP utility. Allows OCSP requests to be generated or
14560 read. The request can be sent to a responder and the output
14561 parsed, outputted or printed in text form. Not complete yet:
14562 still needs to check the OCSP response validity.
14566 * New subcommands for 'openssl ca':
14567 `openssl ca -status <serial>` prints the status of the cert with
14568 the given serial number (according to the index file).
14569 `openssl ca -updatedb` updates the expiry status of certificates
14572 *Massimiliano Pala <madwolf@comune.modena.it>*
14574 * New '-newreq-nodes' command option to CA.pl. This is like
14575 '-newreq', but calls 'openssl req' with the '-nodes' option
14576 so that the resulting key is not encrypted.
14578 *Damien Miller <djm@mindrot.org>*
14580 * New configuration for the GNU Hurd.
14582 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14584 * Initial code to implement OCSP basic response verify. This
14585 is currently incomplete. Currently just finds the signer's
14586 certificate and verifies the signature on the response.
14590 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14591 value of OPENSSLDIR. This is available via the new '-d' option
14592 to 'openssl version', and is also included in 'openssl version -a'.
14596 * Allowing defining memory allocation callbacks that will be given
14597 file name and line number information in additional arguments
14598 (a `const char*` and an int). The basic functionality remains, as
14599 well as the original possibility to just replace malloc(),
14600 realloc() and free() by functions that do not know about these
14601 additional arguments. To register and find out the current
14602 settings for extended allocation functions, the following
14603 functions are provided:
14605 CRYPTO_set_mem_ex_functions
14606 CRYPTO_set_locked_mem_ex_functions
14607 CRYPTO_get_mem_ex_functions
14608 CRYPTO_get_locked_mem_ex_functions
14610 These work the same way as CRYPTO_set_mem_functions and friends.
14611 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14612 extended allocation function is enabled.
14613 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14614 a conventional allocation function is enabled.
14616 *Richard Levitte, Bodo Moeller*
14618 * Finish off removing the remaining LHASH function pointer casts.
14619 There should no longer be any prototype-casting required when using
14620 the LHASH abstraction, and any casts that remain are "bugs". See
14621 the callback types and macros at the head of lhash.h for details
14622 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14626 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14627 If /dev/[u]random devices are not available or do not return enough
14628 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14630 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14631 /etc/entropy will be queried once each in this sequence, querying stops
14632 when enough entropy was collected without querying more sockets.
14636 * Change the Unix RAND_poll() variant to be able to poll several
14637 random devices, as specified by DEVRANDOM, until a sufficient amount
14638 of data has been collected. We spend at most 10 ms on each file
14639 (select timeout) and read in non-blocking mode. DEVRANDOM now
14640 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14641 (previously it was just the string "/dev/urandom"), so on typical
14642 platforms the 10 ms delay will never occur.
14643 Also separate out the Unix variant to its own file, rand_unix.c.
14644 For VMS, there's a currently-empty rand_vms.c.
14648 * Move OCSP client related routines to ocsp_cl.c. These
14649 provide utility functions which an application needing
14650 to issue a request to an OCSP responder and analyse the
14651 response will typically need: as opposed to those which an
14652 OCSP responder itself would need which will be added later.
14654 OCSP_request_sign() signs an OCSP request with an API similar
14655 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14656 response. OCSP_response_get1_basic() extracts basic response
14657 from response. OCSP_resp_find_status(): finds and extracts status
14658 information from an OCSP_CERTID structure (which will be created
14659 when the request structure is built). These are built from lower
14660 level functions which work on OCSP_SINGLERESP structures but
14661 won't normally be used unless the application wishes to examine
14662 extensions in the OCSP response for example.
14664 Replace nonce routines with a pair of functions.
14665 OCSP_request_add1_nonce() adds a nonce value and optionally
14666 generates a random value. OCSP_check_nonce() checks the
14667 validity of the nonce in an OCSP response.
14671 * Change function OCSP_request_add() to OCSP_request_add0_id().
14672 This doesn't copy the supplied OCSP_CERTID and avoids the
14673 need to free up the newly created id. Change return type
14674 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14675 This can then be used to add extensions to the request.
14676 Deleted OCSP_request_new(), since most of its functionality
14677 is now in OCSP_REQUEST_new() (and the case insensitive name
14678 clash) apart from the ability to set the request name which
14679 will be added elsewhere.
14683 * Update OCSP API. Remove obsolete extensions argument from
14684 various functions. Extensions are now handled using the new
14685 OCSP extension code. New simple OCSP HTTP function which
14686 can be used to send requests and parse the response.
14690 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14691 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14692 uses the special reorder version of SET OF to sort the attributes
14693 and reorder them to match the encoded order. This resolves a long
14694 standing problem: a verify on a PKCS7 structure just after signing
14695 it used to fail because the attribute order did not match the
14696 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14697 it uses the received order. This is necessary to tolerate some broken
14698 software that does not order SET OF. This is handled by encoding
14699 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14700 to produce the required SET OF.
14704 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14705 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14706 files to get correct declarations of the ASN.1 item variables.
14710 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14711 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14712 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14713 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14714 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14715 ASN1_ITEM and no wrapper functions.
14719 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14720 replace the old function pointer based I/O routines. Change most of
14721 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14725 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14726 lines, recognize more "algorithms" that can be deselected, and make
14727 it complain about algorithm deselection that isn't recognised.
14731 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14732 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14733 to use new functions. Add NO_ASN1_OLD which can be set to remove
14734 some old style ASN1 functions: this can be used to determine if old
14735 code will still work when these eventually go away.
14739 * New extension functions for OCSP structures, these follow the
14740 same conventions as certificates and CRLs.
14744 * New function X509V3_add1_i2d(). This automatically encodes and
14745 adds an extension. Its behaviour can be customised with various
14746 flags to append, replace or delete. Various wrappers added for
14747 certificates and CRLs.
14751 * Fix to avoid calling the underlying ASN1 print routine when
14752 an extension cannot be parsed. Correct a typo in the
14753 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14757 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14758 entries for variables.
14762 * Add functionality to `apps/openssl.c` for detecting locking
14763 problems: As the program is single-threaded, all we have
14764 to do is register a locking callback using an array for
14765 storing which locks are currently held by the program.
14769 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14770 SSL_get_ex_data_X509_STORE_idx(), which is used in
14771 ssl_verify_cert_chain() and thus can be called at any time
14772 during TLS/SSL handshakes so that thread-safety is essential.
14773 Unfortunately, the ex_data design is not at all suited
14774 for multi-threaded use, so it probably should be abolished.
14778 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14780 *Broadcom, tweaked and integrated by Geoff Thorpe*
14782 * Move common extension printing code to new function
14783 X509V3_print_extensions(). Reorganise OCSP print routines and
14784 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14788 * New function X509_signature_print() to remove duplication in some
14793 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14794 set (this was treated exactly the same as SET OF previously). This
14795 is used to reorder the STACK representing the structure to match the
14796 encoding. This will be used to get round a problem where a PKCS7
14797 structure which was signed could not be verified because the STACK
14798 order did not reflect the encoded order.
14802 * Reimplement the OCSP ASN1 module using the new code.
14806 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14807 for its ASN1 operations. The old style function pointers still exist
14808 for now but they will eventually go away.
14812 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14813 completely replaces the old ASN1 functionality with a table driven
14814 encoder and decoder which interprets an ASN1_ITEM structure describing
14815 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14816 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14817 has also been converted to the new form.
14821 * Change BN_mod_exp_recp so that negative moduli are tolerated
14822 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14823 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14824 for negative moduli.
14828 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14829 of not touching the result's sign bit.
14833 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14838 * Changed the LHASH code to use prototypes for callbacks, and created
14839 macros to declare and implement thin (optionally static) functions
14840 that provide type-safety and avoid function pointer casting for the
14841 type-specific callbacks.
14845 * Added Kerberos Cipher Suites to be used with TLS, as written in
14847 *Veers Staats <staatsvr@asc.hpc.mil>,
14848 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14850 * Reformat the FAQ so the different questions and answers can be divided
14851 in sections depending on the subject.
14855 * Have the zlib compression code load ZLIB.DLL dynamically under
14860 * New function BN_mod_sqrt for computing square roots modulo a prime
14861 (using the probabilistic Tonelli-Shanks algorithm unless
14862 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14863 be handled deterministically).
14865 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14867 * Make BN_mod_inverse faster by explicitly handling small quotients
14868 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14869 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14873 * New function BN_kronecker.
14877 * Fix BN_gcd so that it works on negative inputs; the result is
14878 positive unless both parameters are zero.
14879 Previously something reasonably close to an infinite loop was
14880 possible because numbers could be growing instead of shrinking
14881 in the implementation of Euclid's algorithm.
14885 * Fix BN_is_word() and BN_is_one() macros to take into account the
14886 sign of the number in question.
14888 Fix BN_is_word(a,w) to work correctly for w == 0.
14890 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14891 because its test if the absolute value of 'a' equals 'w'.
14892 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14893 it exists mostly for use in the implementations of BN_is_zero(),
14894 BN_is_one(), and BN_is_word().
14898 * New function BN_swap.
14902 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14903 the exponentiation functions are more likely to produce reasonable
14904 results on negative inputs.
14908 * Change BN_mod_mul so that the result is always non-negative.
14909 Previously, it could be negative if one of the factors was negative;
14910 I don't think anyone really wanted that behaviour.
14914 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14915 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14916 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14917 and add new functions:
14926 BN_mod_lshift1_quick
14928 BN_mod_lshift_quick
14930 These functions always generate non-negative results.
14932 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14933 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14935 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14936 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14937 be reduced modulo `m`.
14939 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14942 The following entry accidentally appeared in the CHANGES file
14943 distributed with OpenSSL 0.9.7. The modifications described in
14944 it do *not* apply to OpenSSL 0.9.7.
14946 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14947 was actually never needed) and in BN_mul(). The removal in BN_mul()
14948 required a small change in bn_mul_part_recursive() and the addition
14949 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14950 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14951 bn_sub_words() and bn_add_words() except they take arrays with
14957 * In 'openssl passwd', verify passwords read from the terminal
14958 unless the '-salt' option is used (which usually means that
14959 verification would just waste user's time since the resulting
14960 hash is going to be compared with some given password hash)
14961 or the new '-noverify' option is used.
14963 This is an incompatible change, but it does not affect
14964 non-interactive use of 'openssl passwd' (passwords on the command
14965 line, '-stdin' option, '-in ...' option) and thus should not
14966 cause any problems.
14970 * Remove all references to RSAref, since there's no more need for it.
14974 * Make DSO load along a path given through an environment variable
14975 (SHLIB_PATH) with shl_load().
14979 * Constify the ENGINE code as a result of BIGNUM constification.
14980 Also constify the RSA code and most things related to it. In a
14981 few places, most notable in the depth of the ASN.1 code, ugly
14982 casts back to non-const were required (to be solved at a later
14987 * Make it so the openssl application has all engines loaded by default.
14991 * Constify the BIGNUM routines a little more.
14995 * Add the following functions:
14997 ENGINE_load_cswift()
14999 ENGINE_load_atalla()
15000 ENGINE_load_nuron()
15001 ENGINE_load_builtin_engines()
15003 That way, an application can itself choose if external engines that
15004 are built-in in OpenSSL shall ever be used or not. The benefit is
15005 that applications won't have to be linked with libdl or other dso
15006 libraries unless it's really needed.
15008 Changed 'openssl engine' to load all engines on demand.
15009 Changed the engine header files to avoid the duplication of some
15010 declarations (they differed!).
15014 * 'openssl engine' can now list capabilities.
15018 * Better error reporting in 'openssl engine'.
15022 * Never call load_dh_param(NULL) in s_server.
15026 * Add engine application. It can currently list engines by name and
15027 identity, and test if they are actually available.
15031 * Improve RPM specification file by forcing symbolic linking and making
15032 sure the installed documentation is also owned by root.root.
15034 *Damien Miller <djm@mindrot.org>*
15036 * Give the OpenSSL applications more possibilities to make use of
15037 keys (public as well as private) handled by engines.
15041 * Add OCSP code that comes from CertCo.
15045 * Add VMS support for the Rijndael code.
15049 * Added untested support for Nuron crypto accelerator.
15053 * Add support for external cryptographic devices. This code was
15054 previously distributed separately as the "engine" branch.
15056 *Geoff Thorpe, Richard Levitte*
15058 * Rework the filename-translation in the DSO code. It is now possible to
15059 have far greater control over how a "name" is turned into a filename
15060 depending on the operating environment and any oddities about the
15061 different shared library filenames on each system.
15065 * Support threads on FreeBSD-elf in Configure.
15069 * Fix for SHA1 assembly problem with MASM: it produces
15070 warnings about corrupt line number information when assembling
15071 with debugging information. This is caused by the overlapping
15074 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15077 NCONF_get_number() has no error checking at all. As a replacement,
15078 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15079 promoted strongly. The old NCONF_get_number is kept around for
15080 binary backward compatibility.
15081 Make it possible for methods to load from something other than a BIO,
15082 by providing a function pointer that is given a name instead of a BIO.
15083 For example, this could be used to load configuration data from an
15088 * Fix for non blocking accept BIOs. Added new I/O special reason
15089 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15090 with non blocking I/O was not possible because no retry code was
15091 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15096 * Added the beginnings of Rijndael support.
15100 * Fix for bug in DirectoryString mask setting. Add support for
15101 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15102 to allow certificate printing to more controllable, additional
15103 'certopt' option to 'x509' to allow new printing options to be
15108 * Clean old EAY MD5 hack from e_os.h.
15112 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15114 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15115 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15117 *Joe Orton, Steve Henson*
15119 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15121 * Fix additional bug revealed by the NISCC test suite:
15123 Stop bug triggering large recursion when presented with
15124 certain ASN.1 tags ([CVE-2003-0851])
15128 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15130 * Fix various bugs revealed by running the NISCC test suite:
15132 Stop out of bounds reads in the ASN1 code when presented with
15133 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15135 If verify callback ignores invalid public key errors don't try to check
15136 certificate signature with the NULL public key.
15140 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15141 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15146 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15147 extra data after the compression methods not only for TLS 1.0
15148 but also for SSL 3.0 (as required by the specification).
15150 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15152 * Change X509_certificate_type() to mark the key as exported/exportable
15153 when it's 512 *bits* long, not 512 bytes.
15157 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15159 * Countermeasure against the Klima-Pokorny-Rosa extension of
15160 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15161 a protocol version number mismatch like a decryption error
15162 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15166 * Turn on RSA blinding by default in the default implementation
15167 to avoid a timing attack. Applications that don't want it can call
15168 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15169 They would be ill-advised to do so in most cases.
15171 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15173 * Change RSA blinding code so that it works when the PRNG is not
15174 seeded (in this case, the secret RSA exponent is abused as
15175 an unpredictable seed -- if it is not unpredictable, there
15176 is no point in blinding anyway). Make RSA blinding thread-safe
15177 by remembering the creator's thread ID in rsa->blinding and
15178 having all other threads use local one-time blinding factors
15179 (this requires more computation than sharing rsa->blinding, but
15180 avoids excessive locking; and if an RSA object is not shared
15181 between threads, blinding will still be very fast).
15185 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15187 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15188 via timing by performing a MAC computation even if incorrect
15189 block cipher padding has been found. This is a countermeasure
15190 against active attacks where the attacker has to distinguish
15191 between bad padding and a MAC verification error. ([CVE-2003-0078])
15193 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15194 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15195 Martin Vuagnoux (EPFL, Ilion)*
15197 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15199 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15200 memory from its contents. This is done with a counter that will
15201 place alternating values in each byte. This can be used to solve
15202 two issues: 1) the removal of calls to memset() by highly optimizing
15203 compilers, and 2) cleansing with other values than 0, since those can
15204 be read through on certain media, for example a swap space on disk.
15208 * Bugfix: client side session caching did not work with external caching,
15209 because the session->cipher setting was not restored when reloading
15210 from the external cache. This problem was masked, when
15211 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15212 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15216 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15217 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15219 *Zeev Lieber <zeev-l@yahoo.com>*
15221 * Undo an undocumented change introduced in 0.9.6e which caused
15222 repeated calls to OpenSSL_add_all_ciphers() and
15223 OpenSSL_add_all_digests() to be ignored, even after calling
15228 * Change the default configuration reader to deal with last line not
15229 being properly terminated.
15233 * Change X509_NAME_cmp() so it applies the special rules on handling
15234 DN values that are of type PrintableString, as well as RDNs of type
15235 emailAddress where the value has the type ia5String.
15237 *stefank@valicert.com via Richard Levitte*
15239 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15240 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15241 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15242 the bitwise-OR of the two for use by the majority of applications
15243 wanting this behaviour, and update the docs. The documented
15244 behaviour and actual behaviour were inconsistent and had been
15245 changing anyway, so this is more a bug-fix than a behavioural
15248 *Geoff Thorpe, diagnosed by Nadav Har'El*
15250 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15251 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15255 * Fix initialization code race conditions in
15256 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15257 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15258 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15259 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15260 ssl2_get_cipher_by_char(),
15261 ssl3_get_cipher_by_char().
15263 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15265 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15266 the cached sessions are flushed, as the remove_cb() might use ex_data
15267 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15268 (see [openssl.org #212]).
15270 *Geoff Thorpe, Lutz Jaenicke*
15272 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15273 length, instead of the encoding length to d2i_ASN1_OBJECT.
15277 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15279 * [In 0.9.6g-engine release:]
15280 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15282 *Lynn Gazis <lgazis@rainbow.com>*
15284 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15286 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15287 and get fix the header length calculation.
15288 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15289 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15291 * Use proper error handling instead of 'assertions' in buffer
15292 overflow checks added in 0.9.6e. This prevents DoS (the
15293 assertions could call abort()).
15295 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15297 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15299 * Add various sanity checks to asn1_get_length() to reject
15300 the ASN1 length bytes if they exceed sizeof(long), will appear
15301 negative or the content length exceeds the length of the
15304 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15306 * Fix cipher selection routines: ciphers without encryption had no flags
15307 for the cipher strength set and where therefore not handled correctly
15308 by the selection routines (PR #130).
15312 * Fix EVP_dsa_sha macro.
15317 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15318 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15319 that was added in OpenSSL 0.9.6d.
15321 As the countermeasure turned out to be incompatible with some
15322 broken SSL implementations, the new option is part of SSL_OP_ALL.
15323 SSL_OP_ALL is usually employed when compatibility with weird SSL
15324 implementations is desired (e.g. '-bugs' option to 's_client' and
15325 's_server'), so the new option is automatically set in many
15330 * Changes in security patch:
15332 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15333 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15334 Air Force Materiel Command, USAF, under agreement number
15337 * Add various sanity checks to asn1_get_length() to reject
15338 the ASN1 length bytes if they exceed sizeof(long), will appear
15339 negative or the content length exceeds the length of the
15340 supplied buffer. ([CVE-2002-0659])
15342 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15344 * Assertions for various potential buffer overflows, not known to
15345 happen in practice.
15347 *Ben Laurie (CHATS)*
15349 * Various temporary buffers to hold ASCII versions of integers were
15350 too small for 64 bit platforms. ([CVE-2002-0655])
15351 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15353 * Remote buffer overflow in SSL3 protocol - an attacker could
15354 supply an oversized session ID to a client. ([CVE-2002-0656])
15356 *Ben Laurie (CHATS)*
15358 * Remote buffer overflow in SSL2 protocol - an attacker could
15359 supply an oversized client master key. ([CVE-2002-0656])
15361 *Ben Laurie (CHATS)*
15363 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15365 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15366 encoded as NULL) with id-dsa-with-sha1.
15368 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15370 * Check various `X509_...()` return values in `apps/req.c`.
15372 *Nils Larsch <nla@trustcenter.de>*
15374 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15375 an end-of-file condition would erroneously be flagged, when the CRLF
15376 was just at the end of a processed block. The bug was discovered when
15377 processing data through a buffering memory BIO handing the data to a
15378 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15379 <ptsekov@syntrex.com> and Nedelcho Stanev.
15383 * Implement a countermeasure against a vulnerability recently found
15384 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15385 before application data chunks to avoid the use of known IVs
15386 with data potentially chosen by the attacker.
15390 * Fix length checks in ssl3_get_client_hello().
15394 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15395 to prevent ssl3_read_internal() from incorrectly assuming that
15396 ssl3_read_bytes() found application data while handshake
15397 processing was enabled when in fact s->s3->in_read_app_data was
15398 merely automatically cleared during the initial handshake.
15400 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15402 * Fix object definitions for Private and Enterprise: they were not
15403 recognized in their shortname (=lowercase) representation. Extend
15404 obj_dat.pl to issue an error when using undefined keywords instead
15405 of silently ignoring the problem (Svenning Sorensen
15406 <sss@sss.dnsalias.net>).
15410 * Fix DH_generate_parameters() so that it works for 'non-standard'
15411 generators, i.e. generators other than 2 and 5. (Previously, the
15412 code did not properly initialise the 'add' and 'rem' values to
15413 BN_generate_prime().)
15415 In the new general case, we do not insist that 'generator' is
15416 actually a primitive root: This requirement is rather pointless;
15417 a generator of the order-q subgroup is just as good, if not
15422 * Map new X509 verification errors to alerts. Discovered and submitted by
15423 Tom Wu <tom@arcot.com>.
15427 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15428 returning non-zero before the data has been completely received
15429 when using non-blocking I/O.
15431 *Bodo Moeller; problem pointed out by John Hughes*
15433 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15435 *Ben Laurie, Lutz Jaenicke*
15437 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15438 Yoram Zahavi <YoramZ@gilian.com>).
15442 * Add information about CygWin 1.3 and on, and preserve proper
15443 configuration for the versions before that.
15445 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15447 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15448 check whether we deal with a copy of a session and do not delete from
15449 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15450 <izhar@checkpoint.com>.
15454 * Do not store session data into the internal session cache, if it
15455 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15456 flag is set). Proposed by Aslam <aslam@funk.com>.
15460 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15465 * [In 0.9.6d-engine release:]
15466 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15468 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15470 * Add the configuration target linux-s390x.
15472 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15474 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15475 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15476 variable as an indication that a ClientHello message has been
15477 received. As the flag value will be lost between multiple
15478 invocations of ssl3_accept when using non-blocking I/O, the
15479 function may not be aware that a handshake has actually taken
15480 place, thus preventing a new session from being added to the
15483 To avoid this problem, we now set s->new_session to 2 instead of
15484 using a local variable.
15486 *Lutz Jaenicke, Bodo Moeller*
15488 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15489 if the SSL_R_LENGTH_MISMATCH error is detected.
15491 *Geoff Thorpe, Bodo Moeller*
15493 * New 'shared_ldflag' column in Configure platform table.
15497 * Fix EVP_CIPHER_mode macro.
15499 *"Dan S. Camper" <dan@bti.net>*
15501 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15502 type, we must throw them away by setting rr->length to 0.
15504 *D P Chang <dpc@qualys.com>*
15506 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15508 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15509 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15510 worked incorrectly for those cases where range = `10..._2` and
15511 `3*range` is two bits longer than range.)
15515 * Only add signing time to PKCS7 structures if it is not already
15520 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15521 OBJ_ld_ce should be OBJ_id_ce.
15522 Also some ip-pda OIDs in crypto/objects/objects.txt were
15523 incorrect (cf. RFC 3039).
15525 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15527 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15528 returns early because it has nothing to do.
15530 *Andy Schneider <andy.schneider@bjss.co.uk>*
15532 * [In 0.9.6c-engine release:]
15533 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15535 *Andy Schneider <andy.schneider@bjss.co.uk>*
15537 * [In 0.9.6c-engine release:]
15538 Add support for Cryptographic Appliance's keyserver technology.
15539 (Use engine 'keyclient')
15541 *Cryptographic Appliances and Geoff Thorpe*
15543 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15544 is called via tools/c89.sh because arguments have to be
15545 rearranged (all '-L' options must appear before the first object
15548 *Richard Shapiro <rshapiro@abinitio.com>*
15550 * [In 0.9.6c-engine release:]
15551 Add support for Broadcom crypto accelerator cards, backported
15554 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15556 * [In 0.9.6c-engine release:]
15557 Add support for SureWare crypto accelerator cards from
15558 Baltimore Technologies. (Use engine 'sureware')
15560 *Baltimore Technologies and Mark Cox*
15562 * [In 0.9.6c-engine release:]
15563 Add support for crypto accelerator cards from Accelerated
15564 Encryption Processing, www.aep.ie. (Use engine 'aep')
15566 *AEP Inc. and Mark Cox*
15568 * Add a configuration entry for gcc on UnixWare.
15570 *Gary Benson <gbenson@redhat.com>*
15572 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15573 messages are stored in a single piece (fixed-length part and
15574 variable-length part combined) and fix various bugs found on the way.
15578 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15579 instead. BIO_gethostbyname() does not know what timeouts are
15580 appropriate, so entries would stay in cache even when they have
15582 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15584 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15585 faced with a pathologically small ClientHello fragment that does
15586 not contain client_version: Instead of aborting with an error,
15587 simply choose the highest available protocol version (i.e.,
15588 TLS 1.0 unless it is disabled). In practice, ClientHello
15589 messages are never sent like this, but this change gives us
15590 strictly correct behaviour at least for TLS.
15594 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15595 never resets s->method to s->ctx->method when called from within
15596 one of the SSL handshake functions.
15598 *Bodo Moeller; problem pointed out by Niko Baric*
15600 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15601 (sent using the client's version number) if client_version is
15602 smaller than the protocol version in use. Also change
15603 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15604 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15605 the client will at least see that alert.
15609 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15614 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15615 client receives HelloRequest while in a handshake.
15617 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15619 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15620 should end in 'break', not 'goto end' which circumvents various
15621 cleanups done in state SSL_ST_OK. But session related stuff
15622 must be disabled for SSL_ST_OK in the case that we just sent a
15625 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15626 before just sending a HelloRequest.
15628 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15630 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15631 reveal whether illegal block cipher padding was found or a MAC
15632 verification error occurred. (Neither SSLerr() codes nor alerts
15633 are directly visible to potential attackers, but the information
15634 may leak via logfiles.)
15636 Similar changes are not required for the SSL 2.0 implementation
15637 because the number of padding bytes is sent in clear for SSL 2.0,
15638 and the extra bytes are just ignored. However ssl/s2_pkt.c
15639 failed to verify that the purported number of padding bytes is in
15644 * Add OpenUNIX-8 support including shared libraries
15645 (Boyd Lynn Gerber <gerberb@zenez.com>).
15649 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15650 'wristwatch attack' using huge encoding parameters (cf.
15651 James H. Manger's CRYPTO 2001 paper). Note that the
15652 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15653 encoding parameters and hence was not vulnerable.
15657 * BN_sqr() bug fix.
15659 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15661 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15662 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15663 followed by modular reduction.
15665 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15667 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15668 equivalent based on BN_pseudo_rand() instead of BN_rand().
15672 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15673 This function was broken, as the check for a new client hello message
15674 to handle SGC did not allow these large messages.
15675 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15679 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15683 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15684 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15688 * Rework the configuration and shared library support for Tru64 Unix.
15689 The configuration part makes use of modern compiler features and
15690 still retains old compiler behavior for those that run older versions
15691 of the OS. The shared library support part includes a variant that
15692 uses the RPATH feature, and is available through the special
15693 configuration target "alpha-cc-rpath", which will never be selected
15696 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15698 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15699 with the same message size as in ssl3_get_certificate_request().
15700 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15701 messages might inadvertently be reject as too long.
15703 *Petr Lampa <lampa@fee.vutbr.cz>*
15705 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15709 * Modified SSL library such that the verify_callback that has been set
15710 specifically for an SSL object with SSL_set_verify() is actually being
15711 used. Before the change, a verify_callback set with this function was
15712 ignored and the verify_callback() set in the SSL_CTX at the time of
15713 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15714 to allow the necessary settings.
15718 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15719 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15720 done automatically (in contradiction to the requirements of the C
15721 standard). This made problems when used from OpenSSH.
15725 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15726 dh->length and always used
15728 BN_rand_range(priv_key, dh->p).
15730 BN_rand_range() is not necessary for Diffie-Hellman, and this
15731 specific range makes Diffie-Hellman unnecessarily inefficient if
15732 dh->length (recommended exponent length) is much smaller than the
15733 length of dh->p. We could use BN_rand_range() if the order of
15734 the subgroup was stored in the DH structure, but we only have
15739 BN_rand(priv_key, l, ...)
15741 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15748 RSA_eay_public_encrypt
15749 RSA_eay_private_decrypt
15750 RSA_eay_private_encrypt (signing)
15751 RSA_eay_public_decrypt (signature verification)
15753 (default implementations for RSA_public_encrypt,
15754 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15755 always reject numbers >= n.
15759 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15760 to synchronize access to 'locking_thread'. This is necessary on
15761 systems where access to 'locking_thread' (an 'unsigned long'
15762 variable) is not atomic.
15766 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15767 *before* setting the 'crypto_lock_rand' flag. The previous code had
15768 a race condition if 0 is a valid thread ID.
15770 *Travis Vitek <vitek@roguewave.com>*
15772 * Add support for shared libraries under Irix.
15774 *Albert Chin-A-Young <china@thewrittenword.com>*
15776 * Add configuration option to build on Linux on both big-endian and
15777 little-endian MIPS.
15779 *Ralf Baechle <ralf@uni-koblenz.de>*
15781 * Add the possibility to create shared libraries on HP-UX.
15785 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15787 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15788 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15789 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15790 PRNG state recovery was possible based on the output of
15791 one PRNG request appropriately sized to gain knowledge on
15792 'md' followed by enough consecutive 1-byte PRNG requests
15793 to traverse all of 'state'.
15795 1. When updating 'md_local' (the current thread's copy of 'md')
15796 during PRNG output generation, hash all of the previous
15797 'md_local' value, not just the half used for PRNG output.
15799 2. Make the number of bytes from 'state' included into the hash
15800 independent from the number of PRNG bytes requested.
15802 The first measure alone would be sufficient to avoid
15803 Markku-Juhani's attack. (Actually it had never occurred
15804 to me that the half of 'md_local' used for chaining was the
15805 half from which PRNG output bytes were taken -- I had always
15806 assumed that the secret half would be used.) The second
15807 measure makes sure that additional data from 'state' is never
15808 mixed into 'md_local' in small portions; this heuristically
15809 further strengthens the PRNG.
15813 * Fix crypto/bn/asm/mips3.s.
15817 * When only the key is given to "enc", the IV is undefined. Print out
15818 an error message in this case.
15822 * Handle special case when X509_NAME is empty in X509 printing routines.
15826 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15827 positive and less than q.
15831 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15832 used: it isn't thread safe and the add_lock_callback should handle
15835 *Paul Rose <Paul.Rose@bridge.com>*
15837 * Verify that incoming data obeys the block size in
15838 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15844 *Ulf Möller, Bodo Möller*
15846 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15847 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15848 when fixing the server behaviour for backwards-compatible 'client
15849 hello' messages. (Note that the attack is impractical against
15850 SSL 3.0 and TLS 1.0 anyway because length and version checking
15851 means that the probability of guessing a valid ciphertext is
15852 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15855 Before 0.9.5, the countermeasure (hide the error by generating a
15856 random 'decryption result') did not work properly because
15857 ERR_clear_error() was missing, meaning that SSL_get_error() would
15858 detect the supposedly ignored error.
15860 Both problems are now fixed.
15864 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15865 (previously it was 1024).
15869 * Fix for compatibility mode trust settings: ignore trust settings
15870 unless some valid trust or reject settings are present.
15874 * Fix for blowfish EVP: its a variable length cipher.
15878 * Fix various bugs related to DSA S/MIME verification. Handle missing
15879 parameters in DSA public key structures and return an error in the
15880 DSA routines if parameters are absent.
15884 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15885 in the current directory if neither $RANDFILE nor $HOME was set.
15886 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15887 caused some confusion to Windows users who haven't defined $HOME.
15888 Thus RAND_file_name() is changed again: e_os.h can define a
15889 DEFAULT_HOME, which will be used if $HOME is not set.
15890 For Windows, we use "C:"; on other platforms, we still require
15891 environment variables.
15893 * Move 'if (!initialized) RAND_poll()' into regions protected by
15894 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15895 having multiple threads call RAND_poll() concurrently.
15899 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15900 combination of a flag and a thread ID variable.
15901 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15902 flag), *other* threads can enter ssleay_add_bytes without obeying
15903 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15904 that they do not hold after the first thread unsets add_do_not_lock).
15908 * Change bctest again: '-x' expressions are not available in all
15909 versions of 'test'.
15913 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15915 * Fix a couple of memory leaks in PKCS7_dataDecode()
15917 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15919 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15920 the default extension for executables, if any. Also, make the perl
15921 scripts that use symlink() to test if it really exists and use "cp"
15922 if it doesn't. All this made OpenSSL compilable and installable in
15927 * Fix for asn1_GetSequence() for indefinite length constructed data.
15928 If SEQUENCE is length is indefinite just set c->slen to the total
15929 amount of data available.
15931 *Steve Henson, reported by shige@FreeBSD.org*
15933 *This change does not apply to 0.9.7.*
15935 * Change bctest to avoid here-documents inside command substitution
15936 (workaround for FreeBSD /bin/sh bug).
15937 For compatibility with Ultrix, avoid shell functions (introduced
15938 in the bctest version that searches along $PATH).
15942 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15943 with des_encrypt() defined on some operating systems, like Solaris
15948 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15949 On the Importance of Eliminating Errors in Cryptographic
15950 Computations, J. Cryptology 14 (2001) 2, 101-119,
15951 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15955 * MIPS assembler BIGNUM division bug fix.
15959 * Disabled incorrect Alpha assembler code.
15963 * Fix PKCS#7 decode routines so they correctly update the length
15964 after reading an EOC for the EXPLICIT tag.
15968 *This change does not apply to 0.9.7.*
15970 * Fix bug in PKCS#12 key generation routines. This was triggered
15971 if a 3DES key was generated with a 0 initial byte. Include
15972 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15973 (but broken) behaviour.
15977 * Enhance bctest to search for a working bc along $PATH and print
15980 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15982 * Fix memory leaks in err.c: free err_data string if necessary;
15983 don't write to the wrong index in ERR_set_error_data.
15987 * Implement ssl23_peek (analogous to ssl23_read), which previously
15992 * Replace rdtsc with `_emit` statements for VC++ version 5.
15994 *Jeremy Cooper <jeremy@baymoo.org>*
15996 * Make it possible to reuse SSLv2 sessions.
16000 * In copy_email() check for >= 0 as a return value for
16001 X509_NAME_get_index_by_NID() since 0 is a valid index.
16003 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16005 * Avoid coredump with unsupported or invalid public keys by checking if
16006 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16007 PKCS7_verify() fails with non detached data.
16011 * Don't use getenv in library functions when run as setuid/setgid.
16012 New function OPENSSL_issetugid().
16016 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16017 due to incorrect handling of multi-threading:
16019 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16021 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16023 3. Count how many times MemCheck_off() has been called so that
16024 nested use can be treated correctly. This also avoids
16025 inband-signalling in the previous code (which relied on the
16026 assumption that thread ID 0 is impossible).
16030 * Add "-rand" option also to s_client and s_server.
16034 * Fix CPU detection on Irix 6.x.
16035 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16036 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16038 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16043 *This change does not apply to 0.9.7.*
16045 * Use the cached encoding of an X509_NAME structure rather than
16046 copying it. This is apparently the reason for the libsafe "errors"
16047 but the code is actually correct.
16051 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16052 Bleichenbacher's DSA attack.
16053 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16054 to be set and top=0 forces the highest bit to be set; top=-1 is new
16055 and leaves the highest bit random.
16057 *Ulf Moeller, Bodo Moeller*
16059 * In the `NCONF_...`-based implementations for `CONF_...` queries
16060 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16061 a temporary CONF structure with the data component set to NULL
16062 (which gives segmentation faults in lh_retrieve).
16063 Instead, use NULL for the CONF pointer in CONF_get_string and
16064 CONF_get_number (which may use environment variables) and directly
16065 return NULL from CONF_get_section.
16069 * Fix potential buffer overrun for EBCDIC.
16073 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16074 keyUsage if basicConstraints absent for a CA.
16078 * Make SMIME_write_PKCS7() write mail header values with a format that
16079 is more generally accepted (no spaces before the semicolon), since
16080 some programs can't parse those values properly otherwise. Also make
16081 sure BIO's that break lines after each write do not create invalid
16086 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16087 macros previously used would not encode an empty SEQUENCE OF
16088 and break the signature.
16092 *This change does not apply to 0.9.7.*
16094 * Zero the premaster secret after deriving the master secret in
16099 * Add some EVP_add_digest_alias registrations (as found in
16100 OpenSSL_add_all_digests()) to SSL_library_init()
16101 aka OpenSSL_add_ssl_algorithms(). This provides improved
16102 compatibility with peers using X.509 certificates
16103 with unconventional AlgorithmIdentifier OIDs.
16107 * Fix for Irix with NO_ASM.
16109 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16111 * ./config script fixes.
16113 *Ulf Moeller, Richard Levitte*
16115 * Fix 'openssl passwd -1'.
16119 * Change PKCS12_key_gen_asc() so it can cope with non null
16120 terminated strings whose length is passed in the passlen
16121 parameter, for example from PEM callbacks. This was done
16122 by adding an extra length parameter to asc2uni().
16124 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16126 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16127 call failed, free the DSA structure.
16131 * Fix to uni2asc() to cope with zero length Unicode strings.
16132 These are present in some PKCS#12 files.
16136 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16137 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16138 when writing a 32767 byte record.
16140 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16142 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16143 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16145 (RSA objects have a reference count access to which is protected
16146 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16147 so they are meant to be shared between threads.)
16148 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16149 "Reddie, Steven" <Steven.Reddie@ca.com>*
16151 * Fix a deadlock in CRYPTO_mem_leaks().
16155 * Use better test patterns in bntest.
16159 * rand_win.c fix for Borland C.
16163 * BN_rshift bugfix for n == 0.
16167 * Add a 'bctest' script that checks for some known 'bc' bugs
16168 so that 'make test' does not abort just because 'bc' is broken.
16172 * Store verify_result within SSL_SESSION also for client side to
16173 avoid potential security hole. (Re-used sessions on the client side
16174 always resulted in verify_result==X509_V_OK, not using the original
16175 result of the server certificate verification.)
16179 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16180 SSL3_RT_APPLICATION_DATA, return 0.
16181 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16186 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16187 releases, have been re-implemented by renaming the previous
16188 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16189 and ssl3_read_internal, respectively, and adding 'peek' parameters
16190 to them. The new ssl[23]_{read,peek} functions are calls to
16191 ssl[23]_read_internal with the 'peek' flag set appropriately.
16192 A 'peek' parameter has also been added to ssl3_read_bytes, which
16193 does the actual work for ssl3_read_internal.
16197 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16198 the method-specific "init()" handler. Also clean up ex_data after
16199 calling the method-specific "finish()" handler. Previously, this was
16200 happening the other way round.
16204 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16205 The previous value, 12, was not always sufficient for BN_mod_exp().
16209 * Make sure that shared libraries get the internal name engine with
16210 the full version number and not just 0. This should mark the
16211 shared libraries as not backward compatible. Of course, this should
16212 be changed again when we can guarantee backward binary compatibility.
16216 * Fix typo in get_cert_by_subject() in by_dir.c
16218 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16220 * Rework the system to generate shared libraries:
16222 - Make note of the expected extension for the shared libraries and
16223 if there is a need for symbolic links from for example libcrypto.so.0
16224 to libcrypto.so.0.9.7. There is extended info in Configure for
16227 - Make as few rebuilds of the shared libraries as possible.
16229 - Still avoid linking the OpenSSL programs with the shared libraries.
16231 - When installing, install the shared libraries separately from the
16236 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16238 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16239 and not in SSL_clear because the latter is also used by the
16240 accept/connect functions; previously, the settings made by
16241 SSL_set_read_ahead would be lost during the handshake.
16243 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16245 * Correct util/mkdef.pl to be selective about disabled algorithms.
16246 Previously, it would create entries for disabled algorithms no
16251 * Added several new manual pages for SSL_* function.
16255 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16257 * In ssl23_get_client_hello, generate an error message when faced
16258 with an initial SSL 3.0/TLS record that is too small to contain the
16259 first two bytes of the ClientHello message, i.e. client_version.
16260 (Note that this is a pathologic case that probably has never happened
16261 in real life.) The previous approach was to use the version number
16262 from the record header as a substitute; but our protocol choice
16263 should not depend on that one because it is not authenticated
16264 by the Finished messages.
16268 * More robust randomness gathering functions for Windows.
16270 *Jeffrey Altman <jaltman@columbia.edu>*
16272 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16273 not set then we don't setup the error code for issuer check errors
16274 to avoid possibly overwriting other errors which the callback does
16275 handle. If an application does set the flag then we assume it knows
16276 what it is doing and can handle the new informational codes
16281 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16282 a general "ANY" type, as such it should be able to decode anything
16283 including tagged types. However it didn't check the class so it would
16284 wrongly interpret tagged types in the same way as their universal
16285 counterpart and unknown types were just rejected. Changed so that the
16286 tagged and unknown types are handled in the same way as a SEQUENCE:
16287 that is the encoding is stored intact. There is also a new type
16288 "V_ASN1_OTHER" which is used when the class is not universal, in this
16289 case we have no idea what the actual type is so we just lump them all
16294 * On VMS, stdout may very well lead to a file that is written to
16295 in a record-oriented fashion. That means that every write() will
16296 write a separate record, which will be read separately by the
16297 programs trying to read from it. This can be very confusing.
16299 The solution is to put a BIO filter in the way that will buffer
16300 text until a linefeed is reached, and then write everything a
16301 line at a time, so every record written will be an actual line,
16302 not chunks of lines and not (usually doesn't happen, but I've
16303 seen it once) several lines in one record. BIO_f_linebuffer() is
16306 Currently, it's a VMS-only method, because that's where it has
16307 been tested well enough.
16311 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16312 it can return incorrect results.
16313 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16314 but it was in 0.9.6-beta[12].)
16318 * Disable the check for content being present when verifying detached
16319 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16320 include zero length content when signing messages.
16324 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16325 BIO_ctrl (for BIO pairs).
16329 * Add DSO method for VMS.
16333 * Bug fix: Montgomery multiplication could produce results with the
16338 * Add RPM specification openssl.spec and modify it to build three
16339 packages. The default package contains applications, application
16340 documentation and run-time libraries. The devel package contains
16341 include files, static libraries and function documentation. The
16342 doc package contains the contents of the doc directory. The original
16343 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16347 * Add a large number of documentation files for many SSL routines.
16349 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16351 * Add a configuration entry for Sony News 4.
16353 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16355 * Don't set the two most significant bits to one when generating a
16356 random number < q in the DSA library.
16360 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16361 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16362 the underlying transport is blocking) if a handshake took place.
16363 (The default behaviour is needed by applications such as s_client
16364 and s_server that use select() to determine when to use SSL_read;
16365 but for applications that know in advance when to expect data, it
16366 just makes things more complicated.)
16370 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16375 * Add a few more EBCDIC conditionals that make `req` and `x509`
16376 work better on such systems.
16378 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16380 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16381 Update PKCS12_parse() so it copies the friendlyName and the
16382 keyid to the certificates aux info.
16386 * Fix bug in PKCS7_verify() which caused an infinite loop
16387 if there was more than one signature.
16389 *Sven Uszpelkat <su@celocom.de>*
16391 * Major change in util/mkdef.pl to include extra information
16392 about each symbol, as well as presenting variables as well
16393 as functions. This change means that there's n more need
16394 to rebuild the .num files when some algorithms are excluded.
16398 * Allow the verify time to be set by an application,
16399 rather than always using the current time.
16403 * Phase 2 verify code reorganisation. The certificate
16404 verify code now looks up an issuer certificate by a
16405 number of criteria: subject name, authority key id
16406 and key usage. It also verifies self signed certificates
16407 by the same criteria. The main comparison function is
16408 X509_check_issued() which performs these checks.
16410 Lot of changes were necessary in order to support this
16411 without completely rewriting the lookup code.
16413 Authority and subject key identifier are now cached.
16415 The LHASH 'certs' is X509_STORE has now been replaced
16416 by a STACK_OF(X509_OBJECT). This is mainly because an
16417 LHASH can't store or retrieve multiple objects with
16418 the same hash value.
16420 As a result various functions (which were all internal
16421 use only) have changed to handle the new X509_STORE
16422 structure. This will break anything that messed round
16423 with X509_STORE internally.
16425 The functions X509_STORE_add_cert() now checks for an
16426 exact match, rather than just subject name.
16428 The X509_STORE API doesn't directly support the retrieval
16429 of multiple certificates matching a given criteria, however
16430 this can be worked round by performing a lookup first
16431 (which will fill the cache with candidate certificates)
16432 and then examining the cache for matches. This is probably
16433 the best we can do without throwing out X509_LOOKUP
16434 entirely (maybe later...).
16436 The X509_VERIFY_CTX structure has been enhanced considerably.
16438 All certificate lookup operations now go via a get_issuer()
16439 callback. Although this currently uses an X509_STORE it
16440 can be replaced by custom lookups. This is a simple way
16441 to bypass the X509_STORE hackery necessary to make this
16442 work and makes it possible to use more efficient techniques
16443 in future. A very simple version which uses a simple
16444 STACK for its trusted certificate store is also provided
16445 using X509_STORE_CTX_trusted_stack().
16447 The verify_cb() and verify() callbacks now have equivalents
16448 in the X509_STORE_CTX structure.
16450 X509_STORE_CTX also has a 'flags' field which can be used
16451 to customise the verify behaviour.
16455 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16456 excludes S/MIME capabilities.
16460 * When a certificate request is read in keep a copy of the
16461 original encoding of the signed data and use it when outputting
16462 again. Signatures then use the original encoding rather than
16463 a decoded, encoded version which may cause problems if the
16464 request is improperly encoded.
16468 * For consistency with other BIO_puts implementations, call
16469 buffer_write(b, ...) directly in buffer_puts instead of calling
16472 In BIO_puts, increment b->num_write as in BIO_write.
16474 *Peter.Sylvester@EdelWeb.fr*
16476 * Fix BN_mul_word for the case where the word is 0. (We have to use
16477 BN_zero, we may not return a BIGNUM with an array consisting of
16478 words set to zero.)
16482 * Avoid calling abort() from within the library when problems are
16483 detected, except if preprocessor symbols have been defined
16484 (such as REF_CHECK, BN_DEBUG etc.).
16488 * New openssl application 'rsautl'. This utility can be
16489 used for low-level RSA operations. DER public key
16490 BIO/fp routines also added.
16494 * New Configure entry and patches for compiling on QNX 4.
16496 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16498 * A demo state-machine implementation was sponsored by
16499 Nuron (<http://www.nuron.com/>) and is now available in
16500 demos/state_machine.
16504 * New options added to the 'dgst' utility for signature
16505 generation and verification.
16509 * Unrecognized PKCS#7 content types are now handled via a
16510 catch all ASN1_TYPE structure. This allows unsupported
16511 types to be stored as a "blob" and an application can
16512 encode and decode it manually.
16516 * Fix various signed/unsigned issues to make a_strex.c
16517 compile under VC++.
16519 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16521 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16522 length if passed a buffer. ASN1_INTEGER_to_BN failed
16523 if passed a NULL BN and its argument was negative.
16525 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16527 * Modification to PKCS#7 encoding routines to output definite
16528 length encoding. Since currently the whole structures are in
16529 memory there's not real point in using indefinite length
16530 constructed encoding. However if OpenSSL is compiled with
16531 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16535 * Added BIO_vprintf() and BIO_vsnprintf().
16539 * Added more prefixes to parse for in the strings written
16540 through a logging bio, to cover all the levels that are available
16541 through syslog. The prefixes are now:
16543 PANIC, EMERG, EMR => LOG_EMERG
16544 ALERT, ALR => LOG_ALERT
16545 CRIT, CRI => LOG_CRIT
16546 ERROR, ERR => LOG_ERR
16547 WARNING, WARN, WAR => LOG_WARNING
16548 NOTICE, NOTE, NOT => LOG_NOTICE
16549 INFO, INF => LOG_INFO
16550 DEBUG, DBG => LOG_DEBUG
16552 and as before, if none of those prefixes are present at the
16553 beginning of the string, LOG_ERR is chosen.
16555 On Win32, the `LOG_*` levels are mapped according to this:
16557 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16558 LOG_WARNING => EVENTLOG_WARNING_TYPE
16559 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16563 * Made it possible to reconfigure with just the configuration
16564 argument "reconf" or "reconfigure". The command line arguments
16565 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16566 and are retrieved from there when reconfiguring.
16572 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16574 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16578 * The obj_dat.pl script was messing up the sorting of object
16579 names. The reason was that it compared the quoted version
16580 of strings as a result "OCSP" > "OCSP Signing" because
16581 " > SPACE. Changed script to store unquoted versions of
16582 names and add quotes on output. It was also omitting some
16583 names from the lookup table if they were given a default
16584 value (that is if SN is missing it is given the same
16585 value as LN and vice versa), these are now added on the
16586 grounds that if an object has a name we should be able to
16587 look it up. Finally added warning output when duplicate
16588 short or long names are found.
16592 * Changes needed for Tandem NSK.
16594 *Scott Uroff <scott@xypro.com>*
16596 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16597 RSA_padding_check_SSLv23(), special padding was never detected
16598 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16599 version rollback attacks was not effective.
16601 In s23_clnt.c, don't use special rollback-attack detection padding
16602 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16603 client; similarly, in s23_srvr.c, don't do the rollback check if
16604 SSL 2.0 is the only protocol enabled in the server.
16608 * Make it possible to get hexdumps of unprintable data with 'openssl
16609 asn1parse'. By implication, the functions ASN1_parse_dump() and
16610 BIO_dump_indent() are added.
16614 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16615 these print out strings and name structures based on various
16616 flags including RFC2253 support and proper handling of
16617 multibyte characters. Added options to the 'x509' utility
16618 to allow the various flags to be set.
16622 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16623 Also change the functions X509_cmp_current_time() and
16624 X509_gmtime_adj() work with an ASN1_TIME structure,
16625 this will enable certificates using GeneralizedTime in validity
16626 dates to be checked.
16630 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16631 negative public key encodings) on by default,
16632 NO_NEG_PUBKEY_BUG can be set to disable it.
16636 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16637 content octets. An i2c_ASN1_OBJECT is unnecessary because
16638 the encoding can be trivially obtained from the structure.
16642 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16643 not read locks (`CRYPTO_r_[un]lock`).
16647 * A first attempt at creating official support for shared
16648 libraries through configuration. I've kept it so the
16649 default is static libraries only, and the OpenSSL programs
16650 are always statically linked for now, but there are
16651 preparations for dynamic linking in place.
16652 This has been tested on Linux and Tru64.
16656 * Randomness polling function for Win9x, as described in:
16657 Peter Gutmann, Software Generation of Practically Strong
16662 * Fix so PRNG is seeded in req if using an already existing
16667 * New options to smime application. -inform and -outform
16668 allow alternative formats for the S/MIME message including
16669 PEM and DER. The -content option allows the content to be
16670 specified separately. This should allow things like Netscape
16671 form signing output easier to verify.
16675 * Fix the ASN1 encoding of tags using the 'long form'.
16679 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16680 STRING types. These convert content octets to and from the
16681 underlying type. The actual tag and length octets are
16682 already assumed to have been read in and checked. These
16683 are needed because all other string types have virtually
16684 identical handling apart from the tag. By having versions
16685 of the ASN1 functions that just operate on content octets
16686 IMPLICIT tagging can be handled properly. It also allows
16687 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16688 and ASN1_INTEGER are identical apart from the tag.
16692 * Change the handling of OID objects as follows:
16694 - New object identifiers are inserted in objects.txt, following
16695 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16696 - objects.pl is used to process obj_mac.num and create a new
16698 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16701 This is currently kind of a hack, and the perl code in objects.pl
16702 isn't very elegant, but it works as I intended. The simplest way
16703 to check that it worked correctly is to look in obj_dat.h and
16704 check the array nid_objs and make sure the objects haven't moved
16705 around (this is important!). Additions are OK, as well as
16706 consistent name changes.
16710 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16714 * Addition of the command line parameter '-rand file' to 'openssl req'.
16715 The given file adds to whatever has already been seeded into the
16716 random pool through the RANDFILE configuration file option or
16717 environment variable, or the default random state file.
16721 * mkstack.pl now sorts each macro group into lexical order.
16722 Previously the output order depended on the order the files
16723 appeared in the directory, resulting in needless rewriting
16728 * Patches to make OpenSSL compile under Win32 again. Mostly
16729 work arounds for the VC++ problem that it treats func() as
16730 func(void). Also stripped out the parts of mkdef.pl that
16731 added extra typesafe functions: these no longer exist.
16735 * Reorganisation of the stack code. The macros are now all
16736 collected in safestack.h . Each macro is defined in terms of
16737 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16738 DEBUG_SAFESTACK is now handled in terms of function casts,
16739 this has the advantage of retaining type safety without the
16740 use of additional functions. If DEBUG_SAFESTACK is not defined
16741 then the non typesafe macros are used instead. Also modified the
16742 mkstack.pl script to handle the new form. Needs testing to see
16743 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16744 the default if no major problems. Similar behaviour for ASN1_SET_OF
16745 and PKCS12_STACK_OF.
16749 * When some versions of IIS use the 'NET' form of private key the
16750 key derivation algorithm is different. Normally MD5(password) is
16751 used as a 128 bit RC4 key. In the modified case
16752 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16753 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16754 as the old Netscape_RSA functions except they have an additional
16755 'sgckey' parameter which uses the modified algorithm. Also added
16756 an -sgckey command line option to the rsa utility. Thanks to
16757 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16758 algorithm to openssl-dev.
16762 * The evp_local.h macros were using 'c.##kname' which resulted in
16763 invalid expansion on some systems (SCO 5.0.5 for example).
16764 Corrected to 'c.kname'.
16766 *Phillip Porch <root@theporch.com>*
16768 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16769 a STACK of email addresses from a certificate or request, these look
16770 in the subject name and the subject alternative name extensions and
16771 omit any duplicate addresses.
16775 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16776 This makes DSA verification about 2 % faster.
16780 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16781 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16782 plus overhead for 1024 bit moduli).
16783 This makes exponentiations about 0.5 % faster for 1024 bit
16784 exponents (as measured by "openssl speed rsa2048").
16788 * Rename memory handling macros to avoid conflicts with other
16790 Malloc => OPENSSL_malloc
16791 Malloc_locked => OPENSSL_malloc_locked
16792 Realloc => OPENSSL_realloc
16793 Free => OPENSSL_free
16797 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16798 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16802 * CygWin32 support.
16804 *John Jarvie <jjarvie@newsguy.com>*
16806 * The type-safe stack code has been rejigged. It is now only compiled
16807 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16808 by default all type-specific stack functions are "#define"d back to
16809 standard stack functions. This results in more streamlined output
16810 but retains the type-safety checking possibilities of the original
16815 * The STACK code has been cleaned up, and certain type declarations
16816 that didn't make a lot of sense have been brought in line. This has
16817 also involved a cleanup of sorts in safestack.h to more correctly
16818 map type-safe stack functions onto their plain stack counterparts.
16819 This work has also resulted in a variety of "const"ifications of
16820 lots of the code, especially `_cmp` operations which should normally
16821 be prototyped with "const" parameters anyway.
16825 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16826 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16827 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16828 where all of 'md' is used each time the PRNG is used, but 'state'
16829 is used only indexed by a cyclic counter. As entropy may not be
16830 well distributed from the beginning, 'md' is important as a
16831 chaining variable. However, the output function chains only half
16832 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16833 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16834 in all of 'state' being rewritten, with the new values depending
16835 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16839 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16840 the handshake is continued after ssl_verify_cert_chain();
16841 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16842 can lead to 'unexplainable' connection aborts later.
16844 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16846 * Major EVP API cipher revision.
16847 Add hooks for extra EVP features. This allows various cipher
16848 parameters to be set in the EVP interface. Support added for variable
16849 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16850 setting of RC2 and RC5 parameters.
16852 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16855 Remove lots of duplicated code from the EVP library. For example *every*
16856 cipher init() function handles the 'iv' in the same way according to the
16857 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16858 for CFB and OFB modes they zero ctx->num.
16860 New functionality allows removal of S/MIME code RC2 hack.
16862 Most of the routines have the same form and so can be declared in terms
16865 By shifting this to the top level EVP_CipherInit() it can be removed from
16866 all individual ciphers. If the cipher wants to handle IVs or keys
16867 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16870 Change lots of functions like EVP_EncryptUpdate() to now return a
16871 value: although software versions of the algorithms cannot fail
16872 any installed hardware versions can.
16876 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16877 this option is set, tolerate broken clients that send the negotiated
16878 protocol version number instead of the requested protocol version
16883 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16884 i.e. non-zero for export ciphersuites, zero otherwise.
16885 Previous versions had this flag inverted, inconsistent with
16886 rsa_tmp_cb (..._TMP_RSA_CB).
16888 *Bodo Moeller; problem reported by Amit Chopra*
16890 * Add missing DSA library text string. Work around for some IIS
16891 key files with invalid SEQUENCE encoding.
16895 * Add a document (doc/standards.txt) that list all kinds of standards
16896 and so on that are implemented in OpenSSL.
16900 * Enhance c_rehash script. Old version would mishandle certificates
16901 with the same subject name hash and wouldn't handle CRLs at all.
16902 Added -fingerprint option to crl utility, to support new c_rehash
16907 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16911 * Fix for SSL server purpose checking. Server checking was
16912 rejecting certificates which had extended key usage present
16913 but no ssl client purpose.
16915 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16917 * Make PKCS#12 code work with no password. The PKCS#12 spec
16918 is a little unclear about how a blank password is handled.
16919 Since the password in encoded as a BMPString with terminating
16920 double NULL a zero length password would end up as just the
16921 double NULL. However no password at all is different and is
16922 handled differently in the PKCS#12 key generation code. NS
16923 treats a blank password as zero length. MSIE treats it as no
16924 password on export: but it will try both on import. We now do
16925 the same: PKCS12_parse() tries zero length and no password if
16926 the password is set to "" or NULL (NULL is now a valid password:
16927 it wasn't before) as does the pkcs12 application.
16931 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16932 perror when PEM_read_bio_X509_REQ fails, the error message must
16933 be obtained from the error queue.
16937 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16938 it in ERR_remove_state if appropriate, and change ERR_get_state
16939 accordingly to avoid race conditions (this is necessary because
16940 thread_hash is no longer constant once set).
16944 * Bugfix for linux-elf makefile.one.
16948 * RSA_get_default_method() will now cause a default
16949 RSA_METHOD to be chosen if one doesn't exist already.
16950 Previously this was only set during a call to RSA_new()
16951 or RSA_new_method(NULL) meaning it was possible for
16952 RSA_get_default_method() to return NULL.
16956 * Added native name translation to the existing DSO code
16957 that will convert (if the flag to do so is set) filenames
16958 that are sufficiently small and have no path information
16959 into a canonical native form. Eg. "blah" converted to
16960 "libblah.so" or "blah.dll" etc.
16964 * New function ERR_error_string_n(e, buf, len) which is like
16965 ERR_error_string(e, buf), but writes at most 'len' bytes
16966 including the 0 terminator. For ERR_error_string_n, 'buf'
16969 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16971 * CONF library reworked to become more general. A new CONF
16972 configuration file reader "class" is implemented as well as a
16973 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16974 old `CONF_*` functions are still there, but are reimplemented to
16975 work in terms of the new functions. Also, a set of functions
16976 to handle the internal storage of the configuration data is
16977 provided to make it easier to write new configuration file
16978 reader "classes" (I can definitely see something reading a
16979 configuration file in XML format, for example), called `_CONF_*`,
16980 or "the configuration storage API"...
16982 The new configuration file reading functions are:
16984 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16985 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16987 NCONF_default, NCONF_WIN32
16989 NCONF_dump_fp, NCONF_dump_bio
16991 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16992 NCONF_new creates a new CONF object. This works in the same way
16993 as other interfaces in OpenSSL, like the BIO interface.
16994 `NCONF_dump_*` dump the internal storage of the configuration file,
16995 which is useful for debugging. All other functions take the same
16996 arguments as the old `CONF_*` functions with the exception of the
16997 first that must be a `CONF *` instead of a `LHASH *`.
16999 To make it easier to use the new classes with the old `CONF_*` functions,
17000 the function CONF_set_default_method is provided.
17004 * Add '-tls1' option to 'openssl ciphers', which was already
17005 mentioned in the documentation but had not been implemented.
17006 (This option is not yet really useful because even the additional
17007 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17011 * Initial DSO code added into libcrypto for letting OpenSSL (and
17012 OpenSSL-based applications) load shared libraries and bind to
17013 them in a portable way.
17015 *Geoff Thorpe, with contributions from Richard Levitte*
17017 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17019 * Make sure _lrotl and _lrotr are only used with MSVC.
17021 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17022 (the default implementation of RAND_status).
17024 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17025 to '-clrext' (= clear extensions), as intended and documented.
17026 *Bodo Moeller; inconsistency pointed out by Michael Attili
17027 <attili@amaxo.com>*
17029 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17030 was larger than the MD block size.
17032 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17034 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17035 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17036 using the passed key: if the passed key was a private key the result
17037 of X509_print(), for example, would be to print out all the private key
17042 * des_quad_cksum() byte order bug fix.
17043 *Ulf Möller, using the problem description in krb4-0.9.7, where
17044 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17046 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17049 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17051 * For easily testing in shell scripts whether some command
17052 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17053 returns with exit code 0 iff no command of the given name is available.
17054 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17055 the output goes to stdout and nothing is printed to stderr.
17056 Additional arguments are always ignored.
17058 Since for each cipher there is a command of the same name,
17059 the 'no-cipher' compilation switches can be tested this way.
17061 ('openssl no-XXX' is not able to detect pseudo-commands such
17062 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17066 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17070 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17071 is set; it will be thrown away anyway because each handshake creates
17073 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17074 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17075 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17076 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17080 * New s_client option -ign_eof: EOF at stdin is ignored, and
17081 'Q' and 'R' lose their special meanings (quit/renegotiate).
17082 This is part of what -quiet does; unlike -quiet, -ign_eof
17083 does not suppress any output.
17087 * Add compatibility options to the purpose and trust code. The
17088 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17089 accepts a certificate or CA, this was the previous behaviour,
17090 with all the associated security issues.
17092 X509_TRUST_COMPAT is the old trust behaviour: only and
17093 automatically trust self signed roots in certificate store. A
17094 new trust setting X509_TRUST_DEFAULT is used to specify that
17095 a purpose has no associated trust setting and it should instead
17096 use the value in the default purpose.
17100 * Fix the PKCS#8 DSA private key code so it decodes keys again
17101 and fix a memory leak.
17105 * In util/mkerr.pl (which implements 'make errors'), preserve
17106 reason strings from the previous version of the .c file, as
17107 the default to have only downcase letters (and digits) in
17108 automatically generated reasons codes is not always appropriate.
17112 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17113 using strerror. Previously, ERR_reason_error_string() returned
17114 library names as reason strings for SYSerr; but SYSerr is a special
17115 case where small numbers are errno values, not library numbers.
17119 * Add '-dsaparam' option to 'openssl dhparam' application. This
17120 converts DSA parameters into DH parameters. (When creating parameters,
17121 DSA_generate_parameters is used.)
17125 * Include 'length' (recommended exponent length) in C code generated
17126 by 'openssl dhparam -C'.
17130 * The second argument to set_label in perlasm was already being used
17131 so couldn't be used as a "file scope" flag. Moved to third argument
17136 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17137 instead of RAND_bytes for encryption IVs and salts.
17141 * Include RAND_status() into RAND_METHOD instead of implementing
17142 it only for md_rand.c Otherwise replacing the PRNG by calling
17143 RAND_set_rand_method would be impossible.
17147 * Don't let DSA_generate_key() enter an infinite loop if the random
17148 number generation fails.
17152 * New 'rand' application for creating pseudo-random output.
17156 * Added configuration support for Linux/IA64
17158 *Rolf Haberrecker <rolf@suse.de>*
17160 * Assembler module support for Mingw32.
17164 * Shared library support for HPUX (in shlib/).
17166 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17168 * Shared library support for Solaris gcc.
17170 *Lutz Behnke <behnke@trustcenter.de>*
17172 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17174 * PKCS7_encrypt() was adding text MIME headers twice because they
17175 were added manually and by SMIME_crlf_copy().
17179 * In bntest.c don't call BN_rand with zero bits argument.
17181 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17183 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17184 case was implemented. This caused BN_div_recp() to fail occasionally.
17188 * Add an optional second argument to the set_label() in the perl
17189 assembly language builder. If this argument exists and is set
17190 to 1 it signals that the assembler should use a symbol whose
17191 scope is the entire file, not just the current function. This
17192 is needed with MASM which uses the format label:: for this scope.
17194 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17196 * Change the ASN1 types so they are typedefs by default. Before
17197 almost all types were #define'd to ASN1_STRING which was causing
17198 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17203 * Change names of new functions to the new get1/get0 naming
17204 convention: After 'get1', the caller owns a reference count
17205 and has to call `..._free`; 'get0' returns a pointer to some
17206 data structure without incrementing reference counters.
17207 (Some of the existing 'get' functions increment a reference
17208 counter, some don't.)
17209 Similarly, 'set1' and 'add1' functions increase reference
17210 counters or duplicate objects.
17214 * Allow for the possibility of temp RSA key generation failure:
17215 the code used to assume it always worked and crashed on failure.
17219 * Fix potential buffer overrun problem in BIO_printf().
17220 *Ulf Möller, using public domain code by Patrick Powell; problem
17221 pointed out by David Sacerdote <das33@cornell.edu>*
17223 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17224 RAND_egd() and RAND_status(). In the command line application,
17225 the EGD socket can be specified like a seed file using RANDFILE
17230 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17231 Some CAs (e.g. Verisign) distribute certificates in this form.
17235 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17236 list to exclude them. This means that no special compilation option
17237 is needed to use anonymous DH: it just needs to be included in the
17242 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17243 EVP_MD_type. The old functionality is available in a new macro called
17244 EVP_MD_md(). Change code that uses it and update docs.
17248 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17249 where the `void *` argument is replaced by a function pointer argument.
17250 Previously `void *` was abused to point to functions, which works on
17251 many platforms, but is not correct. As these functions are usually
17252 called by macros defined in OpenSSL header files, most source code
17253 should work without changes.
17257 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17258 sections with information on -D... compiler switches used for
17259 compiling the library so that applications can see them. To enable
17260 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17261 must be defined. E.g.,
17262 #define OPENSSL_ALGORITHM_DEFINES
17263 #include <openssl/opensslconf.h>
17264 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17266 *Richard Levitte, Ulf and Bodo Möller*
17268 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17273 * Change the 'other' type in certificate aux info to a STACK_OF
17274 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17275 the required ASN1 format: arbitrary types determined by an OID.
17279 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17280 argument to 'req'. This is not because the function is newer or
17281 better than others it just uses the work 'NEW' in the certificate
17282 request header lines. Some software needs this.
17286 * Reorganise password command line arguments: now passwords can be
17287 obtained from various sources. Delete the PEM_cb function and make
17288 it the default behaviour: i.e. if the callback is NULL and the
17289 usrdata argument is not NULL interpret it as a null terminated pass
17290 phrase. If usrdata and the callback are NULL then the pass phrase
17291 is prompted for as usual.
17295 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17296 the support is automatically enabled. The resulting binaries will
17297 autodetect the card and use it if present.
17299 *Ben Laurie and Compaq Inc.*
17301 * Work around for Netscape hang bug. This sends certificate request
17302 and server done in one record. Since this is perfectly legal in the
17303 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17304 the bugs/SSLv3 entry for more info.
17308 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17312 * Add -rand argument to smime and pkcs12 applications and read/write
17317 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17321 * Add command line password options to the remaining applications.
17325 * Bug fix for BN_div_recp() for numerators with an even number of
17330 * More tests in bntest.c, and changed test_bn output.
17334 * ./config recognizes MacOS X now.
17338 * Bug fix for BN_div() when the first words of num and divisor are
17339 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17343 * Add support for various broken PKCS#8 formats, and command line
17344 options to produce them.
17348 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17349 get temporary BIGNUMs from a BN_CTX.
17353 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17358 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17359 include a #define from the old name to the new. The original intent
17360 was that statically linked binaries could for example just call
17361 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17362 link with digests. This never worked because SSLeay_add_all_digests()
17363 and SSLeay_add_all_ciphers() were in the same source file so calling
17364 one would link with the other. They are now in separate source files.
17368 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17372 * Use a less unusual form of the Miller-Rabin primality test (it used
17373 a binary algorithm for exponentiation integrated into the Miller-Rabin
17374 loop, our standard modexp algorithms are faster).
17378 * Support for the EBCDIC character set completed.
17380 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17382 * Source code cleanups: use const where appropriate, eliminate casts,
17383 use `void *` instead of `char *` in lhash.
17387 * Bugfix: ssl3_send_server_key_exchange was not restartable
17388 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17389 this the server could overwrite ephemeral keys that the client
17394 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17395 using 50 iterations of the Rabin-Miller test.
17397 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17398 iterations of the Rabin-Miller test as required by the appendix
17399 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17400 As BN_is_prime_fasttest includes trial division, DSA parameter
17401 generation becomes much faster.
17403 This implies a change for the callback functions in DSA_is_prime
17404 and DSA_generate_parameters: The callback function is called once
17405 for each positive witness in the Rabin-Miller test, not just
17406 occasionally in the inner loop; and the parameters to the
17407 callback function now provide an iteration count for the outer
17408 loop rather than for the current invocation of the inner loop.
17409 DSA_generate_parameters additionally can call the callback
17410 function with an 'iteration count' of -1, meaning that a
17411 candidate has passed the trial division test (when q is generated
17412 from an application-provided seed, trial division is skipped).
17416 * New function BN_is_prime_fasttest that optionally does trial
17417 division before starting the Rabin-Miller test and has
17418 an additional BN_CTX * argument (whereas BN_is_prime always
17419 has to allocate at least one BN_CTX).
17420 'callback(1, -1, cb_arg)' is called when a number has passed the
17421 trial division stage.
17425 * Fix for bug in CRL encoding. The validity dates weren't being handled
17430 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17434 * New function BN_pseudo_rand().
17438 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17439 bignum version of BN_from_montgomery() with the working code from
17440 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17445 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17446 made it impossible to use the same SSL_SESSION data structure in
17447 SSL2 clients in multiple threads.
17451 * The return value of RAND_load_file() no longer counts bytes obtained
17452 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17453 to seed the PRNG (previously an explicit byte count was required).
17455 *Ulf Möller, Bodo Möller*
17457 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17458 used `char *` instead of `void *` and had casts all over the place.
17462 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17466 * Retain source code compatibility for BN_prime_checks macro:
17467 BN_is_prime(..., BN_prime_checks, ...) now uses
17468 BN_prime_checks_for_size to determine the appropriate number of
17469 Rabin-Miller iterations.
17473 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17474 DH_CHECK_P_NOT_SAFE_PRIME.
17475 (Check if this is true? OpenPGP calls them "strong".)
17479 * Merge the functionality of "dh" and "gendh" programs into a new program
17480 "dhparam". The old programs are retained for now but will handle DH keys
17481 (instead of parameters) in future.
17485 * Make the ciphers, s_server and s_client programs check the return values
17486 when a new cipher list is set.
17490 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17491 ciphers. Before when the 56bit ciphers were enabled the sorting was
17494 The syntax for the cipher sorting has been extended to support sorting by
17495 cipher-strength (using the strength_bits hard coded in the tables).
17496 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17498 Fix a bug in the cipher-command parser: when supplying a cipher command
17499 string with an "undefined" symbol (neither command nor alphanumeric
17500 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17501 an error is flagged.
17503 Due to the strength-sorting extension, the code of the
17504 ssl_create_cipher_list() function was completely rearranged. I hope that
17505 the readability was also increased :-)
17507 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17509 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17510 for the first serial number and places 2 in the serial number file. This
17511 avoids problems when the root CA is created with serial number zero and
17512 the first user certificate has the same issuer name and serial number
17517 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17518 the new code. Add documentation for this stuff.
17522 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17523 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17524 structures and behave in an analogous way to the X509v3 functions:
17525 they shouldn't be called directly but wrapper functions should be used
17528 So we also now have some wrapper functions that call the X509at functions
17529 when passed certificate requests. (TO DO: similar things can be done with
17530 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17531 things. Some of these need some d2i or i2d and print functionality
17532 because they handle more complex structures.)
17536 * Add missing #ifndefs that caused missing symbols when building libssl
17537 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17538 NO_RSA in `ssl/s2*.c`.
17540 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17542 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17543 has a return value which indicates the quality of the random data
17544 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17545 error queue. New function RAND_pseudo_bytes() generates output that is
17546 guaranteed to be unique but not unpredictable. RAND_add is like
17547 RAND_seed, but takes an extra argument for an entropy estimate
17548 (RAND_seed always assumes full entropy).
17552 * Do more iterations of Rabin-Miller probable prime test (specifically,
17553 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17554 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17555 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17556 false-positive rate of at most 2^-80 for random input.
17560 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17564 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17565 in the 0.9.5 release), this returns the chain
17566 from an X509_CTX structure with a dup of the stack and all
17567 the X509 reference counts upped: so the stack will exist
17568 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17571 Also make SSL_SESSION_print() print out the verify return
17576 * Add manpage for the pkcs12 command. Also change the default
17577 behaviour so MAC iteration counts are used unless the new
17578 -nomaciter option is used. This improves file security and
17579 only older versions of MSIE (4.0 for example) need it.
17583 * Honor the no-xxx Configure options when creating .DEF files.
17587 * Add PKCS#10 attributes to field table: challengePassword,
17588 unstructuredName and unstructuredAddress. These are taken from
17589 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17590 international characters are used.
17592 More changes to X509_ATTRIBUTE code: allow the setting of types
17593 based on strings. Remove the 'loc' parameter when adding
17594 attributes because these will be a SET OF encoding which is sorted
17599 * Initial changes to the 'req' utility to allow request generation
17600 automation. This will allow an application to just generate a template
17601 file containing all the field values and have req construct the
17604 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17605 used all over the place including certificate requests and PKCS#7
17606 structures. They are currently handled manually where necessary with
17607 some primitive wrappers for PKCS#7. The new functions behave in a
17608 manner analogous to the X509 extension functions: they allow
17609 attributes to be looked up by NID and added.
17611 Later something similar to the X509V3 code would be desirable to
17612 automatically handle the encoding, decoding and printing of the
17613 more complex types. The string types like challengePassword can
17614 be handled by the string table functions.
17616 Also modified the multi byte string table handling. Now there is
17617 a 'global mask' which masks out certain types. The table itself
17618 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17619 is useful when for example there is only one permissible type
17620 (as in countryName) and using the mask might result in no valid
17625 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17626 SSL_get_peer_finished to allow applications to obtain the latest
17627 Finished messages sent to the peer or expected from the peer,
17628 respectively. (SSL_get_peer_finished is usually the Finished message
17629 actually received from the peer, otherwise the protocol will be aborted.)
17631 As the Finished message are message digests of the complete handshake
17632 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17633 be used for external authentication procedures when the authentication
17634 provided by SSL/TLS is not desired or is not enough.
17638 * Enhanced support for Alpha Linux is added. Now ./config checks if
17639 the host supports BWX extension and if Compaq C is present on the
17640 $PATH. Just exploiting of the BWX extension results in 20-30%
17641 performance kick for some algorithms, e.g. DES and RC4 to mention
17642 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17647 * Add support for MS "fast SGC". This is arguably a violation of the
17648 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17649 weak crypto and after checking the certificate is SGC a second one
17650 with strong crypto. MS SGC stops the first handshake after receiving
17651 the server certificate message and sends a second client hello. Since
17652 a server will typically do all the time consuming operations before
17653 expecting any further messages from the client (server key exchange
17654 is the most expensive) there is little difference between the two.
17656 To get OpenSSL to support MS SGC we have to permit a second client
17657 hello message after we have sent server done. In addition we have to
17658 reset the MAC if we do get this second client hello.
17662 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17663 if a DER encoded private key is RSA or DSA traditional format. Changed
17664 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17665 format DER encoded private key. Newer code should use PKCS#8 format which
17666 has the key type encoded in the ASN1 structure. Added DER private key
17667 support to pkcs8 application.
17671 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17672 ciphersuites has been selected (as required by the SSL 3/TLS 1
17673 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17674 is set, we interpret this as a request to violate the specification
17675 (the worst that can happen is a handshake failure, and 'correct'
17676 behaviour would result in a handshake failure anyway).
17680 * In SSL_CTX_add_session, take into account that there might be multiple
17681 SSL_SESSION structures with the same session ID (e.g. when two threads
17682 concurrently obtain them from an external cache).
17683 The internal cache can handle only one SSL_SESSION with a given ID,
17684 so if there's a conflict, we now throw out the old one to achieve
17689 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17690 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17691 some routines that use cipher OIDs: some ciphers do not have OIDs
17692 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17697 * Simplify the trust setting structure and code. Now we just have
17698 two sequences of OIDs for trusted and rejected settings. These will
17699 typically have values the same as the extended key usage extension
17700 and any application specific purposes.
17702 The trust checking code now has a default behaviour: it will just
17703 check for an object with the same NID as the passed id. Functions can
17704 be provided to override either the default behaviour or the behaviour
17705 for a given id. SSL client, server and email already have functions
17706 in place for compatibility: they check the NID and also return "trusted"
17707 if the certificate is self signed.
17711 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17712 traditional format into an EVP_PKEY structure.
17716 * Add a password callback function PEM_cb() which either prompts for
17717 a password if usr_data is NULL or otherwise assumes it is a null
17718 terminated password. Allow passwords to be passed on command line
17719 environment or config files in a few more utilities.
17723 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17724 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17725 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17726 Update documentation.
17730 * Support for ASN1 "NULL" type. This could be handled before by using
17731 ASN1_TYPE but there wasn't any function that would try to read a NULL
17732 and produce an error if it couldn't. For compatibility we also have
17733 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17734 don't allocate anything because they don't need to.
17738 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17741 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17743 * Rebuild of the memory allocation routines used by OpenSSL code and
17744 possibly others as well. The purpose is to make an interface that
17745 provide hooks so anyone can build a separate set of allocation and
17746 deallocation routines to be used by OpenSSL, for example memory
17747 pool implementations, or something else, which was previously hard
17748 since Malloc(), Realloc() and Free() were defined as macros having
17749 the values malloc, realloc and free, respectively (except for Win32
17750 compilations). The same is provided for memory debugging code.
17751 OpenSSL already comes with functionality to find memory leaks, but
17752 this gives people a chance to debug other memory problems.
17754 With these changes, a new set of functions and macros have appeared:
17756 CRYPTO_set_mem_debug_functions() [F]
17757 CRYPTO_get_mem_debug_functions() [F]
17758 CRYPTO_dbg_set_options() [F]
17759 CRYPTO_dbg_get_options() [F]
17760 CRYPTO_malloc_debug_init() [M]
17762 The memory debug functions are NULL by default, unless the library
17763 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17764 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17765 gives the standard debugging functions that come with OpenSSL) or
17766 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17767 provided by the library user) must be used. When the standard
17768 debugging functions are used, CRYPTO_dbg_set_options can be used to
17769 request additional information:
17770 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17771 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17773 Also, things like CRYPTO_set_mem_functions will always give the
17774 expected result (the new set of functions is used for allocation
17775 and deallocation) at all times, regardless of platform and compiler
17778 To finish it up, some functions that were never use in any other
17779 way than through macros have a new API and new semantic:
17781 CRYPTO_dbg_malloc()
17782 CRYPTO_dbg_realloc()
17785 All macros of value have retained their old syntax.
17787 *Richard Levitte and Bodo Moeller*
17789 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17790 ordering of SMIMECapabilities wasn't in "strength order" and there
17791 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17796 * Some ASN1 types with illegal zero length encoding (INTEGER,
17797 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17799 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17801 * Merge in my S/MIME library for OpenSSL. This provides a simple
17802 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17803 functionality to handle multipart/signed properly) and a utility
17804 called 'smime' to call all this stuff. This is based on code I
17805 originally wrote for Celo who have kindly allowed it to be
17806 included in OpenSSL.
17810 * Add variants des_set_key_checked and des_set_key_unchecked of
17811 des_set_key (aka des_key_sched). Global variable des_check_key
17812 decides which of these is called by des_set_key; this way
17813 des_check_key behaves as it always did, but applications and
17814 the library itself, which was buggy for des_check_key == 1,
17815 have a cleaner way to pick the version they need.
17819 * New function PKCS12_newpass() which changes the password of a
17824 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17825 dynamic mix. In both cases the ids can be used as an index into the
17826 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17827 functions so they accept a list of the field values and the
17828 application doesn't need to directly manipulate the X509_TRUST
17833 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17838 * Modify the way the V3 extension code looks up extensions. This now
17839 works in a similar way to the object code: we have some "standard"
17840 extensions in a static table which is searched with OBJ_bsearch()
17841 and the application can add dynamic ones if needed. The file
17842 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17843 updated whenever a new extension is added to the core code and kept
17844 in ext_nid order. There is a simple program 'tabtest.c' which checks
17845 this. New extensions are not added too often so this file can readily
17846 be maintained manually.
17848 There are two big advantages in doing things this way. The extensions
17849 can be looked up immediately and no longer need to be "added" using
17850 X509V3_add_standard_extensions(): this function now does nothing.
17851 Side note: I get *lots* of email saying the extension code doesn't
17852 work because people forget to call this function.
17853 Also no dynamic allocation is done unless new extensions are added:
17854 so if we don't add custom extensions there is no need to call
17855 X509V3_EXT_cleanup().
17859 * Modify enc utility's salting as follows: make salting the default. Add a
17860 magic header, so unsalted files fail gracefully instead of just decrypting
17861 to garbage. This is because not salting is a big security hole, so people
17862 should be discouraged from doing it.
17866 * Fixes and enhancements to the 'x509' utility. It allowed a message
17867 digest to be passed on the command line but it only used this
17868 parameter when signing a certificate. Modified so all relevant
17869 operations are affected by the digest parameter including the
17870 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17871 DSA key was used because it didn't fix the digest.
17875 * Initial certificate chain verify code. Currently tests the untrusted
17876 certificates for consistency with the verify purpose (which is set
17877 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17879 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17880 this is because it will reject chains with invalid extensions whereas
17881 every previous version of OpenSSL and SSLeay made no checks at all.
17883 Trust code: checks the root CA for the relevant trust settings. Trust
17884 settings have an initial value consistent with the verify purpose: e.g.
17885 if the verify purpose is for SSL client use it expects the CA to be
17886 trusted for SSL client use. However the default value can be changed to
17887 permit custom trust settings: one example of this would be to only trust
17888 certificates from a specific "secure" set of CAs.
17890 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17891 which should be used for version portability: especially since the
17892 verify structure is likely to change more often now.
17894 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17895 to set them. If not set then assume SSL clients will verify SSL servers
17898 Two new options to the verify program: -untrusted allows a set of
17899 untrusted certificates to be passed in and -purpose which sets the
17900 intended purpose of the certificate. If a purpose is set then the
17901 new chain verify code is used to check extension consistency.
17905 * Support for the authority information access extension.
17909 * Modify RSA and DSA PEM read routines to transparently handle
17910 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17911 public keys in a format compatible with certificate
17912 SubjectPublicKeyInfo structures. Unfortunately there were already
17913 functions called *_PublicKey_* which used various odd formats so
17914 these are retained for compatibility: however the DSA variants were
17915 never in a public release so they have been deleted. Changed dsa/rsa
17916 utilities to handle the new format: note no releases ever handled public
17917 keys so we should be OK.
17919 The primary motivation for this change is to avoid the same fiasco
17920 that dogs private keys: there are several incompatible private key
17921 formats some of which are standard and some OpenSSL specific and
17922 require various evil hacks to allow partial transparent handling and
17923 even then it doesn't work with DER formats. Given the option anything
17924 other than PKCS#8 should be dumped: but the other formats have to
17925 stay in the name of compatibility.
17927 With public keys and the benefit of hindsight one standard format
17928 is used which works with EVP_PKEY, RSA or DSA structures: though
17929 it clearly returns an error if you try to read the wrong kind of key.
17931 Added a -pubkey option to the 'x509' utility to output the public key.
17932 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17933 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17934 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17935 that do the same as the `EVP_PKEY_assign_*()` except they up the
17936 reference count of the added key (they don't "swallow" the
17941 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17942 CRLs would fail if the file contained no certificates or no CRLs:
17943 added a new function to read in both types and return the number
17944 read: this means that if none are read it will be an error. The
17945 DER versions of the certificate and CRL reader would always fail
17946 because it isn't possible to mix certificates and CRLs in DER format
17947 without choking one or the other routine. Changed this to just read
17948 a certificate: this is the best we can do. Also modified the code
17949 in `apps/verify.c` to take notice of return codes: it was previously
17950 attempting to read in certificates from NULL pointers and ignoring
17951 any errors: this is one reason why the cert and CRL reader seemed
17952 to work. It doesn't check return codes from the default certificate
17953 routines: these may well fail if the certificates aren't installed.
17957 * Code to support otherName option in GeneralName.
17961 * First update to verify code. Change the verify utility
17962 so it warns if it is passed a self signed certificate:
17963 for consistency with the normal behaviour. X509_verify
17964 has been modified to it will now verify a self signed
17965 certificate if *exactly* the same certificate appears
17966 in the store: it was previously impossible to trust a
17967 single self signed certificate. This means that:
17968 openssl verify ss.pem
17969 now gives a warning about a self signed certificate but
17970 openssl verify -CAfile ss.pem ss.pem
17975 * For servers, store verify_result in SSL_SESSION data structure
17976 (and add it to external session representation).
17977 This is needed when client certificate verifications fails,
17978 but an application-provided verification callback (set by
17979 SSL_CTX_set_cert_verify_callback) allows accepting the session
17980 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17981 but returns 1): When the session is reused, we have to set
17982 ssl->verify_result to the appropriate error code to avoid
17985 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17987 * Fix a bug in the new PKCS#7 code: it didn't consider the
17988 case in PKCS7_dataInit() where the signed PKCS7 structure
17989 didn't contain any existing data because it was being created.
17991 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17993 * Add a salt to the key derivation routines in enc.c. This
17994 forms the first 8 bytes of the encrypted file. Also add a
17995 -S option to allow a salt to be input on the command line.
17999 * New function X509_cmp(). Oddly enough there wasn't a function
18000 to compare two certificates. We do this by working out the SHA1
18001 hash and comparing that. X509_cmp() will be needed by the trust
18006 * SSL_get1_session() is like SSL_get_session(), but increments
18007 the reference count in the SSL_SESSION returned.
18009 *Geoff Thorpe <geoff@eu.c2.net>*
18011 * Fix for 'req': it was adding a null to request attributes.
18012 Also change the X509_LOOKUP and X509_INFO code to handle
18013 certificate auxiliary information.
18017 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18022 * Add the possibility to add extra information to the memory leak
18023 detecting output, to form tracebacks, showing from where each
18024 allocation was originated: CRYPTO_push_info("constant string") adds
18025 the string plus current file name and line number to a per-thread
18026 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18027 is like calling CYRPTO_pop_info() until the stack is empty.
18028 Also updated memory leak detection code to be multi-thread-safe.
18032 * Add options -text and -noout to pkcs7 utility and delete the
18033 encryption options which never did anything. Update docs.
18037 * Add options to some of the utilities to allow the pass phrase
18038 to be included on either the command line (not recommended on
18039 OSes like Unix) or read from the environment. Update the
18040 manpages and fix a few bugs.
18044 * Add a few manpages for some of the openssl commands.
18048 * Fix the -revoke option in ca. It was freeing up memory twice,
18049 leaking and not finding already revoked certificates.
18053 * Extensive changes to support certificate auxiliary information.
18054 This involves the use of X509_CERT_AUX structure and X509_AUX
18055 functions. An X509_AUX function such as PEM_read_X509_AUX()
18056 can still read in a certificate file in the usual way but it
18057 will also read in any additional "auxiliary information". By
18058 doing things this way a fair degree of compatibility can be
18059 retained: existing certificates can have this information added
18060 using the new 'x509' options.
18062 Current auxiliary information includes an "alias" and some trust
18063 settings. The trust settings will ultimately be used in enhanced
18064 certificate chain verification routines: currently a certificate
18065 can only be trusted if it is self signed and then it is trusted
18070 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18071 The problem was that one of the replacement routines had not been working
18072 since SSLeay releases. For now the offending routine has been replaced
18073 with non-optimised assembler. Even so, this now gives around 95%
18074 performance improvement for 1024 bit RSA signs.
18078 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18079 handling. Most clients have the effective key size in bits equal to
18080 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18081 A few however don't do this and instead use the size of the decrypted key
18082 to determine the RC2 key length and the AlgorithmIdentifier to determine
18083 the effective key length. In this case the effective key length can still
18084 be 40 bits but the key length can be 168 bits for example. This is fixed
18085 by manually forcing an RC2 key into the EVP_PKEY structure because the
18086 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18087 the key length and effective key length are equal.
18091 * Add a bunch of functions that should simplify the creation of
18092 X509_NAME structures. Now you should be able to do:
18093 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18094 and have it automatically work out the correct field type and fill in
18095 the structures. The more adventurous can try:
18096 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18097 and it will (hopefully) work out the correct multibyte encoding.
18101 * Change the 'req' utility to use the new field handling and multibyte
18102 copy routines. Before the DN field creation was handled in an ad hoc
18103 way in req, ca, and x509 which was rather broken and didn't support
18104 BMPStrings or UTF8Strings. Since some software doesn't implement
18105 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18106 using the dirstring_type option. See the new comment in the default
18107 openssl.cnf for more info.
18111 * Make crypto/rand/md_rand.c more robust:
18112 - Assure unique random numbers after fork().
18113 - Make sure that concurrent threads access the global counter and
18114 md serializably so that we never lose entropy in them
18115 or use exactly the same state in multiple threads.
18116 Access to the large state is not always serializable because
18117 the additional locking could be a performance killer, and
18118 md should be large enough anyway.
18122 * New file `apps/app_rand.c` with commonly needed functionality
18123 for handling the random seed file.
18125 Use the random seed file in some applications that previously did not:
18127 dsaparam -genkey (which also ignored its '-rand' option),
18130 x509 (when signing).
18131 Except on systems with /dev/urandom, it is crucial to have a random
18132 seed file at least for key creation, DSA signing, and for DH exchanges;
18133 for RSA signatures we could do without one.
18135 gendh and gendsa (unlike genrsa) used to read only the first byte
18136 of each file listed in the '-rand' option. The function as previously
18137 found in genrsa is now in app_rand.c and is used by all programs
18138 that support '-rand'.
18142 * In RAND_write_file, use mode 0600 for creating files;
18143 don't just chmod when it may be too late.
18147 * Report an error from X509_STORE_load_locations
18148 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18152 * New function ASN1_mbstring_copy() this copies a string in either
18153 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18154 into an ASN1_STRING type. A mask of permissible types is passed
18155 and it chooses the "minimal" type to use or an error if not type
18160 * Add function equivalents to the various macros in asn1.h. The old
18161 macros are retained with an `M_` prefix. Code inside the library can
18162 use the `M_` macros. External code (including the openssl utility)
18163 should *NOT* in order to be "shared library friendly".
18167 * Add various functions that can check a certificate's extensions
18168 to see if it usable for various purposes such as SSL client,
18169 server or S/MIME and CAs of these types. This is currently
18170 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18171 verification. Also added a -purpose flag to x509 utility to
18172 print out all the purposes.
18176 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18181 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18182 for, obtain and decode and extension and obtain its critical flag.
18183 This allows all the necessary extension code to be handled in a
18184 single function call.
18188 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18189 platforms. See crypto/rc4/rc4_enc.c for further details.
18193 * New -noout option to asn1parse. This causes no output to be produced
18194 its main use is when combined with -strparse and -out to extract data
18195 from a file (which may not be in ASN.1 format).
18199 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18200 when producing the local key id.
18202 *Richard Levitte <levitte@stacken.kth.se>*
18204 * New option -dhparam in s_server. This allows a DH parameter file to be
18205 stated explicitly. If it is not stated then it tries the first server
18206 certificate file. The previous behaviour hard coded the filename
18211 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18212 a public key to be input or output. For example:
18213 openssl rsa -in key.pem -pubout -out pubkey.pem
18214 Also added necessary DSA public key functions to handle this.
18218 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18219 in the message. This was handled by allowing
18220 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18222 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18224 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18225 to the end of the strings whereas this didn't. This would cause problems
18226 if strings read with d2i_ASN1_bytes() were later modified.
18228 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18230 * Fix for base64 decode bug. When a base64 bio reads only one line of
18231 data and it contains EOF it will end up returning an error. This is
18232 caused by input 46 bytes long. The cause is due to the way base64
18233 BIOs find the start of base64 encoded data. They do this by trying a
18234 trial decode on each line until they find one that works. When they
18235 do a flag is set and it starts again knowing it can pass all the
18236 data directly through the decoder. Unfortunately it doesn't reset
18237 the context it uses. This means that if EOF is reached an attempt
18238 is made to pass two EOFs through the context and this causes the
18239 resulting error. This can also cause other problems as well. As is
18240 usual with these problems it takes *ages* to find and the fix is
18241 trivial: move one line.
18243 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18245 * Ugly workaround to get s_client and s_server working under Windows. The
18246 old code wouldn't work because it needed to select() on sockets and the
18247 tty (for keypresses and to see if data could be written). Win32 only
18248 supports select() on sockets so we select() with a 1s timeout on the
18249 sockets and then see if any characters are waiting to be read, if none
18250 are present then we retry, we also assume we can always write data to
18251 the tty. This isn't nice because the code then blocks until we've
18252 received a complete line of data and it is effectively polling the
18253 keyboard at 1s intervals: however it's quite a bit better than not
18254 working at all :-) A dedicated Windows application might handle this
18255 with an event loop for example.
18259 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18260 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18261 will be called when RSA_sign() and RSA_verify() are used. This is useful
18262 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18263 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18264 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18265 This necessitated the support of an extra signature type NID_md5_sha1
18266 for SSL signatures and modifications to the SSL library to use it instead
18267 of calling RSA_public_decrypt() and RSA_private_encrypt().
18271 * Add new -verify -CAfile and -CApath options to the crl program, these
18272 will lookup a CRL issuers certificate and verify the signature in a
18273 similar way to the verify program. Tidy up the crl program so it
18274 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18275 less strict. It will now permit CRL extensions even if it is not
18276 a V2 CRL: this will allow it to tolerate some broken CRLs.
18280 * Initialize all non-automatic variables each time one of the openssl
18281 sub-programs is started (this is necessary as they may be started
18282 multiple times from the "OpenSSL>" prompt).
18284 *Lennart Bang, Bodo Moeller*
18286 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18287 removing all other RSA functionality (this is what NO_RSA does). This
18288 is so (for example) those in the US can disable those operations covered
18289 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18294 * Non-copying interface to BIO pairs.
18295 (still largely untested)
18299 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18300 ASCII string. This was handled independently in various places before.
18304 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18305 UTF8 strings a character at a time.
18309 * Use client_version from client hello to select the protocol
18310 (s23_srvr.c) and for RSA client key exchange verification
18311 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18315 * Add various utility functions to handle SPKACs, these were previously
18316 handled by poking round in the structure internals. Added new function
18317 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18318 print, verify and generate SPKACs. Based on an original idea from
18319 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18323 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18327 * Allow the config file extension section to be overwritten on the
18328 command line. Based on an original idea from Massimiliano Pala
18329 <madwolf@comune.modena.it>. The new option is called -extensions
18330 and can be applied to ca, req and x509. Also -reqexts to override
18331 the request extensions in req and -crlexts to override the crl extensions
18336 * Add new feature to the SPKAC handling in ca. Now you can include
18337 the same field multiple times by preceding it by "XXXX." for example:
18340 this is the same syntax as used in the req config file.
18344 * Allow certificate extensions to be added to certificate requests. These
18345 are specified in a 'req_extensions' option of the req section of the
18346 config file. They can be printed out with the -text option to req but
18347 are otherwise ignored at present.
18351 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18352 data read consists of only the final block it would not decrypted because
18353 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18354 A misplaced 'break' also meant the decrypted final block might not be
18355 copied until the next read.
18359 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18360 a few extra parameters to the DH structure: these will be useful if
18361 for example we want the value of 'q' or implement X9.42 DH.
18365 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18366 provides hooks that allow the default DSA functions or functions on a
18367 "per key" basis to be replaced. This allows hardware acceleration and
18368 hardware key storage to be handled without major modification to the
18369 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18370 associated functions.
18374 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18375 as "read only": it can't be written to and the buffer it points to will
18376 not be freed. Reading from a read only BIO is much more efficient than
18377 a normal memory BIO. This was added because there are several times when
18378 an area of memory needs to be read from a BIO. The previous method was
18379 to create a memory BIO and write the data to it, this results in two
18380 copies of the data and an O(n^2) reading algorithm. There is a new
18381 function BIO_new_mem_buf() which creates a read only memory BIO from
18382 an area of memory. Also modified the PKCS#7 routines to use read only
18387 * Bugfix: ssl23_get_client_hello did not work properly when called in
18388 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18389 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18390 but a retry condition occurred while trying to read the rest.
18394 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18395 NID_pkcs7_encrypted by default: this was wrong since this should almost
18396 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18397 the encrypted data type: this is a more sensible place to put it and it
18398 allows the PKCS#12 code to be tidied up that duplicated this
18403 * Changed obj_dat.pl script so it takes its input and output files on
18404 the command line. This should avoid shell escape redirection problems
18409 * Initial support for certificate extension requests, these are included
18410 in things like Xenroll certificate requests. Included functions to allow
18411 extensions to be obtained and added.
18415 * -crlf option to s_client and s_server for sending newlines as
18416 CRLF (as required by many protocols).
18420 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18422 * Install libRSAglue.a when OpenSSL is built with RSAref.
18424 *Ralf S. Engelschall*
18426 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18428 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18430 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18435 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18436 DH parameters/keys (q is lost during that conversion, but the resulting
18437 DH parameters contain its length).
18439 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18440 much faster than DH_generate_parameters (which creates parameters
18441 where `p = 2*q + 1`), and also the smaller q makes DH computations
18442 much more efficient (160-bit exponentiation instead of 1024-bit
18443 exponentiation); so this provides a convenient way to support DHE
18444 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18445 utter importance to use
18446 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18448 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18449 when such DH parameters are used, because otherwise small subgroup
18450 attacks may become possible!
18454 * Avoid memory leak in i2d_DHparams.
18458 * Allow the -k option to be used more than once in the enc program:
18459 this allows the same encrypted message to be read by multiple recipients.
18463 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18464 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18465 it will always use the numerical form of the OID, even if it has a short
18470 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18471 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18472 otherwise bn_mod_exp was called. In the case of hardware keys for example
18473 no private key components need be present and it might store extra data
18474 in the RSA structure, which cannot be accessed from bn_mod_exp.
18475 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18476 private key operations.
18480 * Added support for SPARC Linux.
18484 * pem_password_cb function type incompatibly changed from
18485 typedef int pem_password_cb(char *buf, int size, int rwflag);
18487 ....(char *buf, int size, int rwflag, void *userdata);
18488 so that applications can pass data to their callbacks:
18489 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18490 additional void * argument, which is just handed through whenever
18491 the password callback is called.
18493 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18495 New function SSL_CTX_set_default_passwd_cb_userdata.
18497 Compatibility note: As many C implementations push function arguments
18498 onto the stack in reverse order, the new library version is likely to
18499 interoperate with programs that have been compiled with the old
18500 pem_password_cb definition (PEM_whatever takes some data that
18501 happens to be on the stack as its last argument, and the callback
18502 just ignores this garbage); but there is no guarantee whatsoever that
18505 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18506 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18507 problems not only on Windows, but also on some Unix platforms.
18508 To avoid problematic command lines, these definitions are now in an
18509 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18510 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18514 * MIPS III/IV assembler module is reimplemented.
18518 * More DES library cleanups: remove references to srand/rand and
18519 delete an unused file.
18523 * Add support for the free Netwide assembler (NASM) under Win32,
18524 since not many people have MASM (ml) and it can be hard to obtain.
18525 This is currently experimental but it seems to work OK and pass all
18526 the tests. Check out INSTALL.W32 for info.
18530 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18531 without temporary keys kept an extra copy of the server key,
18532 and connections with temporary keys did not free everything in case
18537 * New function RSA_check_key and new openssl rsa option -check
18538 for verifying the consistency of RSA keys.
18540 *Ulf Moeller, Bodo Moeller*
18542 * Various changes to make Win32 compile work:
18543 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18544 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18545 comparison" warnings.
18546 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18550 * Add a debugging option to PKCS#5 v2 key generation function: when
18551 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18552 derived keys are printed to stderr.
18556 * Copy the flags in ASN1_STRING_dup().
18558 *Roman E. Pavlov <pre@mo.msk.ru>*
18560 * The x509 application mishandled signing requests containing DSA
18561 keys when the signing key was also DSA and the parameters didn't match.
18563 It was supposed to omit the parameters when they matched the signing key:
18564 the verifying software was then supposed to automatically use the CA's
18565 parameters if they were absent from the end user certificate.
18567 Omitting parameters is no longer recommended. The test was also
18568 the wrong way round! This was probably due to unusual behaviour in
18569 EVP_cmp_parameters() which returns 1 if the parameters match.
18570 This meant that parameters were omitted when they *didn't* match and
18571 the certificate was useless. Certificates signed with 'ca' didn't have
18574 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18576 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18577 The interface is as follows:
18578 Applications can use
18579 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18580 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18581 "off" is now the default.
18582 The library internally uses
18583 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18584 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18585 to disable memory-checking temporarily.
18587 Some inconsistent states that previously were possible (and were
18588 even the default) are now avoided.
18590 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18591 with each memory chunk allocated; this is occasionally more helpful
18592 than just having a counter.
18594 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18596 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18601 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18602 which largely parallels "options", but is for changing API behaviour,
18603 whereas "options" are about protocol behaviour.
18604 Initial "mode" flags are:
18606 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18607 a single record has been written.
18608 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18609 retries use the same buffer location.
18610 (But all of the contents must be
18615 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18618 * Fix problems with no-hmac etc.
18620 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18622 * New functions RSA_get_default_method(), RSA_set_method() and
18623 RSA_get_method(). These allows replacement of RSA_METHODs without having
18624 to mess around with the internals of an RSA structure.
18628 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18629 Also really enable memory leak checks in openssl.c and in some
18632 *Chad C. Mulligan, Bodo Moeller*
18634 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18635 up the length of negative integers. This has now been simplified to just
18636 store the length when it is first determined and use it later, rather
18637 than trying to keep track of where data is copied and updating it to
18639 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18641 * Add a new function PKCS7_signatureVerify. This allows the verification
18642 of a PKCS#7 signature but with the signing certificate passed to the
18643 function itself. This contrasts with PKCS7_dataVerify which assumes the
18644 certificate is present in the PKCS#7 structure. This isn't always the
18645 case: certificates can be omitted from a PKCS#7 structure and be
18646 distributed by "out of band" means (such as a certificate database).
18650 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18651 function prototypes in pem.h, also change util/mkdef.pl to add the
18652 necessary function names.
18656 * mk1mf.pl (used by Windows builds) did not properly read the
18657 options set by Configure in the top level Makefile, and Configure
18658 was not even able to write more than one option correctly.
18659 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18663 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18664 file to be loaded from a BIO or FILE pointer. The BIO version will
18665 for example allow memory BIOs to contain config info.
18669 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18670 Whoever hopes to achieve shared-library compatibility across versions
18671 must use this, not the compile-time macro.
18672 (Exercise 0.9.4: Which is the minimum library version required by
18674 Note: All this applies only to multi-threaded programs, others don't
18679 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18680 through a BIO pair triggered the default case, i.e.
18681 SSLerr(...,SSL_R_UNKNOWN_STATE).
18685 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18686 can use the SSL library even if none of the specific BIOs is
18691 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18692 for the encoded length.
18694 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18696 * Add initial documentation of the X509V3 functions.
18700 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18701 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18702 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18703 secure PKCS#8 private key format with a high iteration count.
18707 * Fix determination of Perl interpreter: A perl or perl5
18708 *directory* in $PATH was also accepted as the interpreter.
18710 *Ralf S. Engelschall*
18712 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18713 wrong with it but it was very old and did things like calling
18714 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18715 unusual formatting.
18719 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18720 to use the new extension code.
18724 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18725 with macros. This should make it easier to change their form, add extra
18726 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18731 * Add to configuration table a new entry that can specify an alternative
18732 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18733 according to Mark Crispin <MRC@Panda.COM>.
18737 * DES CBC did not update the IV. Weird.
18741 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18742 Changing the behaviour of the former might break existing programs --
18743 where IV updating is needed, des_ncbc_encrypt can be used.
18746 * When bntest is run from "make test" it drives bc to check its
18747 calculations, as well as internally checking them. If an internal check
18748 fails, it needs to cause bc to give a non-zero result or make test carries
18749 on without noticing the failure. Fixed.
18753 * DES library cleanups.
18757 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18758 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18759 ciphers. NOTE: although the key derivation function has been verified
18760 against some published test vectors it has not been extensively tested
18761 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18766 * Instead of "mkdir -p", which is not fully portable, use new
18767 Perl script "util/mkdir-p.pl".
18771 * Rewrite the way password based encryption (PBE) is handled. It used to
18772 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18773 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18774 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18775 the 'parameter' field of the AlgorithmIdentifier is passed to the
18776 underlying key generation function so it must do its own ASN1 parsing.
18777 This has also changed the EVP_PBE_CipherInit() function which now has a
18778 'parameter' argument instead of literal salt and iteration count values
18779 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18783 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18784 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18785 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18786 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18787 value was just used as a "magic string" and not used directly its
18788 value doesn't matter.
18792 * Introduce some semblance of const correctness to BN. Shame C doesn't
18797 * "linux-sparc64" configuration (ultrapenguin).
18799 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18800 "linux-sparc" configuration.
18802 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18804 * config now generates no-xxx options for missing ciphers.
18808 * Support the EBCDIC character set (work in progress).
18809 File ebcdic.c not yet included because it has a different license.
18811 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18813 * Support BS2000/OSD-POSIX.
18815 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18817 * Make callbacks for key generation use `void *` instead of `char *`.
18821 * Make S/MIME samples compile (not yet tested).
18825 * Additional typesafe stacks.
18829 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18833 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18835 * New configuration variant "sco5-gcc".
18837 * Updated some demos.
18839 *Sean O Riordain, Wade Scholine*
18841 * Add missing BIO_free at exit of pkcs12 application.
18845 * Fix memory leak in conf.c.
18849 * Updates for Win32 to assembler version of MD5.
18853 * Set #! path to perl in `apps/der_chop` to where we found it
18854 instead of using a fixed path.
18858 * SHA library changes for irix64-mips4-cc.
18862 * Improvements for VMS support.
18866 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18868 * Bignum library bug fix. IRIX 6 passes "make test" now!
18869 This also avoids the problems with SC4.2 and unpatched SC5.
18871 *Andy Polyakov <appro@fy.chalmers.se>*
18873 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18874 These are required because of the typesafe stack would otherwise break
18875 existing code. If old code used a structure member which used to be STACK
18876 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18877 sk_num or sk_value it would produce an error because the num, data members
18878 are not present in STACK_OF. Now it just produces a warning. sk_set
18879 replaces the old method of assigning a value to sk_value
18880 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18881 that does this will no longer work (and should use sk_set instead) but
18882 this could be regarded as a "questionable" behaviour anyway.
18886 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18887 correctly handle encrypted S/MIME data.
18891 * Change type of various DES function arguments from des_cblock
18892 (which means, in function argument declarations, pointer to char)
18893 to des_cblock * (meaning pointer to array with 8 char elements),
18894 which allows the compiler to do more typechecking; it was like
18895 that back in SSLeay, but with lots of ugly casts.
18897 Introduce new type const_des_cblock.
18901 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18902 problems: find RecipientInfo structure that matches recipient certificate
18903 and initialise the ASN1 structures properly based on passed cipher.
18907 * Belatedly make the BN tests actually check the results.
18911 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18912 to and from BNs: it was completely broken. New compilation option
18913 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18914 key elements as negative integers.
18918 * Reorganize and speed up MD5.
18920 *Andy Polyakov <appro@fy.chalmers.se>*
18924 *Richard Levitte <richard@levitte.org>*
18926 * New option -out to asn1parse to allow the parsed structure to be
18927 output to a file. This is most useful when combined with the -strparse
18928 option to examine the output of things like OCTET STRINGS.
18932 * Make SSL library a little more fool-proof by not requiring any longer
18933 that `SSL_set_{accept,connect}_state` be called before
18934 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18935 in many applications because usually everything *appeared* to work as
18936 intended anyway -- now it really works as intended).
18940 * Move openssl.cnf out of lib/.
18944 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18945 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18946 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18948 *Ralf S. Engelschall*
18950 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18951 handle PKCS#7 enveloped data properly.
18953 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18955 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18956 copying pointers. The cert_st handling is changed by this in
18957 various ways (and thus what used to be known as ctx->default_cert
18958 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18959 any longer when s->cert does not give us what we need).
18960 ssl_cert_instantiate becomes obsolete by this change.
18961 As soon as we've got the new code right (possibly it already is?),
18962 we have solved a couple of bugs of the earlier code where s->cert
18963 was used as if it could not have been shared with other SSL structures.
18965 Note that using the SSL API in certain dirty ways now will result
18966 in different behaviour than observed with earlier library versions:
18967 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18968 does not influence s as it used to.
18970 In order to clean up things more thoroughly, inside SSL_SESSION
18971 we don't use CERT any longer, but a new structure SESS_CERT
18972 that holds per-session data (if available); currently, this is
18973 the peer's certificate chain and, for clients, the server's certificate
18974 and temporary key. CERT holds only those values that can have
18975 meaningful defaults in an SSL_CTX.
18979 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18980 from the internal representation. Various PKCS#7 fixes: remove some
18981 evil casts and set the enc_dig_alg field properly based on the signing
18986 * Allow PKCS#12 password to be set from the command line or the
18987 environment. Let 'ca' get its config file name from the environment
18988 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18993 * Allow certificate policies extension to use an IA5STRING for the
18994 organization field. This is contrary to the PKIX definition but
18995 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19000 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19001 without disallowing inline assembler and the like for non-pedantic builds.
19005 * Support Borland C++ builder.
19007 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19013 * SHA-1 cleanups and performance enhancements.
19015 *Andy Polyakov <appro@fy.chalmers.se>*
19017 * Sparc v8plus assembler for the bignum library.
19019 *Andy Polyakov <appro@fy.chalmers.se>*
19021 * Accept any -xxx and +xxx compiler options in Configure.
19025 * Update HPUX configuration.
19029 * Add missing `sk_<type>_unshift()` function to safestack.h
19031 *Ralf S. Engelschall*
19033 * New function SSL_CTX_use_certificate_chain_file that sets the
19034 "extra_cert"s in addition to the certificate. (This makes sense
19035 only for "PEM" format files, as chains as a whole are not
19040 * Support verify_depth from the SSL API.
19041 x509_vfy.c had what can be considered an off-by-one-error:
19042 Its depth (which was not part of the external interface)
19043 was actually counting the number of certificates in a chain;
19044 now it really counts the depth.
19048 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19049 instead of X509err, which often resulted in confusing error
19050 messages since the error codes are not globally unique
19051 (e.g. an alleged error in ssl3_accept when a certificate
19052 didn't match the private key).
19054 * New function SSL_CTX_set_session_id_context that allows to set a default
19055 value (so that you don't need SSL_set_session_id_context for each
19056 connection using the SSL_CTX).
19060 * OAEP decoding bug fix.
19064 * Support INSTALL_PREFIX for package builders, as proposed by
19069 * New Configure options "threads" and "no-threads". For systems
19070 where the proper compiler options are known (currently Solaris
19071 and Linux), "threads" is the default.
19075 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19079 * Install various scripts to $(OPENSSLDIR)/misc, not to
19080 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19081 such as /usr/local/bin.
19085 * "make linux-shared" to build shared libraries.
19087 *Niels Poppe <niels@netbox.org>*
19089 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19093 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19094 extension adding in x509 utility.
19098 * Remove NOPROTO sections and error code comments.
19102 * Partial rewrite of the DEF file generator to now parse the ANSI
19107 * New Configure options --prefix=DIR and --openssldir=DIR.
19111 * Complete rewrite of the error code script(s). It is all now handled
19112 by one script at the top level which handles error code gathering,
19113 header rewriting and C source file generation. It should be much better
19114 than the old method: it now uses a modified version of Ulf's parser to
19115 read the ANSI prototypes in all header files (thus the old K&R definitions
19116 aren't needed for error creation any more) and do a better job of
19117 translating function codes into names. The old 'ASN1 error code embedded
19118 in a comment' is no longer necessary and it doesn't use .err files which
19119 have now been deleted. Also the error code call doesn't have to appear all
19120 on one line (which resulted in some large lines...).
19124 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19128 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19129 0 (which usually indicates a closed connection), but continue reading.
19133 * Fix some race conditions.
19137 * Add support for CRL distribution points extension. Add Certificate
19138 Policies and CRL distribution points documentation.
19142 * Move the autogenerated header file parts to crypto/opensslconf.h.
19146 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19147 8 of keying material. Merlin has also confirmed interop with this fix
19148 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19150 *Merlin Hughes <merlin@baltimore.ie>*
19152 * Fix lots of warnings.
19154 *Richard Levitte <levitte@stacken.kth.se>*
19156 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19157 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19159 *Richard Levitte <levitte@stacken.kth.se>*
19161 * Fix problems with sizeof(long) == 8.
19163 *Andy Polyakov <appro@fy.chalmers.se>*
19165 * Change functions to ANSI C.
19169 * Fix typos in error codes.
19171 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19173 * Remove defunct assembler files from Configure.
19177 * SPARC v8 assembler BIGNUM implementation.
19179 *Andy Polyakov <appro@fy.chalmers.se>*
19181 * Support for Certificate Policies extension: both print and set.
19182 Various additions to support the r2i method this uses.
19186 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19187 return a const string when you are expecting an allocated buffer.
19191 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19192 types DirectoryString and DisplayText.
19196 * Add code to allow r2i extensions to access the configuration database,
19197 add an LHASH database driver and add several ctx helper functions.
19201 * Fix an evil bug in bn_expand2() which caused various BN functions to
19202 fail when they extended the size of a BIGNUM.
19206 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19207 support typesafe stack.
19211 * Fix typo in SSL_[gs]et_options().
19213 *Nils Frostberg <nils@medcom.se>*
19215 * Delete various functions and files that belonged to the (now obsolete)
19216 old X509V3 handling code.
19220 * New Configure option "rsaref".
19224 * Don't auto-generate pem.h.
19228 * Introduce type-safe ASN.1 SETs.
19232 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19234 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19236 * Introduce type-safe STACKs. This will almost certainly break lots of code
19237 that links with OpenSSL (well at least cause lots of warnings), but fear
19238 not: the conversion is trivial, and it eliminates loads of evil casts. A
19239 few STACKed things have been converted already. Feel free to convert more.
19240 In the fullness of time, I'll do away with the STACK type altogether.
19244 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19245 specified in `<certfile>` by updating the entry in the index.txt file.
19246 This way one no longer has to edit the index.txt file manually for
19247 revoking a certificate. The -revoke option does the gory details now.
19249 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19251 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19252 `-text` option at all and this way the `-noout -text` combination was
19253 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19255 *Ralf S. Engelschall*
19257 * Make sure a corresponding plain text error message exists for the
19258 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19259 verify callback function determined that a certificate was revoked.
19261 *Ralf S. Engelschall*
19263 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19264 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19265 all available ciphers including rc5, which was forgotten until now.
19266 In order to let the testing shell script know which algorithms
19267 are available, a new (up to now undocumented) command
19268 `openssl list-cipher-commands` is used.
19272 * Bugfix: s_client occasionally would sleep in select() when
19273 it should have checked SSL_pending() first.
19277 * New functions DSA_do_sign and DSA_do_verify to provide access to
19278 the raw DSA values prior to ASN.1 encoding.
19282 * Tweaks to Configure
19284 *Niels Poppe <niels@netbox.org>*
19286 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19291 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19295 * New config option to avoid instructions that are illegal on the 80386.
19296 The default code is faster, but requires at least a 486.
19300 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19301 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19302 same as SSL2_VERSION anyway.
19306 * New "-showcerts" option for s_client.
19310 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19311 application. Various cleanups and fixes.
19315 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19316 modify error routines to work internally. Add error codes and PBE init
19317 to library startup routines.
19321 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19322 packing functions to asn1 and evp. Changed function names and error
19323 codes along the way.
19327 * PKCS12 integration: and so it begins... First of several patches to
19328 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19329 objects to objects.h
19333 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19334 and display support for Thawte strong extranet extension.
19338 * Add LinuxPPC support.
19340 *Jeff Dubrule <igor@pobox.org>*
19342 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19343 bn_div_words in alpha.s.
19345 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19347 * Make sure the RSA OAEP test is skipped under -DRSAref because
19348 OAEP isn't supported when OpenSSL is built with RSAref.
19350 *Ulf Moeller <ulf@fitug.de>*
19352 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19353 so they no longer are missing under -DNOPROTO.
19355 *Soren S. Jorvang <soren@t.dk>*
19357 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19359 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19360 doesn't work when the session is reused. Coming soon!
19364 * Fix a security hole, that allows sessions to be reused in the wrong
19365 context thus bypassing client cert protection! All software that uses
19366 client certs and session caches in multiple contexts NEEDS PATCHING to
19367 allow session reuse! A fuller solution is in the works.
19369 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19371 * Some more source tree cleanups (removed obsolete files
19372 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19373 permission on "config" script to be executable) and a fix for the INSTALL
19376 *Ulf Moeller <ulf@fitug.de>*
19378 * Remove some legacy and erroneous uses of malloc, free instead of
19381 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19383 * Make rsa_oaep_test return non-zero on error.
19385 *Ulf Moeller <ulf@fitug.de>*
19387 * Add support for native Solaris shared libraries. Configure
19388 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19389 if someone would make that last step automatic.
19391 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19393 * ctx_size was not built with the right compiler during "make links". Fixed.
19397 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19398 except NULL ciphers". This means the default cipher list will no longer
19399 enable NULL ciphers. They need to be specifically enabled e.g. with
19400 the string "DEFAULT:eNULL".
19404 * Fix to RSA private encryption routines: if p < q then it would
19405 occasionally produce an invalid result. This will only happen with
19406 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19410 * Be less restrictive and allow also `perl util/perlpath.pl
19411 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19412 because this way one can also use an interpreter named `perl5` (which is
19413 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19414 installed as `perl`).
19416 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19418 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19420 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19422 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19423 advapi32.lib to Win32 build and change the pem test comparison
19424 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19425 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19426 and crypto/des/ede_cbcm_enc.c.
19430 * DES quad checksum was broken on big-endian architectures. Fixed.
19434 * Comment out two functions in bio.h that aren't implemented. Fix up the
19435 Win32 test batch file so it (might) work again. The Win32 test batch file
19436 is horrible: I feel ill....
19440 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19441 in e_os.h. Audit of header files to check ANSI and non ANSI
19442 sections: 10 functions were absent from non ANSI section and not exported
19443 from Windows DLLs. Fixed up libeay.num for new functions.
19447 * Make `openssl version` output lines consistent.
19449 *Ralf S. Engelschall*
19451 * Fix Win32 symbol export lists for BIO functions: Added
19452 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19453 to ms/libeay{16,32}.def.
19455 *Ralf S. Engelschall*
19457 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19458 fine under Unix and passes some trivial tests I've now added. But the
19459 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19460 added to make sure no one expects that this stuff really works in the
19461 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19462 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19465 *Ralf S. Engelschall*
19467 * Fix the generation of two part addresses in perl.
19469 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19471 * Add config entry for Linux on MIPS.
19473 *John Tobey <jtobey@channel1.com>*
19475 * Make links whenever Configure is run, unless we are on Windoze.
19479 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19480 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19485 * Add a useful kludge to allow package maintainers to specify compiler and
19486 other platforms details on the command line without having to patch the
19487 Configure script every time: One now can use
19488 `perl Configure <id>:<details>`,
19489 i.e. platform ids are allowed to have details appended
19490 to them (separated by colons). This is treated as there would be a static
19491 pre-configured entry in Configure's %table under key `<id>` with value
19492 `<details>` and `perl Configure <id>` is called. So, when you want to
19493 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19494 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19495 now, which overrides the FreeBSD-elf entry on-the-fly.
19497 *Ralf S. Engelschall*
19499 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19503 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19504 on the `perl Configure ...` command line. This way one can compile
19505 OpenSSL libraries with Position Independent Code (PIC) which is needed
19506 for linking it into DSOs.
19508 *Ralf S. Engelschall*
19510 * Remarkably, export ciphers were totally broken and no-one had noticed!
19515 * Cleaned up the LICENSE document: The official contact for any license
19516 questions now is the OpenSSL core team under openssl-core@openssl.org.
19517 And add a paragraph about the dual-license situation to make sure people
19518 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19519 to the OpenSSL toolkit.
19521 *Ralf S. Engelschall*
19523 * General source tree makefile cleanups: Made `making xxx in yyy...`
19524 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19525 Additionally cleaned up the `make links` target: Remove unnecessary
19526 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19527 to speed processing and no longer clutter the display with confusing
19528 stuff. Instead only the actually done links are displayed.
19530 *Ralf S. Engelschall*
19532 * Permit null encryption ciphersuites, used for authentication only. It used
19533 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19534 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19539 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19540 signed attributes when verifying signatures (this would break them),
19541 the detached data encoding was wrong and public keys obtained using
19542 X509_get_pubkey() weren't freed.
19546 * Add text documentation for the BUFFER functions. Also added a work around
19547 to a Win95 console bug. This was triggered by the password read stuff: the
19548 last character typed gets carried over to the next fread(). If you were
19549 generating a new cert request using 'req' for example then the last
19550 character of the passphrase would be CR which would then enter the first
19555 * Added the new 'Includes OpenSSL Cryptography Software' button as
19556 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19557 button and can be used by applications based on OpenSSL to show the
19558 relationship to the OpenSSL project.
19560 *Ralf S. Engelschall*
19562 * Remove confusing variables in function signatures in files
19563 ssl/ssl_lib.c and ssl/ssl.h.
19565 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19567 * Don't install bss_file.c under PREFIX/include/
19569 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19571 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19572 functions that return function pointers and has support for NT specific
19573 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19574 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19575 unsigned to signed types: this was killing the Win32 compile.
19579 * Add new certificate file to stack functions,
19580 SSL_add_dir_cert_subjects_to_stack() and
19581 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19582 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19583 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19584 This means that Apache-SSL and similar packages don't have to mess around
19585 to add as many CAs as they want to the preferred list.
19589 * Experiment with doxygen documentation. Currently only partially applied to
19591 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19592 openssl.doxy as the configuration file.
19596 * Get rid of remaining C++-style comments which strict C compilers hate.
19598 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19600 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19601 compiled in by default: it has problems with large keys.
19605 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19606 DH private keys and/or callback functions which directly correspond to
19607 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19608 is needed for applications which have to configure certificates on a
19609 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19611 For the RSA certificate situation is makes no difference, but
19612 for the DSA certificate situation this fixes the "no shared cipher"
19613 problem where the OpenSSL cipher selection procedure failed because the
19614 temporary keys were not overtaken from the context and the API provided
19615 no way to reconfigure them.
19616 The new functions now let applications reconfigure the stuff and they
19617 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19618 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19619 non-public-API function ssl_cert_instantiate() is used as a helper
19620 function and also to reduce code redundancy inside ssl_rsa.c.
19622 *Ralf S. Engelschall*
19624 * Move s_server -dcert and -dkey options out of the undocumented feature
19625 area because they are useful for the DSA situation and should be
19626 recognized by the users.
19628 *Ralf S. Engelschall*
19630 * Fix the cipher decision scheme for export ciphers: the export bits are
19631 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19632 SSL_EXP_MASK. So, the original variable has to be used instead of the
19633 already masked variable.
19635 *Richard Levitte <levitte@stacken.kth.se>*
19637 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19639 *Richard Levitte <levitte@stacken.kth.se>*
19641 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19642 from `int` to `unsigned int` because it is a length and initialized by
19643 EVP_DigestFinal() which expects an `unsigned int *`.
19645 *Richard Levitte <levitte@stacken.kth.se>*
19647 * Don't hard-code path to Perl interpreter on shebang line of Configure
19648 script. Instead use the usual Shell->Perl transition trick.
19650 *Ralf S. Engelschall*
19652 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19653 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19654 -noout -modulus` as it's already the case for `openssl rsa -noout
19655 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19656 currently the public key is printed (a decision which was already done by
19657 `openssl dsa -modulus` in the past) which serves a similar purpose.
19658 Additionally the NO_RSA no longer completely removes the whole -modulus
19659 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19662 *Ralf S. Engelschall*
19664 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19665 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19667 *Arne Ansper <arne@ats.cyber.ee>*
19669 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19670 to be added. Now both 'req' and 'ca' can use new objects defined in the
19675 * Add cool BIO that does syslog (or event log on NT).
19677 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19679 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19680 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19681 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19682 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19686 * Add preliminary config info for new extension code.
19690 * Make RSA_NO_PADDING really use no padding.
19692 *Ulf Moeller <ulf@fitug.de>*
19694 * Generate errors when private/public key check is done.
19698 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19699 for some CRL extensions and new objects added.
19703 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19704 key usage extension and fuller support for authority key id.
19708 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19709 padding method for RSA, which is recommended for new applications in PKCS
19710 #1 v2.0 (RFC 2437, October 1998).
19711 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19712 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19713 against Bleichbacher's attack on RSA.
19714 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19717 * Updates to the new SSL compression code
19719 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19721 * Fix so that the version number in the master secret, when passed
19722 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19723 (because the server will not accept higher), that the version number
19724 is 0x03,0x01, not 0x03,0x00
19726 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19728 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19729 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19730 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19734 * Support for RAW extensions where an arbitrary extension can be
19735 created by including its DER encoding. See `apps/openssl.cnf` for
19740 * Make sure latest Perl versions don't interpret some generated C array
19741 code as Perl array code in the crypto/err/err_genc.pl script.
19743 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19745 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19746 not many people have the assembler. Various Win32 compilation fixes and
19747 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19748 build instructions.
19752 * Modify configure script 'Configure' to automatically create crypto/date.h
19753 file under Win32 and also build pem.h from pem.org. New script
19754 util/mkfiles.pl to create the MINFO file on environments that can't do a
19755 'make files': perl util/mkfiles.pl >MINFO should work.
19759 * Major rework of DES function declarations, in the pursuit of correctness
19760 and purity. As a result, many evil casts evaporated, and some weirdness,
19761 too. You may find this causes warnings in your code. Zapping your evil
19762 casts will probably fix them. Mostly.
19766 * Fix for a typo in asn1.h. Bug fix to object creation script
19767 obj_dat.pl. It considered a zero in an object definition to mean
19768 "end of object": none of the objects in objects.h have any zeros
19769 so it wasn't spotted.
19771 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19773 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19774 Masking (CBCM). In the absence of test vectors, the best I have been able
19775 to do is check that the decrypt undoes the encrypt, so far. Send me test
19776 vectors if you have them.
19780 * Correct calculation of key length for export ciphers (too much space was
19781 allocated for null ciphers). This has not been tested!
19785 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19786 message is now correct (it understands "crypto" and "ssl" on its
19787 command line). There is also now an "update" option. This will update
19788 the util/ssleay.num and util/libeay.num files with any new functions.
19790 perl util/mkdef.pl crypto ssl update
19791 it will update them.
19795 * Overhauled the Perl interface:
19796 - ported BN stuff to OpenSSL's different BN library
19797 - made the perl/ source tree CVS-aware
19798 - renamed the package from SSLeay to OpenSSL (the files still contain
19799 their history because I've copied them in the repository)
19800 - removed obsolete files (the test scripts will be replaced
19801 by better Test::Harness variants in the future)
19803 *Ralf S. Engelschall*
19805 * First cut for a very conservative source tree cleanup:
19806 1. merge various obsolete readme texts into doc/ssleay.txt
19807 where we collect the old documents and readme texts.
19808 2. remove the first part of files where I'm already sure that we no
19809 longer need them because of three reasons: either they are just temporary
19810 files which were left by Eric or they are preserved original files where
19811 I've verified that the diff is also available in the CVS via "cvs diff
19812 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19813 the crypto/md/ stuff).
19815 *Ralf S. Engelschall*
19817 * More extension code. Incomplete support for subject and issuer alt
19818 name, issuer and authority key id. Change the i2v function parameters
19819 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19820 what that's for :-) Fix to ASN1 macro which messed up
19821 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19825 * Preliminary support for ENUMERATED type. This is largely copied from the
19830 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19832 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19834 * Make sure `make rehash` target really finds the `openssl` program.
19836 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19838 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19839 like to hear about it if this slows down other processors.
19843 * Add CygWin32 platform information to Configure script.
19845 *Alan Batie <batie@aahz.jf.intel.com>*
19847 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19849 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19851 * New program nseq to manipulate netscape certificate sequences
19855 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19860 * Fixes to BN code. Previously the default was to define BN_RECURSION
19861 but the BN code had some problems that would cause failures when
19862 doing certificate verification and some other functions.
19864 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19866 * Add ASN1 and PEM code to support netscape certificate sequences.
19870 * Add ASN1 and PEM code to support netscape certificate sequences.
19874 * Add several PKIX and private extended key usage OIDs.
19878 * Modify the 'ca' program to handle the new extension code. Modify
19879 openssl.cnf for new extension format, add comments.
19883 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19884 and add a sample to openssl.cnf so req -x509 now adds appropriate
19889 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19890 error code, add initial support to X509_print() and x509 application.
19894 * Takes a deep breath and start adding X509 V3 extension support code. Add
19895 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19896 stuff is currently isolated and isn't even compiled yet.
19900 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19901 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19902 Removed the versions check from X509 routines when loading extensions:
19903 this allows certain broken certificates that don't set the version
19904 properly to be processed.
19908 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19909 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19910 can still be regenerated with "make depend".
19914 * Spelling mistake in C version of CAST-128.
19916 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19918 * Changes to the error generation code. The perl script err-code.pl
19919 now reads in the old error codes and retains the old numbers, only
19920 adding new ones if necessary. It also only changes the .err files if new
19921 codes are added. The makefiles have been modified to only insert errors
19922 when needed (to avoid needlessly modifying header files). This is done
19923 by only inserting errors if the .err file is newer than the auto generated
19924 C file. To rebuild all the error codes from scratch (the old behaviour)
19925 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19926 or delete all the .err files.
19930 * CAST-128 was incorrectly implemented for short keys. The C version has
19931 been fixed, but is untested. The assembler versions are also fixed, but
19932 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19933 to regenerate it if needed.
19934 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19935 Hagino <itojun@kame.net>*
19937 * File was opened incorrectly in randfile.c.
19939 *Ulf Möller <ulf@fitug.de>*
19941 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19942 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19943 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19944 al: it's just almost always a UTCTime. Note this patch adds new error
19945 codes so do a "make errors" if there are problems.
19949 * Correct Linux 1 recognition in config.
19951 *Ulf Möller <ulf@fitug.de>*
19953 * Remove pointless MD5 hash when using DSA keys in ca.
19955 *Anonymous <nobody@replay.com>*
19957 * Generate an error if given an empty string as a cert directory. Also
19958 generate an error if handed NULL (previously returned 0 to indicate an
19959 error, but didn't set one).
19961 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19963 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19967 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19968 parameters. This was causing a warning which killed off the Win32 compile.
19972 * Remove C++ style comments from crypto/bn/bn_local.h.
19974 *Neil Costigan <neil.costigan@celocom.com>*
19976 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19977 based on a text string, looking up short and long names and finally
19978 "dot" format. The "dot" format stuff didn't work. Added new function
19979 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19980 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19981 OID is not part of the table.
19985 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19986 X509_LOOKUP_by_alias().
19990 * Sort openssl functions by name.
19994 * Get the `gendsa` command working and add it to the `list` command. Remove
19995 encryption from sample DSA keys (in case anyone is interested the password
20000 * Make *all* `*_free` functions accept a NULL pointer.
20002 *Frans Heymans <fheymans@isaserver.be>*
20004 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20007 *Anonymous <nobody@replay.com>*
20009 * s_server should send the CAfile as acceptable CAs, not its own cert.
20011 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20013 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20015 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20017 * Temp key "for export" tests were wrong in s3_srvr.c.
20019 *Anonymous <nobody@replay.com>*
20021 * Add prototype for temp key callback functions
20022 SSL_CTX_set_tmp_{rsa,dh}_callback().
20026 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20027 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20031 * X509_name_add_entry() freed the wrong thing after an error.
20033 *Arne Ansper <arne@ats.cyber.ee>*
20035 * rsa_eay.c would attempt to free a NULL context.
20037 *Arne Ansper <arne@ats.cyber.ee>*
20039 * BIO_s_socket() had a broken should_retry() on Windoze.
20041 *Arne Ansper <arne@ats.cyber.ee>*
20043 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20045 *Arne Ansper <arne@ats.cyber.ee>*
20047 * Make sure the already existing X509_STORE->depth variable is initialized
20048 in X509_STORE_new(), but document the fact that this variable is still
20049 unused in the certificate verification process.
20051 *Ralf S. Engelschall*
20053 * Fix the various library and `apps/` files to free up pkeys obtained from
20054 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20058 * Fix reference counting in X509_PUBKEY_get(). This makes
20059 demos/maurice/example2.c work, amongst others, probably.
20061 *Steve Henson and Ben Laurie*
20063 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20064 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20065 are no longer created. This way we have a single and consistent command
20066 line interface `openssl <command>`, similar to `cvs <command>`.
20068 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20070 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20071 BIT STRING wrapper always have zero unused bits.
20075 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20079 * Make the top-level INSTALL documentation easier to understand.
20083 * Makefiles updated to exit if an error occurs in a sub-directory
20084 make (including if user presses ^C) [Paul Sutton]
20086 * Make Montgomery context stuff explicit in RSA data structure.
20090 * Fix build order of pem and err to allow for generated pem.h.
20094 * Fix renumbering bug in X509_NAME_delete_entry().
20098 * Enhanced the err-ins.pl script so it makes the error library number
20099 global and can add a library name. This is needed for external ASN1 and
20100 other error libraries.
20104 * Fixed sk_insert which never worked properly.
20108 * Fix ASN1 macros so they can handle indefinite length constructed
20109 EXPLICIT tags. Some non standard certificates use these: they can now
20114 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20115 into a single doc/ssleay.txt bundle. This way the information is still
20116 preserved but no longer messes up this directory. Now it's new room for
20117 the new set of documentation files.
20119 *Ralf S. Engelschall*
20121 * SETs were incorrectly DER encoded. This was a major pain, because they
20122 shared code with SEQUENCEs, which aren't coded the same. This means that
20123 almost everything to do with SETs or SEQUENCEs has either changed name or
20124 number of arguments.
20126 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20128 * Fix test data to work with the above.
20132 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20133 was already fixed by Eric for 0.9.1 it seems.
20135 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20137 * Autodetect FreeBSD3.
20141 * Fix various bugs in Configure. This affects the following platforms:
20145 unixware-2.0-pentium
20150 * Eliminate generated files from CVS. Reorder tests to regenerate files
20151 before they are needed.
20155 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20159 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20161 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20162 changed SSLeay to OpenSSL in version strings.
20164 *Ralf S. Engelschall*
20166 * Some fixups to the top-level documents.
20170 * Fixed the nasty bug where rsaref.h was not found under compile-time
20171 because the symlink to include/ was missing.
20173 *Ralf S. Engelschall*
20175 * Incorporated the popular no-RSA/DSA-only patches
20176 which allow to compile an RSA-free SSLeay.
20178 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20180 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20181 when "ssleay" is still not found.
20183 *Ralf S. Engelschall*
20185 * Added more platforms to Configure: Cray T3E, HPUX 11,
20187 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20189 * Updated the README file.
20191 *Ralf S. Engelschall*
20193 * Added various .cvsignore files in the CVS repository subdirs
20194 to make a "cvs update" really silent.
20196 *Ralf S. Engelschall*
20198 * Recompiled the error-definition header files and added
20199 missing symbols to the Win32 linker tables.
20201 *Ralf S. Engelschall*
20203 * Cleaned up the top-level documents;
20204 o new files: CHANGES and LICENSE
20205 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20206 o merged COPYRIGHT into LICENSE
20207 o removed obsolete TODO file
20208 o renamed MICROSOFT to INSTALL.W32
20210 *Ralf S. Engelschall*
20212 * Removed dummy files from the 0.9.1b source tree:
20213 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20214 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20215 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20216 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20217 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20219 *Ralf S. Engelschall*
20221 * Added various platform portability fixes.
20225 * The Genesis of the OpenSSL rpject:
20226 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20227 Young and Tim J. Hudson created while they were working for C2Net until
20230 *The OpenSSL Project*
20232 ### Changes between 0.9.0b and 0.9.1b [not released]
20234 * Updated a few CA certificates under certs/
20238 * Changed some BIGNUM api stuff.
20242 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20243 DGUX x86, Linux Alpha, etc.
20247 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20248 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20253 * Add -strparse option to asn1pars program which parses nested
20256 *Dr Stephen Henson <shenson@bigfoot.com>*
20258 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20262 * DSA fix for "ca" program.
20266 * Added "-genkey" option to "dsaparam" program.
20270 * Added RIPE MD160 (rmd160) message digest.
20274 * Added -a (all) option to "ssleay version" command.
20278 * Added PLATFORM define which is the id given to Configure.
20282 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20286 * Extended the ASN.1 parser routines.
20290 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20294 * Added a BN_CTX to the BN library.
20298 * Fixed the weak key values in DES library
20302 * Changed API in EVP library for cipher aliases.
20306 * Added support for RC2/64bit cipher.
20310 * Converted the lhash library to the crypto/mem.c functions.
20314 * Added more recognized ASN.1 object ids.
20318 * Added more RSA padding checks for SSL/TLS.
20322 * Added BIO proxy/filter functionality.
20326 * Added extra_certs to SSL_CTX which can be used
20327 send extra CA certificates to the client in the CA cert chain sending
20328 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20332 * Now Fortezza is denied in the authentication phase because
20333 this is key exchange mechanism is not supported by SSLeay at all.
20337 * Additional PKCS1 checks.
20341 * Support the string "TLSv1" for all TLS v1 ciphers.
20345 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20346 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20350 * Fixed a few memory leaks.
20354 * Fixed various code and comment typos.
20358 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20359 bytes sent in the client random.
20361 *Edward Bishop <ebishop@spyglass.com>*
20365 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20366 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20367 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20368 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20369 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20370 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20371 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20372 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20373 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20374 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20375 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20376 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20377 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20378 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20379 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20380 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20381 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20382 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20383 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20384 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20385 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20386 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20387 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20388 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20389 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20390 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20391 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20392 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20393 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20394 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20395 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20396 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20397 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20398 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20399 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20400 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20401 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20402 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20403 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20404 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20405 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20406 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20407 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20408 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20409 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20410 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20411 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20412 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20413 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20414 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20415 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20416 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20417 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20418 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20419 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20420 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20421 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20422 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20423 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20424 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20425 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20426 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20427 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20428 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20429 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20430 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20431 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20432 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20433 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20434 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20435 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20436 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20437 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20438 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20439 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20440 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20441 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20442 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20443 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20444 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20445 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20446 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20447 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20448 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20449 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20450 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20451 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20452 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20453 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20454 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20455 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20456 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20457 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20458 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20459 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20460 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20461 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20462 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20463 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20464 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20465 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20466 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20467 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20468 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20469 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20470 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20471 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20472 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20473 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20474 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20475 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20476 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20477 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20478 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20479 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20480 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20481 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20482 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20483 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20484 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20485 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20486 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20487 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20488 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20489 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20490 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20491 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20492 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20493 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20494 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20495 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20496 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20497 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20498 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20499 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20500 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20501 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20502 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20503 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20504 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20505 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20506 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20507 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20508 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20509 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20510 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20511 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20512 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20513 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20514 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20515 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20516 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20517 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20518 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20519 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20520 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20521 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20522 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20523 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20524 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20525 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20526 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20527 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20528 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20529 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20530 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20531 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20532 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20533 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20534 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20535 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20536 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20537 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20538 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20539 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20540 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20541 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20542 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20543 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20544 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20545 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20546 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20547 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20548 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20549 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655