4 This is a detailed breakdown of significant changes. For a high-level overview
5 of changes in each release, see [NEWS.md](./NEWS.md).
7 For a full list of changes, see the [git commit log][log] and pick the
8 appropriate release branch.
10 [log]: https://github.com/openssl/openssl/commits/
15 - [OpenSSL 3.3](#openssl-33)
16 - [OpenSSL 3.2](#openssl-32)
17 - [OpenSSL 3.1](#openssl-31)
18 - [OpenSSL 3.0](#openssl-30)
19 - [OpenSSL 1.1.1](#openssl-111)
20 - [OpenSSL 1.1.0](#openssl-110)
21 - [OpenSSL 1.0.2](#openssl-102)
22 - [OpenSSL 1.0.1](#openssl-101)
23 - [OpenSSL 1.0.0](#openssl-100)
24 - [OpenSSL 0.9.x](#openssl-09x)
29 ### Changes between 3.2 and 3.3 [xx XXX xxxx]
31 * The activate and soft_load configuration settings for providers in
32 openssl.cnf have been updated to require a value of [1|yes|true|on]
33 (in lower or UPPER case) to enable the setting. Conversely a value
34 of [0|no|false|off] will disable the setting. All other values, or the
35 omission of a value for these settings will result in an error.
39 * Added `-set_issuer` and `-set_subject` options to `openssl x509` to
40 override the Issuer and Subject when creating a certificate. The `-subj`
41 option now is an alias for `-set_subject`.
43 *Job Snijders, George Michaelson*
45 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
46 if called with a NULL stack argument.
50 * In `openssl speed`, changed the default hash function used with `hmac` from
55 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
56 - `certProfile` request message header and respective `-profile` CLI option
57 - support for delayed delivery of all types of response messages
61 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
62 be less hard coded in the build file templates, and to allow easier
63 addition of more exporters. With that, an exporter for CMake is also
68 * The BLAKE2s hash algorithm matches BLAKE2b's support
69 for configurable output length.
73 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
74 server to prefer session resumption using PSK-only key exchange over PSK
75 with DHE, if both are available.
77 *Markus Minichmayr, Tapkey GmbH*
82 ### Changes between 3.2.0 and 3.2.1 [xx XXX xxxx]
84 * When function EVP_PKEY_public_check() is called on RSA public keys,
85 a computation is done to confirm that the RSA modulus, n, is composite.
86 For valid RSA keys, n is a product of two or more large primes and this
87 computation completes quickly. However, if n is an overly large prime,
88 then this computation would take a long time.
90 An application that calls EVP_PKEY_public_check() and supplies an RSA key
91 obtained from an untrusted source could be vulnerable to a Denial of Service
94 The function EVP_PKEY_public_check() is not called from other OpenSSL
95 functions however it is called from the OpenSSL pkey command line
96 application. For that reason that application is also vulnerable if used
97 with the "-pubin" and "-check" options on untrusted data.
99 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
100 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
106 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
107 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
112 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
113 for PowerPC CPUs saves the contents of vector registers in different
114 order than they are restored. Thus the contents of some of these vector
115 registers is corrupted when returning to the caller. The vulnerable code is
116 used only on newer PowerPC processors supporting the PowerISA 2.07
119 The consequences of this kind of internal application state corruption can
120 be various - from no consequences, if the calling application does not
121 depend on the contents of non-volatile XMM registers at all, to the worst
122 consequences, where the attacker could get complete control of the
123 application process. However unless the compiler uses the vector registers
124 for storing pointers, the most likely consequence, if any, would be an
125 incorrect result of some application dependent calculations or a crash
126 leading to a denial of service.
132 * Disable building QUIC server utility when OpenSSL is configured with
137 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
139 * The EVP_PKEY_fromdata function has been augmented to allow for the derivation
140 of CRT (Chinese Remainder Theorem) parameters when requested. See the
141 OSSL_PKEY_PARAM_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
145 * The BLAKE2b hash algorithm supports a configurable output length
146 by setting the "size" parameter.
148 *Čestmír Kalina and Tomáš Mráz*
150 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
154 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
155 and the corresponding provider-storemgmt API function
156 OSSL_FUNC_store_delete().
160 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
161 a passphrase callback when opening a store.
165 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
166 from 8 bytes to 16 bytes.
167 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
168 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
169 requires a salt length of 128 bits. This affects OpenSSL command line
170 applications such as "genrsa" and "pkcs8" and API's such as
171 PEM_write_bio_PrivateKey() that are reliant on the default value.
172 The additional commandline option 'saltlen' has been added to the
173 OpenSSL command line applications for "pkcs8" and "enc" to allow the
174 salt length to be set to a non default value.
178 * Changed the default value of the `ess_cert_id_alg` configuration
179 option which is used to calculate the TSA's public key certificate
180 identifier. The default algorithm is updated to be sha256 instead
183 *Małgorzata Olszówka*
185 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
186 table for point multiplication of the base point, which increases the size of
187 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
188 been added to disable the precomputed table.
192 * Added client side support for QUIC
194 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
196 * Added multiple tutorials on the OpenSSL library and in particular
197 on writing various clients (using TLS and QUIC protocols) with libssl.
201 * Added secp384r1 implementation using Solinas' reduction to improve
202 speed of the NIST P-384 elliptic curve. To enable the implementation
203 the build option `enable-ec_nistp_64_gcc_128` must be used.
207 * Improved RFC7468 compliance of the asn1parse command.
209 *Matthias St. Pierre*
211 * Added SHA256/192 algorithm support.
215 * Added support for securely getting root CA certificate update in
220 * Improved contention on global write locks by using more read locks where
225 * Improved performance of OSSL_PARAM lookups in performance critical
230 * Added the SSL_get0_group_name() function to provide access to the
231 name of the group used for the TLS key exchange.
235 * Provide a new configure option `no-http` that can be used to disable the
236 HTTP support. Provide new configure options `no-apps` and `no-docs` to
237 disable building the openssl command line application and the documentation.
241 * Provide a new configure option `no-ecx` that can be used to disable the
242 X25519, X448, and EdDSA support.
246 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
247 the EVP_KDF_CTX_set_params() function they are now concatenated not just
248 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
252 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
253 the provider context as a parameter.
257 * TLS round-trip time calculation was added by a Brigham Young University
258 Capstone team partnering with Sandia National Laboratories. A new function
259 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
264 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
265 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
266 option. Use of the "advanced" s_client command command via the "-adv" option
271 * Added an "advanced" command mode to s_client. Use this with the "-adv"
272 option. The old "basic" command mode recognises certain letters that must
273 always appear at the start of a line and cannot be escaped. The advanced
274 command mode enables commands to be entered anywhere and there is an
275 escaping mechanism. After starting s_client with "-adv" type "{help}"
276 to show a list of available commands.
280 * Add Raw Public Key (RFC7250) support. Authentication is supported
281 by matching keys against either local policy (TLSA records synthesised
282 from the expected keys) or DANE (TLSA records obtained by the
283 application from DNS). TLSA records will also match the same key in
284 the server certificate, should RPK use not happen to be negotiated.
288 * Added support for modular exponentiation and CRT offloading for the
293 * Added further assembler code for the RISC-V architecture.
297 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
298 from a given EC_GROUP.
302 * Improved support for non-default library contexts and property queries
303 when parsing PKCS#12 files.
307 * Implemented support for all five instances of EdDSA from RFC8032:
308 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
309 The streaming is not yet supported for the HashEdDSA variants
310 (Ed25519ph and Ed448ph).
314 * Added SM4 optimization for ARM processors using ASIMD and AES HW
319 * Implemented SM4-XTS support.
323 * Added platform-agnostic OSSL_sleep() function.
327 * Implemented deterministic ECDSA signatures (RFC6979) support.
331 * Implemented AES-GCM-SIV (RFC8452) support.
335 * Added support for pluggable (provider-based) TLS signature algorithms.
336 This enables TLS 1.3 authentication operations with algorithms embedded
337 in providers not included by default in OpenSSL. In combination with
338 the already available pluggable KEM and X.509 support, this enables
339 for example suitable providers to deliver post-quantum or quantum-safe
340 cryptography to OpenSSL users.
344 * Added support for pluggable (provider-based) CMS signature algorithms.
345 This enables CMS sign and verify operations with algorithms embedded
346 in providers not included by default in OpenSSL.
350 * Added support for Hybrid Public Key Encryption (HPKE) as defined
351 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
352 Message Layer Security (MLS) and other IETF specifications.
353 HPKE can also be used by other applications that require
354 encrypting "to" an ECDH public key. External APIs are defined in
355 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
359 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
364 * Add support for certificate compression (RFC8879), including
365 library support for Brotli and Zstandard compression.
369 * Add the ability to add custom attributes to PKCS12 files. Add a new API
370 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
371 for a user specified callback and optional argument.
372 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
373 added to the existing STACK_OF attrs.
377 * Major refactor of the libssl record layer.
381 * Add a mac salt length option for the pkcs12 command.
385 * Add more SRTP protection profiles from RFC8723 and RFC8269.
389 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
391 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
393 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
394 supported and enabled.
398 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
399 to the list of ciphersuites providing Perfect Forward Secrecy as
400 required by SECLEVEL >= 3.
402 *Dmitry Belyavskiy, Nicola Tuveri*
404 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
405 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
406 SSL_get0_iana_groups() function-like macro, retrieves the list of
407 supported groups sent by the peer.
408 The function SSL_client_hello_get_extension_order() populates
409 a caller-supplied array with the list of extension types present in the
410 ClientHello, in order of appearance.
414 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
415 to make it possible to use empty passphrase strings.
419 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
423 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
424 to change functions used for allocating the memory of asynchronous call stack.
428 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
432 * A failure exit code is returned when using the openssl x509 command to check
433 certificate attributes and the checks fail.
437 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
438 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
439 of 160 bits and above and less than 224 bits were previously accepted by
440 default but are now no longer allowed. By default TLS compression was
441 already disabled in previous OpenSSL versions. At security level 2 it cannot
446 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
451 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
452 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
453 will need to load the legacy crypto provider.
457 * CCM8 cipher suites in TLS have been downgraded to security level zero
458 because they use a short authentication tag which lowers their strength.
462 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
467 * Add X.509 certificate codeSigning purpose and related checks on key usage and
468 extended key usage of the leaf certificate according to the CA/Browser Forum.
472 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
473 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
474 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
475 X.509 version 3 if the certificate information includes X.509 extensions.
479 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
480 such as adding a trace facility for debugging certificate chain building.
484 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
485 in particular supporting requests for central key generation, generalized
486 polling, and various types of genm/genp exchanges defined in CMP Updates.
490 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
491 like correcting the TLS and proxy support and adding tracing for debugging.
495 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
499 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
500 a certificate to be added is already present. `CMS_sign_ex()` and
501 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
502 and no longer throw an error for them.
506 * Fixed and extended `util/check-format.pl` for checking adherence to the
507 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
508 The checks are meanwhile more complete and yield fewer false positives.
512 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
513 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
514 calls. They can be used as the transport BIOs for QUIC.
516 *Hugo Landau, Matt Caswell and Tomáš Mráz*
518 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
519 sending and receiving multiple messages in a single call. An implementation
520 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
524 * Support for loading root certificates from the Windows certificate store
525 has been added. The support is in the form of a store which recognises the
526 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
527 arguments. This store is built by default and can be disabled using the new
528 compile-time option `no-winstore`. This store is not currently used by
529 default and must be loaded explicitly using the above store URI. It is
530 expected to be loaded by default in the future.
534 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
535 kernel versions that support KTLS have a known bug in CCM processing. That
536 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
537 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
542 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
547 * Zerocopy KTLS sendfile() support on Linux.
551 * The OBJ_ calls are now thread safe using a global lock.
555 * New parameter `-digest` for openssl cms command allowing signing
556 pre-computed digests and new CMS API functions supporting that
561 * OPENSSL_malloc() and other allocation functions now raise errors on
562 allocation failures. The callers do not need to explicitly raise errors
563 unless they want to for tracing purposes.
567 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
568 decryption as a protection against Bleichenbacher-like attacks.
569 The RSA decryption API will now return a randomly generated deterministic
570 message instead of an error in case it detects an error when checking
571 padding during PKCS#1 v1.5 decryption. This is a general protection against
572 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
574 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
575 on the RSA decryption context.
579 * Added support for Brainpool curves in TLS-1.3.
581 *Bernd Edlinger and Matt Caswell*
583 * Added OpenBSD specific build targets.
587 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
588 a basic thread pool implementation for select platforms.
595 ### Changes between 3.1.4 and 3.1.5 [xx XXX xxxx]
597 * Fix excessive time spent in DH check / generation with large Q parameter
600 Applications that use the functions DH_generate_key() to generate an
601 X9.42 DH key may experience long delays. Likewise, applications that use
602 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
603 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
604 Where the key or parameters that are being checked have been obtained from
605 an untrusted source this may lead to a Denial of Service.
611 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
613 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
614 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
615 that alter the key or IV length ([CVE-2023-5363]).
619 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
621 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
623 The POLY1305 MAC (message authentication code) implementation in OpenSSL
624 does not save the contents of non-volatile XMM registers on Windows 64
625 platform when calculating the MAC of data larger than 64 bytes. Before
626 returning to the caller all the XMM registers are set to zero rather than
627 restoring their previous content. The vulnerable code is used only on newer
628 x86_64 processors supporting the AVX512-IFMA instructions.
630 The consequences of this kind of internal application state corruption can
631 be various - from no consequences, if the calling application does not
632 depend on the contents of non-volatile XMM registers at all, to the worst
633 consequences, where the attacker could get complete control of the
634 application process. However given the contents of the registers are just
635 zeroized so the attacker cannot put arbitrary values inside, the most likely
636 consequence, if any, would be an incorrect result of some application
637 dependent calculations or a crash leading to a denial of service.
643 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
645 * Fix excessive time spent checking DH q parameter value.
647 The function DH_check() performs various checks on DH parameters. After
648 fixing CVE-2023-3446 it was discovered that a large q parameter value can
649 also trigger an overly long computation during some of these checks.
650 A correct q value, if present, cannot be larger than the modulus p
651 parameter, thus it is unnecessary to perform these checks if q is larger
654 If DH_check() is called with such q parameter value,
655 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
656 intensive checks are skipped.
662 * Fix DH_check() excessive time with over sized modulus.
664 The function DH_check() performs various checks on DH parameters. One of
665 those checks confirms that the modulus ("p" parameter) is not too large.
666 Trying to use a very large modulus is slow and OpenSSL will not normally use
667 a modulus which is over 10,000 bits in length.
669 However the DH_check() function checks numerous aspects of the key or
670 parameters that have been supplied. Some of those checks use the supplied
671 modulus value even if it has already been found to be too large.
673 A new limit has been added to DH_check of 32,768 bits. Supplying a
674 key/parameters with a modulus over this size will simply cause DH_check() to
681 * Do not ignore empty associated data entries with AES-SIV.
683 The AES-SIV algorithm allows for authentication of multiple associated
684 data entries along with the encryption. To authenticate empty data the
685 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
686 with NULL pointer as the output buffer and 0 as the input buffer length.
687 The AES-SIV implementation in OpenSSL just returns success for such call
688 instead of performing the associated data authentication operation.
689 The empty data thus will not be authenticated. ([CVE-2023-2975])
691 Thanks to Juerg Wullschleger (Google) for discovering the issue.
693 The fix changes the authentication tag value and the ciphertext for
694 applications that use empty associated data entries with AES-SIV.
695 To decrypt data encrypted with previous versions of OpenSSL the application
696 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
701 * When building with the `enable-fips` option and using the resulting
702 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
703 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
704 not operate with truncated digests (FIPS 140-3 IG G.R).
708 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
710 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
711 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
713 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
714 numeric text form. For gigantic sub-identifiers, this would take a very
715 long time, the time complexity being O(n^2) where n is the size of that
716 sub-identifier. ([CVE-2023-2650])
718 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
719 IDENTIFIER to canonical numeric text form if the size of that OBJECT
720 IDENTIFIER is 586 bytes or less, and fail otherwise.
722 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
723 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
724 most 128 sub-identifiers, and that the maximum value that each sub-
725 identifier may have is 2^32-1 (4294967295 decimal).
727 For each byte of every sub-identifier, only the 7 lower bits are part of
728 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
729 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
734 * Multiple algorithm implementation fixes for ARM BE platforms.
738 * Added a -pedantic option to fipsinstall that adjusts the various
739 settings to ensure strict FIPS compliance rather than backwards
744 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
745 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
746 trigger a crash of an application using AES-XTS decryption if the memory
747 just after the buffer being decrypted is not mapped.
748 Thanks to Anton Romanov (Amazon) for discovering the issue.
753 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
754 The previous fix for this timing side channel turned out to cause
755 a severe 2-3x performance regression in the typical use case
756 compared to 3.0.7. The new fix uses existing constant time
757 code paths, and restores the previous performance level while
758 fully eliminating all existing timing side channels.
759 The fix was developed by Bernd Edlinger with testing support
764 * Add FIPS provider configuration option to disallow the use of
765 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
766 The option '-no_drbg_truncated_digests' can optionally be
767 supplied to 'openssl fipsinstall'.
771 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
772 that it does not enable policy checking. Thanks to David Benjamin for
773 discovering this issue.
778 * Fixed an issue where invalid certificate policies in leaf certificates are
779 silently ignored by OpenSSL and other certificate policy checks are skipped
780 for that certificate. A malicious CA could use this to deliberately assert
781 invalid certificate policies in order to circumvent policy checking on the
782 certificate altogether.
787 * Limited the number of nodes created in a policy tree to mitigate
788 against CVE-2023-0464. The default limit is set to 1000 nodes, which
789 should be sufficient for most installations. If required, the limit
790 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
791 time define to a desired maximum number of nodes or zero to allow
797 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
799 * Add FIPS provider configuration option to enforce the
800 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
801 The option '-ems-check' can optionally be supplied to
802 'openssl fipsinstall'.
806 * The FIPS provider includes a few non-approved algorithms for
807 backward compatibility purposes and the "fips=yes" property query
808 must be used for all algorithm fetches to ensure FIPS compliance.
810 The algorithms that are included but not approved are Triple DES ECB,
811 Triple DES CBC and EdDSA.
815 * Added support for KMAC in KBKDF.
819 * RNDR and RNDRRS support in provider functions to provide
820 random number generation for Arm CPUs (aarch64).
824 * s_client and s_server apps now explicitly say when the TLS version
825 does not include the renegotiation mechanism. This avoids confusion
826 between that scenario versus when the TLS version includes secure
827 renegotiation but the peer lacks support for it.
831 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
833 *Tomasz Kantecki, Andrey Matyukov*
835 * The various OBJ_* functions have been made thread safe.
839 * Parallel dual-prime 1536/2048-bit modular exponentiation for
840 AVX512_IFMA capable processors.
842 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
844 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
845 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
846 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
847 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
848 `OPENSSL_NO_DEPRECATED_3_1`.
850 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
851 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
852 definitions for these functions regardless of whether
853 `OPENSSL_NO_DEPRECATED_3_1` is defined.
855 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
856 functions regardless of whether they are using them. It is recommended that
857 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
861 * When generating safe-prime DH parameters set the recommended private key
862 length equivalent to minimum key lengths as in RFC 7919.
866 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
867 maximum size that is smaller or equal to the digest length to comply with
868 FIPS 186-4 section 5. This is implemented by a new option
869 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
870 `rsa_pss_saltlen` parameter, which is now the default. Signature
871 verification is not affected by this change and continues to work as before.
878 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
879 listed here are only a brief description.
880 The migration guide contains more detailed information related to new features,
881 breaking changes, and mappings for the large list of deprecated functions.
883 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
885 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
887 * Fixed NULL dereference during PKCS7 data verification.
889 A NULL pointer can be dereferenced when signatures are being
890 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
891 algorithm used for the signature is known to the OpenSSL library but
892 the implementation of the hash algorithm is not available the digest
893 initialization will fail. There is a missing check for the return
894 value from the initialization function which later leads to invalid
895 usage of the digest API most likely leading to a crash.
898 PKCS7 data is processed by the SMIME library calls and also by the
899 time stamp (TS) library calls. The TLS implementation in OpenSSL does
900 not call these functions however third party applications would be
901 affected if they call these functions to verify signatures on untrusted
906 * Fixed X.400 address type confusion in X.509 GeneralName.
908 There is a type confusion vulnerability relating to X.400 address processing
909 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
910 but the public structure definition for GENERAL_NAME incorrectly specified
911 the type of the x400Address field as ASN1_TYPE. This field is subsequently
912 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
915 When CRL checking is enabled (i.e. the application sets the
916 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
917 pass arbitrary pointers to a memcmp call, enabling them to read memory
918 contents or enact a denial of service.
923 * Fixed NULL dereference validating DSA public key.
925 An invalid pointer dereference on read can be triggered when an
926 application tries to check a malformed DSA public key by the
927 EVP_PKEY_public_check() function. This will most likely lead
928 to an application crash. This function can be called on public
929 keys supplied from untrusted sources which could allow an attacker
930 to cause a denial of service attack.
932 The TLS implementation in OpenSSL does not call this function
933 but applications might call the function if there are additional
934 security requirements imposed by standards such as FIPS 140-3.
937 *Shane Lontis, Tomáš Mráz*
939 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
941 An invalid pointer dereference on read can be triggered when an
942 application tries to load malformed PKCS7 data with the
943 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
945 The result of the dereference is an application crash which could
946 lead to a denial of service attack. The TLS implementation in OpenSSL
947 does not call this function however third party applications might
948 call these functions on untrusted data.
953 * Fixed Use-after-free following BIO_new_NDEF.
955 The public API function BIO_new_NDEF is a helper function used for
956 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
957 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
958 be called directly by end user applications.
960 The function receives a BIO from the caller, prepends a new BIO_f_asn1
961 filter BIO onto the front of it to form a BIO chain, and then returns
962 the new head of the BIO chain to the caller. Under certain conditions,
963 for example if a CMS recipient public key is invalid, the new filter BIO
964 is freed and the function returns a NULL result indicating a failure.
965 However, in this case, the BIO chain is not properly cleaned up and the
966 BIO passed by the caller still retains internal pointers to the previously
967 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
968 then a use-after-free will occur. This will most likely result in a crash.
971 *Viktor Dukhovni, Matt Caswell*
973 * Fixed Double free after calling PEM_read_bio_ex.
975 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
976 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
977 data. If the function succeeds then the "name_out", "header" and "data"
978 arguments are populated with pointers to buffers containing the relevant
979 decoded data. The caller is responsible for freeing those buffers. It is
980 possible to construct a PEM file that results in 0 bytes of payload data.
981 In this case PEM_read_bio_ex() will return a failure code but will populate
982 the header argument with a pointer to a buffer that has already been freed.
983 If the caller also frees this buffer then a double free will occur. This
984 will most likely lead to a crash.
986 The functions PEM_read_bio() and PEM_read() are simple wrappers around
987 PEM_read_bio_ex() and therefore these functions are also directly affected.
989 These functions are also called indirectly by a number of other OpenSSL
990 functions including PEM_X509_INFO_read_bio_ex() and
991 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
992 internal uses of these functions are not vulnerable because the caller does
993 not free the header argument if PEM_read_bio_ex() returns a failure code.
996 *Kurt Roeckx, Matt Caswell*
998 * Fixed Timing Oracle in RSA Decryption.
1000 A timing based side channel exists in the OpenSSL RSA Decryption
1001 implementation which could be sufficient to recover a plaintext across
1002 a network in a Bleichenbacher style attack. To achieve a successful
1003 decryption an attacker would have to be able to send a very large number
1004 of trial messages for decryption. The vulnerability affects all RSA padding
1005 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1008 *Dmitry Belyavsky, Hubert Kario*
1010 * Fixed X.509 Name Constraints Read Buffer Overflow.
1012 A read buffer overrun can be triggered in X.509 certificate verification,
1013 specifically in name constraint checking. The read buffer overrun might
1014 result in a crash which could lead to a denial of service attack.
1015 In a TLS client, this can be triggered by connecting to a malicious
1016 server. In a TLS server, this can be triggered if the server requests
1017 client authentication and a malicious client connects.
1022 * Fixed X.509 Policy Constraints Double Locking security issue.
1024 If an X.509 certificate contains a malformed policy constraint and
1025 policy processing is enabled, then a write lock will be taken twice
1026 recursively. On some operating systems (most widely: Windows) this
1027 results in a denial of service when the affected process hangs. Policy
1028 processing being enabled on a publicly facing server is not considered
1029 to be a common setup.
1034 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1035 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1036 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1037 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1038 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1039 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1040 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1041 for legacy EC and SM2 keys is also changed similarly to honor the
1042 equivalent conversion format flag as specified in the underlying
1043 `EC_KEY` object being exported to a provider, when this function is
1044 called through `EVP_PKEY_export()`.
1048 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1050 * Fixed two buffer overflows in punycode decoding functions.
1052 A buffer overrun can be triggered in X.509 certificate verification,
1053 specifically in name constraint checking. Note that this occurs after
1054 certificate chain signature verification and requires either a CA to
1055 have signed the malicious certificate or for the application to continue
1056 certificate verification despite failure to construct a path to a trusted
1059 In a TLS client, this can be triggered by connecting to a malicious
1060 server. In a TLS server, this can be triggered if the server requests
1061 client authentication and a malicious client connects.
1063 An attacker can craft a malicious email address to overflow
1064 an arbitrary number of bytes containing the `.` character (decimal 46)
1065 on the stack. This buffer overflow could result in a crash (causing a
1069 An attacker can craft a malicious email address to overflow four
1070 attacker-controlled bytes on the stack. This buffer overflow could
1071 result in a crash (causing a denial of service) or potentially remote code
1072 execution depending on stack layout for any given platform/compiler.
1077 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1078 parameters in OpenSSL code.
1079 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1080 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1081 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1082 Using these invalid names may cause algorithms to use slower methods
1083 that ignore the CRT parameters.
1087 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1092 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1093 data to be signed before signing the certificate.
1097 * Added RIPEMD160 to the default provider.
1101 * Ensured that the key share group sent or accepted for the key exchange
1102 is allowed for the protocol version.
1106 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1108 * OpenSSL supports creating a custom cipher via the legacy
1109 EVP_CIPHER_meth_new() function and associated function calls. This function
1110 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1111 to use the new provider mechanism in order to implement custom ciphers.
1113 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1114 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1115 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1116 and decryption initialisation functions). Instead of using the custom cipher
1117 directly it incorrectly tries to fetch an equivalent cipher from the
1118 available providers. An equivalent cipher is found based on the NID passed to
1119 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1120 given cipher. However it is possible for an application to incorrectly pass
1121 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1122 is used in this way the OpenSSL encryption/decryption initialisation function
1123 will match the NULL cipher as being equivalent and will fetch this from the
1124 available providers. This will succeed if the default provider has been
1125 loaded (or if a third party provider has been loaded that offers this
1126 cipher). Using the NULL cipher means that the plaintext is emitted as the
1129 Applications are only affected by this issue if they call
1130 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1131 encryption/decryption initialisation function. Applications that only use
1132 SSL/TLS are not impacted by this issue.
1137 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1142 * Fixed the linux-mips64 Configure target which was missing the
1143 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1148 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1153 * Correctly handle a retransmitted ClientHello in DTLS
1157 * Fixed detection of ktls support in cross-compile environment on Linux
1161 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1166 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1167 report correct results in some cases
1171 * Fix UWP builds by defining VirtualLock
1175 * For known safe primes use the minimum key length according to RFC 7919.
1176 Longer private key sizes unnecessarily raise the cycles needed to compute the
1177 shared secret without any increase of the real security. This fixes a
1178 regression from 1.1.1 where these shorter keys were generated for the known
1183 * Added the loongarch64 target
1187 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1188 only passed to the FIPS provider and not to the default or legacy provider.
1192 * Fixed reported performance degradation on aarch64. Restored the
1193 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1194 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1195 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1196 The new algorithm is still used for 32 bit targets.
1200 * Added a missing header for memcmp that caused compilation failure on some
1205 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1207 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1208 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1209 This issue makes the RSA implementation with 2048 bit private keys
1210 incorrect on such machines and memory corruption will happen during
1211 the computation. As a consequence of the memory corruption an attacker
1212 may be able to trigger a remote code execution on the machine performing
1215 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1216 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1217 are affected by this issue.
1222 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1223 implementation would not encrypt the entirety of the data under some
1224 circumstances. This could reveal sixteen bytes of data that was
1225 preexisting in the memory that wasn't written. In the special case of
1226 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1228 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1229 they are both unaffected.
1232 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1234 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1236 * In addition to the c_rehash shell command injection identified in
1237 CVE-2022-1292, further bugs where the c_rehash script does not
1238 properly sanitise shell metacharacters to prevent command injection have been
1241 When the CVE-2022-1292 was fixed it was not discovered that there
1242 are other places in the script where the file names of certificates
1243 being hashed were possibly passed to a command executed through the shell.
1245 This script is distributed by some operating systems in a manner where
1246 it is automatically executed. On such operating systems, an attacker
1247 could execute arbitrary commands with the privileges of the script.
1249 Use of the c_rehash script is considered obsolete and should be replaced
1250 by the OpenSSL rehash command line tool.
1253 *Daniel Fiala, Tomáš Mráz*
1255 * Case insensitive string comparison no longer uses locales. It has instead
1256 been directly implemented.
1260 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1262 * Case insensitive string comparison is reimplemented via new locale-agnostic
1263 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1264 comparison. The previous implementation had problems when the Turkish locale
1269 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1270 metacharacters to prevent command injection. This script is distributed by
1271 some operating systems in a manner where it is automatically executed. On
1272 such operating systems, an attacker could execute arbitrary commands with the
1273 privileges of the script.
1275 Use of the c_rehash script is considered obsolete and should be replaced
1276 by the OpenSSL rehash command line tool.
1281 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1282 certificate on an OCSP response. The bug caused the function in the case
1283 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1284 response (meaning a successful verification) even in the case where the
1285 response signing certificate fails to verify.
1287 It is anticipated that most users of `OCSP_basic_verify` will not use the
1288 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1289 a negative value (indicating a fatal error) in the case of a certificate
1290 verification failure. The normal expected return value in this case would be
1293 This issue also impacts the command line OpenSSL "ocsp" application. When
1294 verifying an ocsp response with the "-no_cert_checks" option the command line
1295 application will report that the verification is successful even though it
1296 has in fact failed. In this case the incorrect successful response will also
1297 be accompanied by error messages showing the failure and contradicting the
1298 apparently successful result.
1303 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1304 AAD data as the MAC key. This made the MAC key trivially predictable.
1306 An attacker could exploit this issue by performing a man-in-the-middle attack
1307 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1308 that the modified data would still pass the MAC integrity check.
1310 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1311 endpoint will always be rejected by the recipient and the connection will
1312 fail at that point. Many application protocols require data to be sent from
1313 the client to the server first. Therefore, in such a case, only an OpenSSL
1314 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1316 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1317 sent in both directions. In this case both clients and servers could be
1318 affected, regardless of the application protocol.
1320 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1321 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1322 the handshake when using this ciphersuite.
1324 The confidentiality of data is not impacted by this issue, i.e. an attacker
1325 cannot decrypt data that has been encrypted using this ciphersuite - they can
1328 In order for this attack to work both endpoints must legitimately negotiate
1329 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1330 OpenSSL 3.0, and is not available within the default provider or the default
1331 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1332 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1333 following must have occurred:
1335 1) OpenSSL must have been compiled with the (non-default) compile time option
1336 enable-weak-ssl-ciphers
1338 2) OpenSSL must have had the legacy provider explicitly loaded (either
1339 through application code or via configuration)
1341 3) The ciphersuite must have been explicitly added to the ciphersuite list
1343 4) The libssl security level must have been set to 0 (default is 1)
1345 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1347 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1348 others that both endpoints have in common
1353 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1354 occupied by the removed hash table entries.
1356 This function is used when decoding certificates or keys. If a long lived
1357 process periodically decodes certificates or keys its memory usage will
1358 expand without bounds and the process might be terminated by the operating
1359 system causing a denial of service. Also traversing the empty hash table
1360 entries will take increasingly more time.
1362 Typically such long lived processes might be TLS clients or TLS servers
1363 configured to accept client certificate authentication.
1366 *Hugo Landau, Aliaksei Levin*
1368 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1369 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1370 statistics are no longer supported. For compatibility, these statistics are
1371 still listed in the output but are now always reported as zero.
1375 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1377 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1378 for non-prime moduli.
1380 Internally this function is used when parsing certificates that contain
1381 elliptic curve public keys in compressed form or explicit elliptic curve
1382 parameters with a base point encoded in compressed form.
1384 It is possible to trigger the infinite loop by crafting a certificate that
1385 has invalid explicit curve parameters.
1387 Since certificate parsing happens prior to verification of the certificate
1388 signature, any process that parses an externally supplied certificate may thus
1389 be subject to a denial of service attack. The infinite loop can also be
1390 reached when parsing crafted private keys as they can contain explicit
1391 elliptic curve parameters.
1393 Thus vulnerable situations include:
1395 - TLS clients consuming server certificates
1396 - TLS servers consuming client certificates
1397 - Hosting providers taking certificates or private keys from customers
1398 - Certificate authorities parsing certification requests from subscribers
1399 - Anything else which parses ASN.1 elliptic curve parameters
1401 Also any other applications that use the BN_mod_sqrt() where the attacker
1402 can control the parameter values are vulnerable to this DoS issue.
1407 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1408 to the list of ciphersuites providing Perfect Forward Secrecy as
1409 required by SECLEVEL >= 3.
1411 *Dmitry Belyavskiy, Nicola Tuveri*
1413 * Made the AES constant time code for no-asm configurations
1414 optional due to the resulting 95% performance degradation.
1415 The AES constant time code can be enabled, for no assembly
1416 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1420 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1425 * The negative return value handling of the certificate verification callback
1426 was reverted. The replacement is to set the verification retry state with
1427 the SSL_set_retry_verify() function.
1431 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1433 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1434 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1435 verify a certificate supplied by a server. That function may return a
1436 negative return value to indicate an internal error (for example out of
1437 memory). Such a negative return value is mishandled by OpenSSL and will cause
1438 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1439 success and a subsequent call to SSL_get_error() to return the value
1440 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1441 returned by OpenSSL if the application has previously called
1442 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1443 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1444 totally unexpected and applications may not behave correctly as a result. The
1445 exact behaviour will depend on the application but it could result in
1446 crashes, infinite loops or other similar incorrect responses.
1448 This issue is made more serious in combination with a separate bug in OpenSSL
1449 3.0 that will cause X509_verify_cert() to indicate an internal error when
1450 processing a certificate chain. This will occur where a certificate does not
1451 include the Subject Alternative Name extension but where a Certificate
1452 Authority has enforced name constraints. This issue can occur even with valid
1458 * Corrected a few file name and file reference bugs in the build,
1459 installation and setup scripts, which lead to installation verification
1460 failures. Slightly enhanced the installation verification script.
1464 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1469 * Fixed PVK encoder to properly query for the passphrase.
1473 * Multiple fixes in the OSSL_HTTP API functions.
1477 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1478 OSSL_PARAM_INTEGER data type and return error on negative numbers
1479 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1480 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1484 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1488 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1492 * Multiple threading fixes.
1496 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1500 * Allow fetching an operation from the provider that owns an unexportable key
1501 as a fallback if that is still allowed by the property query.
1505 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1507 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1512 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1513 S390X capability vector to zero. This simplifies testing of different code
1514 paths on S390X architecture.
1518 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1519 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1520 SP 800-38D". The communication will fail at this point.
1524 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1525 confidential in EC_GROUP data.
1529 * The byte order mark (BOM) character is ignored if encountered at the
1530 beginning of a PEM-formatted file.
1534 * Added CMS support for the Russian GOST algorithms.
1538 * Due to move of the implementation of cryptographic operations
1539 to the providers, validation of various operation parameters can
1540 be postponed until the actual operation is executed where previously
1541 it happened immediately when an operation parameter was set.
1543 For example when setting an unsupported curve with
1544 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1545 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1547 *OpenSSL team members and many third party contributors*
1549 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1550 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1551 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1552 instead to retrieve these algorithms from a provider.
1556 * On build targets where the multilib postfix is set in the build
1557 configuration the libdir directory was changing based on whether
1558 the lib directory with the multilib postfix exists on the system
1559 or not. This unpredictable behavior was removed and eventual
1560 multilib postfix is now always added to the default libdir. Use
1561 `--libdir=lib` to override the libdir if adding the postfix is
1566 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1567 no longer interoperable with OpenSSL 1.1.1.
1571 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1572 function codes, this function can only cause problems for calling
1577 * Add a configurable flag to output date formats as ISO 8601. Does not
1578 change the default date format.
1582 * Version of MSVC earlier than 1300 could get link warnings, which could
1583 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1584 Support for this flag has been removed.
1588 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1589 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1590 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1591 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1592 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1596 * The signatures of the functions to get and set options on SSL and
1597 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1598 Some source code changes may be required.
1602 * The public definitions of conf_method_st and conf_st have been
1603 deprecated. They will be made opaque in a future release.
1605 *Rich Salz and Tomáš Mráz*
1607 * Client-initiated renegotiation is disabled by default. To allow it, use
1608 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1609 flag, or the "ClientRenegotiation" config parameter as appropriate.
1613 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1614 or modify relative pathname inclusion.
1618 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1619 validated. Please consult the README-FIPS and
1620 README-PROVIDERS files, as well as the migration guide.
1622 *OpenSSL team members and many third party contributors*
1624 * For the key types DH and DHX the allowed settable parameters are now different.
1628 * The openssl commands that read keys, certificates, and CRLs now
1629 automatically detect the PEM or DER format of the input files.
1631 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1633 * Added enhanced PKCS#12 APIs which accept a library context.
1637 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1641 * Added support for Kernel TLS (KTLS).
1643 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1645 * Support for RFC 5746 secure renegotiation is now required by default for
1646 SSL or TLS connections to succeed.
1650 * The signature of the `copy` functional parameter of the
1651 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1652 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1653 the signature of the `pub_decode` functional parameter of the
1654 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1655 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1659 * The error return values from some control calls (ctrl) have changed.
1663 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1667 * Many functions in the EVP_ namespace that are getters of values from
1668 implementations or contexts were renamed to include get or get0 in their
1669 names. Old names are provided as macro aliases for compatibility and
1674 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1675 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1676 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1681 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1684 * The output from the command line applications may have minor
1689 * The output from numerous "printing" may have minor changes.
1693 * Windows thread synchronization uses read/write primitives (SRWLock) when
1694 supported by the OS, otherwise CriticalSection continues to be used.
1698 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1699 work on read only BIO source/sinks that do not support these functions.
1700 This allows piping or redirection of a file BIO using stdin to be buffered
1701 into memory. This is used internally in OSSL_DECODER_from_bio().
1705 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1706 this function would return one of the values OSSL_STORE_INFO_NAME,
1707 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1708 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1709 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1710 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1711 using this function should be amended to handle the changed return value.
1715 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1716 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1717 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1718 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1719 contains more than one certificate identifier: This means that all
1720 certificates referenced there MUST be part of the validation chain.
1724 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1725 RC5, DESX and DES have been moved to the legacy provider.
1729 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1730 RIPEMD-160 have been moved to the legacy provider.
1734 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1739 * The deprecated functions EVP_PKEY_get0_RSA(),
1740 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1741 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1742 well as the similarly named "get1" functions behave differently in
1747 * A number of functions handling low-level keys or engines were deprecated
1748 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1749 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1750 EVP_PKEY_get0_siphash().
1754 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1755 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1756 will need to load the legacy crypto provider. This includes these PBE
1757 algorithms which use this KDF:
1758 - NID_pbeWithMD2AndDES_CBC
1759 - NID_pbeWithMD5AndDES_CBC
1760 - NID_pbeWithSHA1AndRC2_CBC
1761 - NID_pbeWithMD2AndRC2_CBC
1762 - NID_pbeWithMD5AndRC2_CBC
1763 - NID_pbeWithSHA1AndDES_CBC
1767 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1768 BIO_debug_callback() functions.
1772 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1773 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1777 * The RAND_METHOD APIs have been deprecated.
1781 * The SRP APIs have been deprecated.
1785 * Add a compile time option to prevent the caching of provider fetched
1786 algorithms. This is enabled by including the no-cached-fetch option
1787 at configuration time.
1791 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1792 count of PKCS12_DEFAULT_ITER.
1794 *Tomáš Mráz and Sahana Prasad*
1796 * The openssl speed command does not use low-level API calls anymore.
1800 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1803 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1805 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1809 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1810 providers may supply their own group implementations (using either the "key
1811 exchange" or the "key encapsulation" methods) which will automatically be
1812 detected and used by libssl.
1814 *Matt Caswell, Nicola Tuveri*
1816 * The undocumented function X509_certificate_type() has been deprecated;
1820 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1824 * Removed RSA padding mode for SSLv23 (which was only used for
1825 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1826 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1831 * Deprecated the obsolete X9.31 RSA key generation related functions.
1833 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1834 is not allowed to return a value > 1, this is no more taken as failure.
1836 *Viktor Dukhovni and David von Oheimb*
1838 * Deprecated the obsolete X9.31 RSA key generation related functions
1839 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1840 BN_X931_generate_prime_ex().
1844 * The default key generation method for the regular 2-prime RSA keys was
1845 changed to the FIPS 186-4 B.3.6 method.
1849 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1853 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1857 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1858 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1860 *Rich Salz, Richard Levitte, and David von Oheimb*
1862 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1866 * Deprecated `OCSP_parse_url()`.
1870 * Validation of SM2 keys has been separated from the validation of regular EC
1875 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1876 switches: a validation failure triggers an early exit, returning a failure
1877 exit status to the parent process.
1881 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1882 to ignore unknown ciphers.
1886 * The `-cipher-commands` and `-digest-commands` options
1887 of the command line utility `list` have been deprecated.
1888 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1892 * Added convenience functions for generating asymmetric key pairs:
1893 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1894 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1898 * All of the low-level EC_KEY functions have been deprecated.
1900 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1902 * Deprecated all the libcrypto and libssl error string loading
1907 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1908 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1913 * The `-crypt` option to the `passwd` command line tool has been removed.
1917 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1922 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1926 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1927 EVP_PKEY_get1_tls_encodedpoint().
1931 * The security callback, which can be customised by application code, supports
1932 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1933 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1937 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1938 interface. Their functionality remains unchanged.
1942 * Added new option for 'openssl list', '-providers', which will display the
1943 list of loaded providers, their names, version and status. It optionally
1944 displays their gettable parameters.
1948 * Removed EVP_PKEY_set_alias_type().
1952 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1953 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1957 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1958 parameters are still checked for type safety at compile time via helper
1963 * Remove the RAND_DRBG API
1965 *Paul Dale and Matthias St. Pierre*
1967 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1968 as well as actual hostnames.
1972 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1973 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1974 conversely, silently ignore DTLS protocol version bounds when configuring
1975 TLS-based contexts. The commands can be repeated to set bounds of both
1976 types. The same applies with the corresponding "min_protocol" and
1977 "max_protocol" command-line switches, in case some application uses both TLS
1980 SSL_CTX instances that are created for a fixed protocol version (e.g.
1981 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1982 attempts to apply bounds to these protocol versions would result in an
1983 error. Now only the "version-flexible" SSL_CTX instances are subject to
1984 limits in configuration files in command-line options.
1988 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1993 * Reworked the recorded ERR codes to make better space for system errors.
1994 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1995 given code is a system error (true) or an OpenSSL error (false).
1999 * Reworked the test perl framework to better allow parallel testing.
2001 *Nicola Tuveri and David von Oheimb*
2003 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2004 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2008 * 'Configure' has been changed to figure out the configuration target if
2009 none is given on the command line. Consequently, the 'config' script is
2010 now only a mere wrapper. All documentation is changed to only mention
2013 *Rich Salz and Richard Levitte*
2015 * Added a library context `OSSL_LIB_CTX` that applications as well as
2016 other libraries can use to form a separate context within which
2017 libcrypto operations are performed.
2021 * Added various `_ex` functions to the OpenSSL API that support using
2022 a non-default `OSSL_LIB_CTX`.
2026 * Handshake now fails if Extended Master Secret extension is dropped
2031 * Dropped interactive mode from the `openssl` program.
2035 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2037 *David von Oheimb and Shane Lontis*
2039 * Deprecated `EC_METHOD_get_field_type()`.
2043 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2044 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2045 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2049 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2053 * Add CAdES-BES signature verification support, mostly derived
2054 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2056 *Filipe Raimundo da Silva*
2058 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2062 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2063 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2067 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2071 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2072 EC_KEY_precompute_mult().
2076 * Deprecated EC_POINTs_mul().
2080 * Removed FIPS_mode() and FIPS_mode_set().
2084 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2088 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2089 EC_POINT_get_Jprojective_coordinates_GFp().
2093 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2094 arrays to be more easily constructed via a series of utility functions.
2095 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2096 the various push functions and finally convert to a passable OSSL_PARAM
2097 array using OSSL_PARAM_BLD_to_param().
2101 * The security strength of SHA1 and MD5 based signatures in TLS has been
2106 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2107 contain a provider side internal key.
2111 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2115 * Project text documents not yet having a proper file name extension
2116 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2117 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2121 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2122 have been converted to Markdown with the goal to produce documents
2123 which not only look pretty when viewed online in the browser, but
2124 remain well readable inside a plain text editor.
2126 To achieve this goal, a 'minimalistic' Markdown style has been applied
2127 which avoids formatting elements that interfere too much with the
2128 reading flow in the text file. For example, it
2130 * avoids [ATX headings][] and uses [setext headings][] instead
2131 (which works for `<h1>` and `<h2>` headings only).
2132 * avoids [inline links][] and uses [reference links][] instead.
2133 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2135 [ATX headings]: https://github.github.com/gfm/#atx-headings
2136 [setext headings]: https://github.github.com/gfm/#setext-headings
2137 [inline links]: https://github.github.com/gfm/#inline-link
2138 [reference links]: https://github.github.com/gfm/#reference-link
2139 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2140 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2142 *Matthias St. Pierre*
2144 * The test suite is changed to preserve results of each test recipe.
2145 A new directory test-runs/ with subdirectories named like the
2146 test recipes are created in the build tree for this purpose.
2150 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2151 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2152 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2154 *David von Oheimb, Martin Peylo*
2156 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2157 It supports arbitrary request and response content types, GET redirection,
2158 TLS, connections via HTTP(S) proxies, connections and exchange via
2159 user-defined BIOs (allowing implicit connections), persistent connections,
2160 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2161 The legacy OCSP-focused (and only partly documented) API
2162 is retained for backward compatibility, while most of it is deprecated.
2166 * Added `util/check-format.pl`, a tool for checking adherence to the
2167 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2168 The checks performed are incomplete and yield some false positives.
2169 Still the tool should be useful for detecting most typical glitches.
2173 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2174 If domain name resolution yields multiple IP addresses all of them are tried
2175 after `connect()` failures.
2179 * All of the low-level RSA functions have been deprecated.
2183 * X509 certificates signed using SHA1 are no longer allowed at security
2188 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2189 modified to use PKEY APIs. These commands are now in maintenance mode
2190 and no new features will be added to them.
2194 * The command line utility rsautl has been deprecated.
2198 * The command line utilities genrsa and rsa have been modified to use PKEY
2199 APIs. They now write PKCS#8 keys by default. These commands are now in
2200 maintenance mode and no new features will be added to them.
2204 * All of the low-level DH functions have been deprecated.
2206 *Paul Dale and Matt Caswell*
2208 * All of the low-level DSA functions have been deprecated.
2212 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2213 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2217 * Deprecated low-level ECDH and ECDSA functions.
2221 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2225 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2226 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2227 a new formulation to include all the things it can be used for,
2228 as well as words of caution.
2232 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2236 * All of the low-level HMAC functions have been deprecated.
2238 *Paul Dale and David von Oheimb*
2240 * Over two thousand fixes were made to the documentation, including:
2241 - Common options (such as -rand/-writerand, TLS version control, etc)
2242 were refactored and point to newly-enhanced descriptions in openssl.pod.
2243 - Added style conformance for all options (with help from Richard Levitte),
2244 documented all reported missing options, added a CI build to check
2245 that all options are documented and that no unimplemented options
2247 - Documented some internals, such as all use of environment variables.
2248 - Addressed all internal broken L<> references.
2252 * All of the low-level CMAC functions have been deprecated.
2256 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2257 functions have been deprecated.
2259 *Paul Dale and David von Oheimb*
2261 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2262 set of functions. The documentation mentioned negative values for some
2263 errors, but this was never the case, so the mention of negative values
2266 Code that followed the documentation and thereby check with something
2267 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2271 * All of the low-level cipher functions have been deprecated.
2273 *Matt Caswell and Paul Dale*
2275 * Removed include/openssl/opensslconf.h.in and replaced it with
2276 include/openssl/configuration.h.in, which differs in not including
2277 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2278 was added to include both.
2280 This allows internal hacks where one might need to modify the set
2281 of configured macros, for example this if deprecated symbols are
2282 still supposed to be available internally:
2284 #include <openssl/configuration.h>
2286 #undef OPENSSL_NO_DEPRECATED
2287 #define OPENSSL_SUPPRESS_DEPRECATED
2289 #include <openssl/macros.h>
2291 This should not be used by applications that use the exported
2292 symbols, as that will lead to linking errors.
2296 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2297 used in exponentiation with 512-bit moduli. No EC algorithms are
2298 affected. Analysis suggests that attacks against 2-prime RSA1024,
2299 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2300 difficult to perform and are not believed likely. Attacks against DH512
2301 are considered just feasible. However, for an attack the target would
2302 have to reuse the DH512 private key, which is not recommended anyway.
2303 Also applications directly using the low-level API BN_mod_exp may be
2304 affected if they use BN_FLG_CONSTTIME.
2309 * Most memory-debug features have been deprecated, and the functionality
2310 replaced with no-ops.
2314 * Added documentation for the STACK API.
2318 * Introduced a new method type and API, OSSL_ENCODER, to represent
2319 generic encoders. These do the same sort of job that PEM writers
2320 and d2i functions do, but with support for methods supplied by
2321 providers, and the possibility for providers to support other
2326 * Introduced a new method type and API, OSSL_DECODER, to represent
2327 generic decoders. These do the same sort of job that PEM readers
2328 and i2d functions do, but with support for methods supplied by
2329 providers, and the possibility for providers to support other
2334 * Added a .pragma directive to the syntax of configuration files, to
2335 allow varying behavior in a supported and predictable manner.
2336 Currently added pragma:
2340 This allows dollar signs to be a keyword character unless it's
2341 followed by a opening brace or parenthesis. This is useful for
2342 platforms where dollar signs are commonly used in names, such as
2343 volume names and system directory names on VMS.
2347 * Added functionality to create an EVP_PKEY from user data.
2351 * Change the interpretation of the '--api' configuration option to
2352 mean that this is a desired API compatibility level with no
2353 further meaning. The previous interpretation, that this would
2354 also mean to remove all deprecated symbols up to and including
2355 the given version, no requires that 'no-deprecated' is also used
2356 in the configuration.
2358 When building applications, the desired API compatibility level
2359 can be set with the OPENSSL_API_COMPAT macro like before. For
2360 API compatibility version below 3.0, the old style numerical
2361 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2362 For version 3.0 and on, the value is expected to be the decimal
2363 value calculated from the major and minor version like this:
2365 MAJOR * 10000 + MINOR * 100
2369 -DOPENSSL_API_COMPAT=30000 For 3.0
2370 -DOPENSSL_API_COMPAT=30200 For 3.2
2372 To hide declarations that are deprecated up to and including the
2373 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2374 given when building the application as well.
2378 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2379 access to certificate and CRL stores via URIs and OSSL_STORE
2382 This adds the following functions:
2384 - X509_LOOKUP_store()
2385 - X509_STORE_load_file()
2386 - X509_STORE_load_path()
2387 - X509_STORE_load_store()
2388 - SSL_add_store_cert_subjects_to_stack()
2389 - SSL_CTX_set_default_verify_store()
2390 - SSL_CTX_load_verify_file()
2391 - SSL_CTX_load_verify_dir()
2392 - SSL_CTX_load_verify_store()
2396 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2397 The presence of this system service is determined at run-time.
2401 * Added functionality to create an EVP_PKEY context based on data
2402 for methods from providers. This takes an algorithm name and a
2403 property query string and simply stores them, with the intent
2404 that any operation that uses this context will use those strings
2405 to fetch the needed methods implicitly, thereby making the port
2406 of application written for pre-3.0 OpenSSL easier.
2410 * The undocumented function NCONF_WIN32() has been deprecated; for
2411 conversion details see the HISTORY section of doc/man5/config.pod
2415 * Introduced the new functions EVP_DigestSignInit_ex() and
2416 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2417 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2418 pages for further details.
2422 * Over two thousand fixes were made to the documentation, including:
2423 adding missing command flags, better style conformance, documentation
2426 *Rich Salz, Richard Levitte*
2428 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2429 X25519, X448, Ed25519 and Ed448.
2433 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2438 * Deprecated the public definition of `ERR_STATE` as well as the function
2439 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2444 * Added ERR functionality to give callers access to the stored function
2445 names that have replaced the older function code based functions.
2447 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2448 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2449 ERR_peek_error_all() and ERR_peek_last_error_all().
2451 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2452 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2453 ERR_func_error_string().
2457 * Extended testing to be verbose for failing tests only. The make variables
2458 VERBOSE_FAILURE or VF can be used to enable this:
2460 $ make VF=1 test # Unix
2461 $ mms /macro=(VF=1) test ! OpenVMS
2462 $ nmake VF=1 test # Windows
2466 * Added the `-copy_extensions` option to the `x509` command for use with
2467 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2468 all extensions in the request are copied to the certificate or vice versa.
2470 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2472 * Added the `-copy_extensions` option to the `req` command for use with
2473 `-x509`. When given with the `copy` or `copyall` argument,
2474 all extensions in the certification request are copied to the certificate.
2478 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2479 they generate are by default RFC 5280 compliant in the following sense:
2480 There is a subjectKeyIdentifier extension with a hash value of the public key
2481 and for not self-signed certs there is an authorityKeyIdentifier extension
2482 with a keyIdentifier field or issuer information identifying the signing key.
2483 This is done unless some configuration overrides the new default behavior,
2484 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2488 * Added several checks to `X509_verify_cert()` according to requirements in
2489 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2490 (which may be done by using the CLI option `-x509_strict`):
2491 * The basicConstraints of CA certificates must be marked critical.
2492 * CA certificates must explicitly include the keyUsage extension.
2493 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2494 * The issuer name of any certificate must not be empty.
2495 * The subject name of CA certs, certs with keyUsage crlSign,
2496 and certs without subjectAlternativeName must not be empty.
2497 * If a subjectAlternativeName extension is given it must not be empty.
2498 * The signatureAlgorithm field and the cert signature must be consistent.
2499 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2500 must not be marked critical.
2501 * The authorityKeyIdentifier must be given for X.509v3 certs
2502 unless they are self-signed.
2503 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2507 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2508 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2512 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2513 used even when parsing explicit parameters, when loading a encoded key
2514 or calling `EC_GROUP_new_from_ecpkparameters()`/
2515 `EC_GROUP_new_from_ecparameters()`.
2516 This prevents bypass of security hardening and performance gains,
2517 especially for curves with specialized EC_METHODs.
2518 By default, if a key encoded with explicit parameters is loaded and later
2519 encoded, the output is still encoded with explicit parameters, even if
2520 internally a "named" EC_GROUP is used for computation.
2524 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2525 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2526 NULL. After this change, only the cofactor parameter can be NULL. It also
2527 does some minimal sanity checks on the passed order.
2532 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2533 An attack is simple, if the first CMS_recipientInfo is valid but the
2534 second CMS_recipientInfo is chosen ciphertext. If the second
2535 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2536 encryption key will be replaced by garbage, and the message cannot be
2537 decoded, but if the RSA decryption fails, the correct encryption key is
2538 used and the recipient will not notice the attack.
2539 As a work around for this potential attack the length of the decrypted
2540 key must be equal to the cipher default key length, in case the
2541 certificate is not given and all recipientInfo are tried out.
2542 The old behaviour can be re-enabled in the CMS code by setting the
2543 CMS_DEBUG_DECRYPT flag.
2547 * Early start up entropy quality from the DEVRANDOM seed source has been
2548 improved for older Linux systems. The RAND subsystem will wait for
2549 /dev/random to be producing output before seeding from /dev/urandom.
2550 The seeded state is stored for future library initialisations using
2551 a system global shared memory segment. The shared memory identifier
2552 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2553 the desired value. The default identifier is 114.
2557 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2558 when primes for RSA keys are computed.
2559 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2560 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2561 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2562 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2563 This avoids possible fingerprinting of newly generated RSA modules.
2567 * Correct the extended master secret constant on EBCDIC systems. Without this
2568 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2569 negotiate EMS will fail. Unfortunately this also means that TLS connections
2570 between EBCDIC systems with this fix, and EBCDIC systems without this
2571 fix will fail if they negotiate EMS.
2575 * Changed the library initialisation so that the config file is now loaded
2576 by default. This was already the case for libssl. It now occurs for both
2577 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2578 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2582 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2583 where the former acts as a replacement for `ERR_put_error()`, and the
2584 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2585 `ERR_raise_data()` adds more flexibility by taking a format string and
2586 an arbitrary number of arguments following it, to be processed with
2591 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2592 to check if a named provider is loaded and available. When called, it
2593 will also activate all fallback providers if such are still present.
2597 * Enforce a minimum DH modulus size of 512 bits.
2601 * Changed DH parameters to generate the order q subgroup instead of 2q.
2602 Previously generated DH parameters are still accepted by DH_check
2603 but DH_generate_key works around that by clearing bit 0 of the
2604 private key for those. This avoids leaking bit 0 of the private key.
2608 * Significantly reduce secure memory usage by the randomness pools.
2612 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2617 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2618 algorithms. An implementation of a key exchange algorithm can be obtained
2619 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2620 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2621 the older EVP_PKEY_derive_init() function. See the man pages for the new
2622 functions for further details.
2626 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2630 * Removed the function names from error messages and deprecated the
2635 * Removed NextStep support and the macro OPENSSL_UNISTD
2639 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2640 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2641 Also removed "export var as function" capability; we do not export
2642 variables, only functions.
2646 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2647 an error and 1 indicating success. In previous versions of OpenSSL this
2648 was a void type. If a key was set longer than the maximum possible this
2653 * Support SM2 signing and verification schemes with X509 certificate.
2657 * Use SHA256 as the default digest for TS query in the `ts` app.
2661 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2665 * Default cipher lists/suites are now available via a function, the
2666 #defines are deprecated.
2670 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2671 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2672 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2676 * Join the directories crypto/x509 and crypto/x509v3
2680 * Added command 'openssl kdf' that uses the EVP_KDF API.
2684 * Added command 'openssl mac' that uses the EVP_MAC API.
2688 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2689 as default directories. Also added the command 'openssl info'
2690 for scripting purposes.
2694 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2699 * Add prediction resistance to the DRBG reseeding process.
2703 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2704 mandated by IEEE Std 1619-2018.
2708 * Added newline escaping functionality to a filename when using openssl dgst.
2709 This output format is to replicate the output format found in the `*sum`
2710 checksum programs. This aims to preserve backward compatibility.
2712 *Matt Eaton, Richard Levitte, and Paul Dale*
2714 * Removed the heartbeat message in DTLS feature, as it has very
2715 little usage and doesn't seem to fulfill a valuable purpose.
2716 The configuration option is now deprecated.
2720 * Changed the output of 'openssl {digestname} < file' to display the
2721 digest name in its output.
2725 * Added a new generic trace API which provides support for enabling
2726 instrumentation through trace output.
2728 *Richard Levitte & Matthias St. Pierre*
2730 * Added build tests for C++. These are generated files that only do one
2731 thing, to include one public OpenSSL head file each. This tests that
2732 the public header files can be usefully included in a C++ application.
2734 This test isn't enabled by default. It can be enabled with the option
2735 'enable-buildtest-c++'.
2739 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2743 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2747 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2751 * Added KMAC to EVP_MAC.
2755 * Added property based algorithm implementation selection framework to
2760 * Added SCA hardening for modular field inversion in EC_GROUP through
2761 a new dedicated field_inv() pointer in EC_METHOD.
2762 This also addresses a leakage affecting conversions from projective
2763 to affine coordinates.
2765 *Billy Bob Brumley, Nicola Tuveri*
2767 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2768 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2769 those algorithms that were already supported through the EVP_PKEY API
2770 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2771 and scrypt are now wrappers that call EVP_KDF.
2775 * Build devcrypto engine as a dynamic engine.
2777 *Eneas U de Queiroz*
2779 * Add keyed BLAKE2 to EVP_MAC.
2783 * Fix a bug in the computation of the endpoint-pair shared secret used
2784 by DTLS over SCTP. This breaks interoperability with older versions
2785 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2786 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2787 interoperability with such broken implementations. However, enabling
2788 this switch breaks interoperability with correct implementations.
2790 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2791 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2795 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2799 * Changed the license to the Apache License v2.0.
2803 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2805 - Major releases (indicated by incrementing the MAJOR release number)
2806 may introduce incompatible API/ABI changes.
2807 - Minor releases (indicated by incrementing the MINOR release number)
2808 may introduce new features but retain API/ABI compatibility.
2809 - Patch releases (indicated by incrementing the PATCH number)
2810 are intended for bug fixes and other improvements of existing
2811 features only (like improving performance or adding documentation)
2812 and retain API/ABI compatibility.
2816 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2820 * Remove the 'dist' target and add a tarball building script. The
2821 'dist' target has fallen out of use, and it shouldn't be
2822 necessary to configure just to create a source distribution.
2826 * Recreate the OS390-Unix config target. It no longer relies on a
2827 special script like it did for OpenSSL pre-1.1.0.
2831 * Instead of having the source directories listed in Configure, add
2832 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2837 * Add GMAC to EVP_MAC.
2841 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2845 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2846 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2847 to facilitate the continued use of MACs through raw private keys in
2848 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2852 * Deprecate ECDH_KDF_X9_62().
2856 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2857 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2858 are retained for backwards compatibility.
2862 * AES-XTS mode now enforces that its two keys are different to mitigate
2863 the attacked described in "Efficient Instantiations of Tweakable
2864 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2865 Details of this attack can be obtained from:
2866 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2870 * Rename the object files, i.e. give them other names than in previous
2871 versions. Their names now include the name of the final product, as
2872 well as its type mnemonic (bin, lib, shlib).
2876 * Added new option for 'openssl list', '-objects', which will display the
2877 list of built in objects, i.e. OIDs with names.
2881 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2882 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2887 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2888 improves application performance by removing data copies and providing
2889 applications with zero-copy system calls such as sendfile and splice.
2893 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2897 * `PKCS12_parse` now maintains the order of the parsed certificates
2898 when outputting them via `*ca` (rather than reversing it).
2902 * Deprecated pthread fork support methods.
2906 * Added support for FFDHE key exchange in TLS 1.3.
2910 * Added a new concept for OpenSSL plugability: providers. This
2911 functionality is designed to replace the ENGINE API and ENGINE
2912 implementations, and to be much more dynamic, allowing provider
2913 authors to introduce new algorithms among other things, as long as
2914 there's an API that supports the algorithm type.
2916 With this concept comes a new core API for interaction between
2917 libcrypto and provider implementations. Public libcrypto functions
2918 that want to use providers do so through this core API.
2920 The main documentation for this core API is found in
2921 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2922 refer to other manuals describing the API specific for supported
2923 algorithm types (also called operations).
2930 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2932 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2934 * Avoid loading of a dynamic engine twice.
2938 * Prioritise DANE TLSA issuer certs over peer certs
2942 * Fixed random API for MacOS prior to 10.12
2944 These MacOS versions don't support the CommonCrypto APIs
2948 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2950 * Fixed an SM2 Decryption Buffer Overflow.
2952 In order to decrypt SM2 encrypted data an application is expected to
2953 call the API function EVP_PKEY_decrypt(). Typically an application will
2954 call this function twice. The first time, on entry, the "out" parameter
2955 can be NULL and, on exit, the "outlen" parameter is populated with the
2956 buffer size required to hold the decrypted plaintext. The application
2957 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2958 again, but this time passing a non-NULL value for the "out" parameter.
2960 A bug in the implementation of the SM2 decryption code means that the
2961 calculation of the buffer size required to hold the plaintext returned
2962 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2963 size required by the second call. This can lead to a buffer overflow
2964 when EVP_PKEY_decrypt() is called by the application a second time with
2965 a buffer that is too small.
2967 A malicious attacker who is able present SM2 content for decryption to
2968 an application could cause attacker chosen data to overflow the buffer
2969 by up to a maximum of 62 bytes altering the contents of other data held
2970 after the buffer, possibly changing application behaviour or causing
2971 the application to crash. The location of the buffer is application
2972 dependent but is typically heap allocated.
2977 * Fixed various read buffer overruns processing ASN.1 strings
2979 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2980 structure which contains a buffer holding the string data and a field
2981 holding the buffer length. This contrasts with normal C strings which
2982 are represented as a buffer for the string data which is terminated
2983 with a NUL (0) byte.
2985 Although not a strict requirement, ASN.1 strings that are parsed using
2986 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2987 well as any string whose value has been set with the ASN1_STRING_set()
2988 function will additionally NUL terminate the byte array in the
2989 ASN1_STRING structure.
2991 However, it is possible for applications to directly construct valid
2992 ASN1_STRING structures which do not NUL terminate the byte array by
2993 directly setting the "data" and "length" fields in the ASN1_STRING
2994 array. This can also happen by using the ASN1_STRING_set0() function.
2996 Numerous OpenSSL functions that print ASN.1 data have been found to
2997 assume that the ASN1_STRING byte array will be NUL terminated, even
2998 though this is not guaranteed for strings that have been directly
2999 constructed. Where an application requests an ASN.1 structure to be
3000 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3001 been directly constructed by the application without NUL terminating
3002 the "data" field, then a read buffer overrun can occur.
3004 The same thing can also occur during name constraints processing
3005 of certificates (for example if a certificate has been directly
3006 constructed by the application instead of loading it via the OpenSSL
3007 parsing functions, and the certificate contains non NUL terminated
3008 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3009 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3011 If a malicious actor can cause an application to directly construct an
3012 ASN1_STRING and then process it through one of the affected OpenSSL
3013 functions then this issue could be hit. This might result in a crash
3014 (causing a Denial of Service attack). It could also result in the
3015 disclosure of private memory contents (such as private keys, or
3016 sensitive plaintext).
3021 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3023 * Fixed a problem with verifying a certificate chain when using the
3024 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3025 the certificates present in a certificate chain. It is not set by default.
3027 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3028 the chain that have explicitly encoded elliptic curve parameters was added
3029 as an additional strict check.
3031 An error in the implementation of this check meant that the result of a
3032 previous check to confirm that certificates in the chain are valid CA
3033 certificates was overwritten. This effectively bypasses the check
3034 that non-CA certificates must not be able to issue other certificates.
3036 If a "purpose" has been configured then there is a subsequent opportunity
3037 for checks that the certificate is a valid CA. All of the named "purpose"
3038 values implemented in libcrypto perform this check. Therefore, where
3039 a purpose is set the certificate chain will still be rejected even when the
3040 strict flag has been used. A purpose is set by default in libssl client and
3041 server certificate verification routines, but it can be overridden or
3042 removed by an application.
3044 In order to be affected, an application must explicitly set the
3045 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3046 for the certificate verification or, in the case of TLS client or server
3047 applications, override the default purpose.
3052 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3053 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3054 renegotiation ClientHello omits the signature_algorithms extension (where it
3055 was present in the initial ClientHello), but includes a
3056 signature_algorithms_cert extension then a NULL pointer dereference will
3057 result, leading to a crash and a denial of service attack.
3059 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3060 (which is the default configuration). OpenSSL TLS clients are not impacted by
3064 *Peter Kästle and Samuel Sapalski*
3066 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3068 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3069 create a unique hash value based on the issuer and serial number data
3070 contained within an X509 certificate. However, it was failing to correctly
3071 handle any errors that may occur while parsing the issuer field (which might
3072 occur if the issuer field is maliciously constructed). This may subsequently
3073 result in a NULL pointer deref and a crash leading to a potential denial of
3079 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3080 padding mode to correctly check for rollback attacks. This is considered a
3081 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3086 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3087 functions. Previously they could overflow the output length argument in some
3088 cases where the input length is close to the maximum permissible length for
3089 an integer on the platform. In such cases the return value from the function
3090 call would be 1 (indicating success), but the output length value would be
3091 negative. This could cause applications to behave incorrectly or crash.
3096 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3097 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3098 could be exploited in a side channel attack to recover the password. Since
3099 the attack is local host only this is outside of the current OpenSSL
3100 threat model and therefore no CVE is assigned.
3102 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3107 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3109 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3110 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3111 If an attacker can control both items being compared then this could lead
3112 to a possible denial of service attack. OpenSSL itself uses the
3113 GENERAL_NAME_cmp function for two purposes:
3114 1) Comparing CRL distribution point names between an available CRL and a
3115 CRL distribution point embedded in an X509 certificate
3116 2) When verifying that a timestamp response token signer matches the
3117 timestamp authority name (exposed via the API functions
3118 TS_RESP_verify_response and TS_RESP_verify_token)
3123 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3125 * Certificates with explicit curve parameters are now disallowed in
3126 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3130 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3131 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3132 conversely, silently ignore DTLS protocol version bounds when configuring
3133 TLS-based contexts. The commands can be repeated to set bounds of both
3134 types. The same applies with the corresponding "min_protocol" and
3135 "max_protocol" command-line switches, in case some application uses both TLS
3138 SSL_CTX instances that are created for a fixed protocol version (e.g.
3139 TLSv1_server_method()) also silently ignore version bounds. Previously
3140 attempts to apply bounds to these protocol versions would result in an
3141 error. Now only the "version-flexible" SSL_CTX instances are subject to
3142 limits in configuration files in command-line options.
3146 * Handshake now fails if Extended Master Secret extension is dropped
3151 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3153 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3155 * Fixed segmentation fault in SSL_check_chain()
3156 Server or client applications that call the SSL_check_chain() function
3157 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3158 dereference as a result of incorrect handling of the
3159 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3160 or unrecognised signature algorithm is received from the peer. This could
3161 be exploited by a malicious peer in a Denial of Service attack.
3166 * Added AES consttime code for no-asm configurations
3167 an optional constant time support for AES was added
3168 when building openssl for no-asm.
3169 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3170 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3171 At this time this feature is by default disabled.
3172 It will be enabled by default in 3.0.
3176 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3178 * Revert the change of EOF detection while reading in libssl to avoid
3179 regressions in applications depending on the current way of reporting
3180 the EOF. As the existing method is not fully accurate the change to
3181 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3182 branch and will be present in the 3.0 release.
3186 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3187 when primes for RSA keys are computed.
3188 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3189 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3190 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3191 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3192 This avoids possible fingerprinting of newly generated RSA modules.
3196 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3198 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3199 while reading in libssl then we would report an error back to the
3200 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3201 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3202 therefore give a hint as to what went wrong.
3206 * Check that ed25519 and ed448 are allowed by the security level. Previously
3207 signature algorithms not using an MD were not being checked that they were
3208 allowed by the security level.
3212 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3213 was not quite right. The behaviour was not consistent between resumption
3214 and normal handshakes, and also not quite consistent with historical
3215 behaviour. The behaviour in various scenarios has been clarified and
3216 it has been updated to make it match historical behaviour as closely as
3221 * *[VMS only]* The header files that the VMS compilers include automatically,
3222 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3223 that the C++ compiler doesn't understand. This is a shortcoming in the
3224 compiler, but can be worked around with `__cplusplus` guards.
3226 C++ applications that use OpenSSL libraries must be compiled using the
3227 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3228 functions. Otherwise, only functions with symbols of less than 31
3229 characters can be used, as the linker will not be able to successfully
3230 resolve symbols with longer names.
3234 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3235 The presence of this system service is determined at run-time.
3239 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3244 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3246 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3247 number generator (RNG). This was intended to include protection in the
3248 event of a fork() system call in order to ensure that the parent and child
3249 processes did not share the same RNG state. However, this protection was not
3250 being used in the default case.
3252 A partial mitigation for this issue is that the output from a high
3253 precision timer is mixed into the RNG state so the likelihood of a parent
3254 and child process sharing state is significantly reduced.
3256 If an application already calls OPENSSL_init_crypto() explicitly using
3257 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3260 *Matthias St. Pierre*
3262 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3263 used even when parsing explicit parameters, when loading a encoded key
3264 or calling `EC_GROUP_new_from_ecpkparameters()`/
3265 `EC_GROUP_new_from_ecparameters()`.
3266 This prevents bypass of security hardening and performance gains,
3267 especially for curves with specialized EC_METHODs.
3268 By default, if a key encoded with explicit parameters is loaded and later
3269 encoded, the output is still encoded with explicit parameters, even if
3270 internally a "named" EC_GROUP is used for computation.
3274 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3275 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3276 NULL. After this change, only the cofactor parameter can be NULL. It also
3277 does some minimal sanity checks on the passed order.
3282 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3283 An attack is simple, if the first CMS_recipientInfo is valid but the
3284 second CMS_recipientInfo is chosen ciphertext. If the second
3285 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3286 encryption key will be replaced by garbage, and the message cannot be
3287 decoded, but if the RSA decryption fails, the correct encryption key is
3288 used and the recipient will not notice the attack.
3289 As a work around for this potential attack the length of the decrypted
3290 key must be equal to the cipher default key length, in case the
3291 certificate is not given and all recipientInfo are tried out.
3292 The old behaviour can be re-enabled in the CMS code by setting the
3293 CMS_DEBUG_DECRYPT flag.
3298 * Early start up entropy quality from the DEVRANDOM seed source has been
3299 improved for older Linux systems. The RAND subsystem will wait for
3300 /dev/random to be producing output before seeding from /dev/urandom.
3301 The seeded state is stored for future library initialisations using
3302 a system global shared memory segment. The shared memory identifier
3303 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3304 the desired value. The default identifier is 114.
3308 * Correct the extended master secret constant on EBCDIC systems. Without this
3309 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3310 negotiate EMS will fail. Unfortunately this also means that TLS connections
3311 between EBCDIC systems with this fix, and EBCDIC systems without this
3312 fix will fail if they negotiate EMS.
3316 * Use Windows installation paths in the mingw builds
3318 Mingw isn't a POSIX environment per se, which means that Windows
3319 paths should be used for installation.
3324 * Changed DH_check to accept parameters with order q and 2q subgroups.
3325 With order 2q subgroups the bit 0 of the private key is not secret
3326 but DH_generate_key works around that by clearing bit 0 of the
3327 private key for those. This avoids leaking bit 0 of the private key.
3331 * Significantly reduce secure memory usage by the randomness pools.
3335 * Revert the DEVRANDOM_WAIT feature for Linux systems
3337 The DEVRANDOM_WAIT feature added a select() call to wait for the
3338 /dev/random device to become readable before reading from the
3339 /dev/urandom device.
3341 It turned out that this change had negative side effects on
3342 performance which were not acceptable. After some discussion it
3343 was decided to revert this feature and leave it up to the OS
3344 resp. the platform maintainer to ensure a proper initialization
3345 during early boot time.
3347 *Matthias St. Pierre*
3349 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3351 * Add build tests for C++. These are generated files that only do one
3352 thing, to include one public OpenSSL head file each. This tests that
3353 the public header files can be usefully included in a C++ application.
3355 This test isn't enabled by default. It can be enabled with the option
3356 'enable-buildtest-c++'.
3360 * Enable SHA3 pre-hashing for ECDSA and DSA.
3364 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3365 This changes the size when using the `genpkey` command when no size is given.
3366 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3367 generation commands to use 2048 bits by default.
3371 * Reorganize the manual pages to consistently have RETURN VALUES,
3372 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3373 util/fix-doc-nits accordingly.
3375 *Paul Yang, Joshua Lock*
3377 * Add the missing accessor EVP_PKEY_get0_engine()
3381 * Have commands like `s_client` and `s_server` output the signature scheme
3382 along with other cipher suite parameters when debugging.
3386 * Make OPENSSL_config() error agnostic again.
3390 * Do the error handling in RSA decryption constant time.
3394 * Prevent over long nonces in ChaCha20-Poly1305.
3396 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3397 for every encryption operation. RFC 7539 specifies that the nonce value
3398 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3399 and front pads the nonce with 0 bytes if it is less than 12
3400 bytes. However it also incorrectly allows a nonce to be set of up to 16
3401 bytes. In this case only the last 12 bytes are significant and any
3402 additional leading bytes are ignored.
3404 It is a requirement of using this cipher that nonce values are
3405 unique. Messages encrypted using a reused nonce value are susceptible to
3406 serious confidentiality and integrity attacks. If an application changes
3407 the default nonce length to be longer than 12 bytes and then makes a
3408 change to the leading bytes of the nonce expecting the new value to be a
3409 new unique nonce then such an application could inadvertently encrypt
3410 messages with a reused nonce.
3412 Additionally the ignored bytes in a long nonce are not covered by the
3413 integrity guarantee of this cipher. Any application that relies on the
3414 integrity of these ignored leading bytes of a long nonce may be further
3415 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3416 is safe because no such use sets such a long nonce value. However user
3417 applications that use this cipher directly and set a non-default nonce
3418 length to be longer than 12 bytes may be vulnerable.
3420 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3426 * Add DEVRANDOM_WAIT feature for Linux systems
3428 On older Linux systems where the getrandom() system call is not available,
3429 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3430 Contrary to getrandom(), the /dev/urandom device will not block during
3431 early boot when the kernel CSPRNG has not been seeded yet.
3433 To mitigate this known weakness, use select() to wait for /dev/random to
3434 become readable before reading from /dev/urandom.
3436 * Ensure that SM2 only uses SM3 as digest algorithm
3440 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3442 * Change the info callback signals for the start and end of a post-handshake
3443 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3444 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3445 confused by this and assume that a TLSv1.2 renegotiation has started. This
3446 can break KeyUpdate handling. Instead we no longer signal the start and end
3447 of a post handshake message exchange (although the messages themselves are
3448 still signalled). This could break some applications that were expecting
3449 the old signals. However without this KeyUpdate is not usable for many
3454 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3456 * Timing vulnerability in DSA signature generation
3458 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3459 timing side channel attack. An attacker could use variations in the signing
3460 algorithm to recover the private key.
3462 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3467 * Timing vulnerability in ECDSA signature generation
3469 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3470 timing side channel attack. An attacker could use variations in the signing
3471 algorithm to recover the private key.
3473 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3478 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3479 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3480 of two gigabytes and the error handling improved.
3482 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3483 categorized as a normal bug, not a security issue, because the DRBG reseeds
3484 automatically and is fully functional even without additional randomness
3485 provided by the application.
3487 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3489 * Add a new ClientHello callback. Provides a callback interface that gives
3490 the application the ability to adjust the nascent SSL object at the
3491 earliest stage of ClientHello processing, immediately after extensions have
3492 been collected but before they have been processed. In particular, this
3493 callback can adjust the supported TLS versions in response to the contents
3498 * Add SM2 base algorithm support.
3502 * s390x assembly pack: add (improved) hardware-support for the following
3503 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3504 aes-cfb/cfb8, aes-ecb.
3508 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3509 parameter is no longer accepted, as it leads to a corrupt table. NULL
3510 pem_str is reserved for alias entries only.
3514 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3515 step for prime curves. The new implementation is based on formulae from
3516 differential addition-and-doubling in homogeneous projective coordinates
3517 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3518 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3519 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3520 to work in projective coordinates.
3522 *Billy Bob Brumley, Nicola Tuveri*
3524 * Change generating and checking of primes so that the error rate of not
3525 being prime depends on the intended use based on the size of the input.
3526 For larger primes this will result in more rounds of Miller-Rabin.
3527 The maximal error rate for primes with more than 1080 bits is lowered
3530 *Kurt Roeckx, Annie Yousar*
3532 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3536 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3537 moving between systems, and to avoid confusion when a Windows build is
3538 done with mingw vs with MSVC. For POSIX installs, there's still a
3539 symlink or copy named 'tsget' to avoid that confusion as well.
3543 * Revert blinding in ECDSA sign and instead make problematic addition
3544 length-invariant. Switch even to fixed-length Montgomery multiplication.
3548 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3549 step for binary curves. The new implementation is based on formulae from
3550 differential addition-and-doubling in mixed Lopez-Dahab projective
3551 coordinates, modified to independently blind the operands.
3553 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3555 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3556 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3557 EC_METHODs to implement their own specialized "ladder step", to take
3558 advantage of more favorable coordinate systems or more efficient
3559 differential addition-and-doubling algorithms.
3561 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3563 * Modified the random device based seed sources to keep the relevant
3564 file descriptors open rather than reopening them on each access.
3565 This allows such sources to operate in a chroot() jail without
3566 the associated device nodes being available. This behaviour can be
3567 controlled using RAND_keep_random_devices_open().
3571 * Numerous side-channel attack mitigations have been applied. This may have
3572 performance impacts for some algorithms for the benefit of improved
3573 security. Specific changes are noted in this change log by their respective
3578 * AIX shared library support overhaul. Switch to AIX "natural" way of
3579 handling shared libraries, which means collecting shared objects of
3580 different versions and bitnesses in one common archive. This allows to
3581 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3582 doesn't affect the way 3rd party applications are linked, only how
3583 multi-version installation is managed.
3587 * Make ec_group_do_inverse_ord() more robust and available to other
3588 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3589 mitigations are applied to the fallback BN_mod_inverse().
3590 When using this function rather than BN_mod_inverse() directly, new
3591 EC cryptosystem implementations are then safer-by-default.
3595 * Add coordinate blinding for EC_POINT and implement projective
3596 coordinate blinding for generic prime curves as a countermeasure to
3597 chosen point SCA attacks.
3599 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3601 * Add blinding to ECDSA and DSA signatures to protect against side channel
3602 attacks discovered by Keegan Ryan (NCC Group).
3606 * Enforce checking in the `pkeyutl` command to ensure that the input
3607 length does not exceed the maximum supported digest length when performing
3608 a sign, verify or verifyrecover operation.
3612 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3613 I/O in combination with something like select() or poll() will hang. This
3614 can be turned off again using SSL_CTX_clear_mode().
3615 Many applications do not properly handle non-application data records, and
3616 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3617 around the problems in those applications, but can also break some.
3618 It's recommended to read the manpages about SSL_read(), SSL_write(),
3619 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3620 SSL_CTX_set_read_ahead() again.
3624 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3625 now allow empty (zero character) pass phrases.
3629 * Apply blinding to binary field modular inversion and remove patent
3630 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3634 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3635 binary and prime elliptic curves.
3639 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3640 constant time fixed point multiplication.
3644 * Revise elliptic curve scalar multiplication with timing attack
3645 defenses: ec_wNAF_mul redirects to a constant time implementation
3646 when computing fixed point and variable point multiplication (which
3647 in OpenSSL are mostly used with secret scalars in keygen, sign,
3648 ECDH derive operations).
3649 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3652 * Updated CONTRIBUTING
3656 * Updated DRBG / RAND to request nonce and additional low entropy
3657 randomness from the system.
3659 *Matthias St. Pierre*
3661 * Updated 'openssl rehash' to use OpenSSL consistent default.
3665 * Moved the load of the ssl_conf module to libcrypto, which helps
3666 loading engines that libssl uses before libssl is initialised.
3670 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3674 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3676 *Ingo Schwarze, Rich Salz*
3678 * Added output of accepting IP address and port for 'openssl s_server'
3682 * Added a new API for TLSv1.3 ciphersuites:
3683 SSL_CTX_set_ciphersuites()
3684 SSL_set_ciphersuites()
3688 * Memory allocation failures consistently add an error to the error
3693 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3694 in libcrypto when run as setuid/setgid.
3698 * Load any config file by default when libssl is used.
3702 * Added new public header file <openssl/rand_drbg.h> and documentation
3703 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3705 *Matthias St. Pierre*
3707 * QNX support removed (cannot find contributors to get their approval
3708 for the license change).
3712 * TLSv1.3 replay protection for early data has been implemented. See the
3713 SSL_read_early_data() man page for further details.
3717 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3718 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3719 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3720 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3721 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3722 configuration has been separated out. See the ciphers man page or the
3723 SSL_CTX_set_ciphersuites() man page for more information.
3727 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3728 in responder mode now supports the new "-multi" option, which
3729 spawns the specified number of child processes to handle OCSP
3730 requests. The "-timeout" option now also limits the OCSP
3731 responder's patience to wait to receive the full client request
3732 on a newly accepted connection. Child processes are respawned
3733 as needed, and the CA index file is automatically reloaded
3734 when changed. This makes it possible to run the "ocsp" responder
3735 as a long-running service, making the OpenSSL CA somewhat more
3736 feature-complete. In this mode, most diagnostic messages logged
3737 after entering the event loop are logged via syslog(3) rather than
3742 * Added support for X448 and Ed448. Heavily based on original work by
3747 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3748 objects loaded. This adds the functions OSSL_STORE_expect() and
3749 OSSL_STORE_find() as well as needed tools to construct searches and
3750 get the search data out of them.
3754 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3755 version of OpenSSL should review their configuration settings to ensure
3756 that they are still appropriate for TLSv1.3. For further information see:
3757 <https://wiki.openssl.org/index.php/TLS1.3>
3761 * Grand redesign of the OpenSSL random generator
3763 The default RAND method now utilizes an AES-CTR DRBG according to
3764 NIST standard SP 800-90Ar1. The new random generator is essentially
3765 a port of the default random generator from the OpenSSL FIPS 2.0
3766 object module. It is a hybrid deterministic random bit generator
3767 using an AES-CTR bit stream and which seeds and reseeds itself
3768 automatically using trusted system entropy sources.
3770 Some of its new features are:
3771 - Support for multiple DRBG instances with seed chaining.
3772 - The default RAND method makes use of a DRBG.
3773 - There is a public and private DRBG instance.
3774 - The DRBG instances are fork-safe.
3775 - Keep all global DRBG instances on the secure heap if it is enabled.
3776 - The public and private DRBG instance are per thread for lock free
3779 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3781 * Changed Configure so it only says what it does and doesn't dump
3782 so much data. Instead, ./configdata.pm should be used as a script
3783 to display all sorts of configuration data.
3787 * Added processing of "make variables" to Configure.
3791 * Added SHA512/224 and SHA512/256 algorithm support.
3795 * The last traces of Netware support, first removed in 1.1.0, have
3800 * Get rid of Makefile.shared, and in the process, make the processing
3801 of certain files (rc.obj, or the .def/.map/.opt files produced from
3802 the ordinal files) more visible and hopefully easier to trace and
3803 debug (or make silent).
3807 * Make it possible to have environment variable assignments as
3808 arguments to config / Configure.
3812 * Add multi-prime RSA (RFC 8017) support.
3816 * Add SM3 implemented according to GB/T 32905-2016
3817 *Jack Lloyd <jack.lloyd@ribose.com>,*
3818 *Ronald Tse <ronald.tse@ribose.com>,*
3819 *Erick Borsboom <erick.borsboom@ribose.com>*
3821 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3822 as documented in RFC6066.
3823 Based on a patch from Tomasz Moń
3825 *Filipe Raimundo da Silva*
3827 * Add SM4 implemented according to GB/T 32907-2016.
3828 *Jack Lloyd <jack.lloyd@ribose.com>,*
3829 *Ronald Tse <ronald.tse@ribose.com>,*
3830 *Erick Borsboom <erick.borsboom@ribose.com>*
3832 * Reimplement -newreq-nodes and ERR_error_string_n; the
3833 original author does not agree with the license change.
3837 * Add ARIA AEAD TLS support.
3841 * Some macro definitions to support VS6 have been removed. Visual
3842 Studio 6 has not worked since 1.1.0
3846 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3847 without clearing the errors.
3851 * Add "atfork" functions. If building on a system that without
3852 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3853 requirements. The RAND facility now uses/requires this.
3861 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3862 not possible to disable entirely. However, it's still possible to
3863 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3866 To disable, configure with 'no-ui-console'. 'no-ui' is still
3867 possible to use as an alias. Check at compile time with the
3868 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3869 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3873 * Add a STORE module, which implements a uniform and URI based reader of
3874 stores that can contain keys, certificates, CRLs and numerous other
3875 objects. The main API is loosely based on a few stdio functions,
3876 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3877 OSSL_STORE_error and OSSL_STORE_close.
3878 The implementation uses backends called "loaders" to implement arbitrary
3879 URI schemes. There is one built in "loader" for the 'file' scheme.
3883 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3884 then adjusted to work on FreeBSD 8.4 as well.
3885 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3886 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3890 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3891 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3892 error code calls like this:
3894 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3896 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3897 that can be encoded in C. For the foreseeable future, this will only
3900 *Richard Levitte and Tim Hudson*
3902 * Removed BSD cryptodev engine.
3906 * Add a build target 'build_all_generated', to build all generated files
3907 and only that. This can be used to prepare everything that requires
3908 things like perl for a system that lacks perl and then move everything
3909 to that system and do the rest of the build there.
3913 * In the UI interface, make it possible to duplicate the user data. This
3914 can be used by engines that need to retain the data for a longer time
3915 than just the call where this user data is passed.
3919 * Ignore the '-named_curve auto' value for compatibility of applications
3922 *Tomáš Mráz <tmraz@fedoraproject.org>*
3924 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3925 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3926 alerts across multiple records (some of which could be empty). In practice
3927 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3928 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3929 support this at all. Supporting it adds significant complexity to the
3930 record layer, and its removal is unlikely to cause interoperability
3935 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3936 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3937 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3942 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3943 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3945 *Richard Levitte, Andy Polyakov*
3947 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3952 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3953 platform rather than 'mingw'.
3957 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3958 success if they are asked to add an object which already exists
3959 in the store. This change cascades to other functions which load
3960 certificates and CRLs.
3964 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3965 facilitate stack unwinding even from assembly subroutines.
3969 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3970 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3974 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3975 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3976 which is the minimum version we support.
3980 * Certificate time validation (X509_cmp_time) enforces stricter
3981 compliance with RFC 5280. Fractional seconds and timezone offsets
3982 are no longer allowed.
3986 * Add support for ARIA
3990 * s_client will now send the Server Name Indication (SNI) extension by
3991 default unless the new "-noservername" option is used. The server name is
3992 based on the host provided to the "-connect" option unless overridden by
3993 using "-servername".
3997 * Add support for SipHash
4001 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4002 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4003 prevent issues where no progress is being made and the peer continually
4004 sends unrecognised record types, using up resources processing them.
4008 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4009 using the algorithm defined in
4010 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4014 * Heartbeat support has been removed; the ABI is changed for now.
4016 *Richard Levitte, Rich Salz*
4018 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4022 * The RSA "null" method, which was partially supported to avoid patent
4023 issues, has been replaced to always returns NULL.
4030 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4032 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4033 used even when parsing explicit parameters, when loading a encoded key
4034 or calling `EC_GROUP_new_from_ecpkparameters()`/
4035 `EC_GROUP_new_from_ecparameters()`.
4036 This prevents bypass of security hardening and performance gains,
4037 especially for curves with specialized EC_METHODs.
4038 By default, if a key encoded with explicit parameters is loaded and later
4039 encoded, the output is still encoded with explicit parameters, even if
4040 internally a "named" EC_GROUP is used for computation.
4044 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4045 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4046 NULL. After this change, only the cofactor parameter can be NULL. It also
4047 does some minimal sanity checks on the passed order.
4052 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4053 An attack is simple, if the first CMS_recipientInfo is valid but the
4054 second CMS_recipientInfo is chosen ciphertext. If the second
4055 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4056 encryption key will be replaced by garbage, and the message cannot be
4057 decoded, but if the RSA decryption fails, the correct encryption key is
4058 used and the recipient will not notice the attack.
4059 As a work around for this potential attack the length of the decrypted
4060 key must be equal to the cipher default key length, in case the
4061 certificate is not given and all recipientInfo are tried out.
4062 The old behaviour can be re-enabled in the CMS code by setting the
4063 CMS_DEBUG_DECRYPT flag.
4068 * Use Windows installation paths in the mingw builds
4070 Mingw isn't a POSIX environment per se, which means that Windows
4071 paths should be used for installation.
4076 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4078 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4079 This changes the size when using the `genpkey` command when no size is given.
4080 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4081 generation commands to use 2048 bits by default.
4085 * Prevent over long nonces in ChaCha20-Poly1305.
4087 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4088 for every encryption operation. RFC 7539 specifies that the nonce value
4089 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4090 and front pads the nonce with 0 bytes if it is less than 12
4091 bytes. However it also incorrectly allows a nonce to be set of up to 16
4092 bytes. In this case only the last 12 bytes are significant and any
4093 additional leading bytes are ignored.
4095 It is a requirement of using this cipher that nonce values are
4096 unique. Messages encrypted using a reused nonce value are susceptible to
4097 serious confidentiality and integrity attacks. If an application changes
4098 the default nonce length to be longer than 12 bytes and then makes a
4099 change to the leading bytes of the nonce expecting the new value to be a
4100 new unique nonce then such an application could inadvertently encrypt
4101 messages with a reused nonce.
4103 Additionally the ignored bytes in a long nonce are not covered by the
4104 integrity guarantee of this cipher. Any application that relies on the
4105 integrity of these ignored leading bytes of a long nonce may be further
4106 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4107 is safe because no such use sets such a long nonce value. However user
4108 applications that use this cipher directly and set a non-default nonce
4109 length to be longer than 12 bytes may be vulnerable.
4111 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4117 * Added SCA hardening for modular field inversion in EC_GROUP through
4118 a new dedicated field_inv() pointer in EC_METHOD.
4119 This also addresses a leakage affecting conversions from projective
4120 to affine coordinates.
4122 *Billy Bob Brumley, Nicola Tuveri*
4124 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4125 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4129 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4133 * Remove the 'dist' target and add a tarball building script. The
4134 'dist' target has fallen out of use, and it shouldn't be
4135 necessary to configure just to create a source distribution.
4139 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4141 * Timing vulnerability in DSA signature generation
4143 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4144 timing side channel attack. An attacker could use variations in the signing
4145 algorithm to recover the private key.
4147 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4152 * Timing vulnerability in ECDSA signature generation
4154 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4155 timing side channel attack. An attacker could use variations in the signing
4156 algorithm to recover the private key.
4158 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4163 * Add coordinate blinding for EC_POINT and implement projective
4164 coordinate blinding for generic prime curves as a countermeasure to
4165 chosen point SCA attacks.
4167 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4169 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4171 * Client DoS due to large DH parameter
4173 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4174 malicious server can send a very large prime value to the client. This will
4175 cause the client to spend an unreasonably long period of time generating a
4176 key for this prime resulting in a hang until the client has finished. This
4177 could be exploited in a Denial Of Service attack.
4179 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4184 * Cache timing vulnerability in RSA Key Generation
4186 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4187 a cache timing side channel attack. An attacker with sufficient access to
4188 mount cache timing attacks during the RSA key generation process could
4189 recover the private key.
4191 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4192 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4197 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4198 parameter is no longer accepted, as it leads to a corrupt table. NULL
4199 pem_str is reserved for alias entries only.
4203 * Revert blinding in ECDSA sign and instead make problematic addition
4204 length-invariant. Switch even to fixed-length Montgomery multiplication.
4208 * Change generating and checking of primes so that the error rate of not
4209 being prime depends on the intended use based on the size of the input.
4210 For larger primes this will result in more rounds of Miller-Rabin.
4211 The maximal error rate for primes with more than 1080 bits is lowered
4214 *Kurt Roeckx, Annie Yousar*
4216 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4220 * Add blinding to ECDSA and DSA signatures to protect against side channel
4221 attacks discovered by Keegan Ryan (NCC Group).
4225 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4226 now allow empty (zero character) pass phrases.
4230 * Certificate time validation (X509_cmp_time) enforces stricter
4231 compliance with RFC 5280. Fractional seconds and timezone offsets
4232 are no longer allowed.
4236 * Fixed a text canonicalisation bug in CMS
4238 Where a CMS detached signature is used with text content the text goes
4239 through a canonicalisation process first prior to signing or verifying a
4240 signature. This process strips trailing space at the end of lines, converts
4241 line terminators to CRLF and removes additional trailing line terminators
4242 at the end of a file. A bug in the canonicalisation process meant that
4243 some characters, such as form-feed, were incorrectly treated as whitespace
4244 and removed. This is contrary to the specification (RFC5485). This fix
4245 could mean that detached text data signed with an earlier version of
4246 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4247 signed with a fixed OpenSSL may fail to verify with an earlier version of
4248 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4249 and use the "-binary" flag (for the "cms" command line application) or set
4250 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4254 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4256 * Constructed ASN.1 types with a recursive definition could exceed the stack
4258 Constructed ASN.1 types with a recursive definition (such as can be found
4259 in PKCS7) could eventually exceed the stack given malicious input with
4260 excessive recursion. This could result in a Denial Of Service attack. There
4261 are no such structures used within SSL/TLS that come from untrusted sources
4262 so this is considered safe.
4264 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4270 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4272 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4273 effectively reduced to only comparing the least significant bit of each
4274 byte. This allows an attacker to forge messages that would be considered as
4275 authenticated in an amount of tries lower than that guaranteed by the
4276 security claims of the scheme. The module can only be compiled by the
4277 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4279 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4285 * Add a build target 'build_all_generated', to build all generated files
4286 and only that. This can be used to prepare everything that requires
4287 things like perl for a system that lacks perl and then move everything
4288 to that system and do the rest of the build there.
4292 * Backport SSL_OP_NO_RENGOTIATION
4294 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4295 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4296 changes this is no longer possible in 1.1.0. Therefore, the new
4297 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4298 1.1.0 to provide equivalent functionality.
4300 Note that if an application built against 1.1.0h headers (or above) is run
4301 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4302 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4306 * Removed the OS390-Unix config target. It relied on a script that doesn't
4311 * rsaz_1024_mul_avx2 overflow bug on x86_64
4313 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4314 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4315 Analysis suggests that attacks against RSA and DSA as a result of this
4316 defect would be very difficult to perform and are not believed likely.
4317 Attacks against DH1024 are considered just feasible, because most of the
4318 work necessary to deduce information about a private key may be performed
4319 offline. The amount of resources required for such an attack would be
4320 significant. However, for an attack on TLS to be meaningful, the server
4321 would have to share the DH1024 private key among multiple clients, which is
4322 no longer an option since CVE-2016-0701.
4324 This only affects processors that support the AVX2 but not ADX extensions
4325 like Intel Haswell (4th generation).
4327 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4328 was originally found via the OSS-Fuzz project.
4333 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4335 * bn_sqrx8x_internal carry bug on x86_64
4337 There is a carry propagating bug in the x86_64 Montgomery squaring
4338 procedure. No EC algorithms are affected. Analysis suggests that attacks
4339 against RSA and DSA as a result of this defect would be very difficult to
4340 perform and are not believed likely. Attacks against DH are considered just
4341 feasible (although very difficult) because most of the work necessary to
4342 deduce information about a private key may be performed offline. The amount
4343 of resources required for such an attack would be very significant and
4344 likely only accessible to a limited number of attackers. An attacker would
4345 additionally need online access to an unpatched system using the target
4346 private key in a scenario with persistent DH parameters and a private
4347 key that is shared between multiple clients.
4349 This only affects processors that support the BMI1, BMI2 and ADX extensions
4350 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4352 This issue was reported to OpenSSL by the OSS-Fuzz project.
4357 * Malformed X.509 IPAddressFamily could cause OOB read
4359 If an X.509 certificate has a malformed IPAddressFamily extension,
4360 OpenSSL could do a one-byte buffer overread. The most likely result
4361 would be an erroneous display of the certificate in text format.
4363 This issue was reported to OpenSSL by the OSS-Fuzz project.
4368 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4370 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4371 platform rather than 'mingw'.
4375 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4376 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4377 which is the minimum version we support.
4381 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4383 * Encrypt-Then-Mac renegotiation crash
4385 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4386 negotiated where it was not in the original handshake (or vice-versa) then
4387 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4388 and servers are affected.
4390 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4395 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4397 * Truncated packet could crash via OOB read
4399 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4400 cipher is being used, then a truncated packet can cause that host to
4401 perform an out-of-bounds read, usually resulting in a crash.
4403 This issue was reported to OpenSSL by Robert Święcki of Google.
4408 * Bad (EC)DHE parameters cause a client crash
4410 If a malicious server supplies bad parameters for a DHE or ECDHE key
4411 exchange then this can result in the client attempting to dereference a
4412 NULL pointer leading to a client crash. This could be exploited in a Denial
4415 This issue was reported to OpenSSL by Guido Vranken.
4420 * BN_mod_exp may produce incorrect results on x86_64
4422 There is a carry propagating bug in the x86_64 Montgomery squaring
4423 procedure. No EC algorithms are affected. Analysis suggests that attacks
4424 against RSA and DSA as a result of this defect would be very difficult to
4425 perform and are not believed likely. Attacks against DH are considered just
4426 feasible (although very difficult) because most of the work necessary to
4427 deduce information about a private key may be performed offline. The amount
4428 of resources required for such an attack would be very significant and
4429 likely only accessible to a limited number of attackers. An attacker would
4430 additionally need online access to an unpatched system using the target
4431 private key in a scenario with persistent DH parameters and a private
4432 key that is shared between multiple clients. For example this can occur by
4433 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4434 similar to CVE-2015-3193 but must be treated as a separate problem.
4436 This issue was reported to OpenSSL by the OSS-Fuzz project.
4441 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4443 * ChaCha20/Poly1305 heap-buffer-overflow
4445 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4446 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4447 crash. This issue is not considered to be exploitable beyond a DoS.
4449 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4454 * CMS Null dereference
4456 Applications parsing invalid CMS structures can crash with a NULL pointer
4457 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4458 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4459 structure callback if an attempt is made to free certain invalid encodings.
4460 Only CHOICE structures using a callback which do not handle NULL value are
4463 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4468 * Montgomery multiplication may produce incorrect results
4470 There is a carry propagating bug in the Broadwell-specific Montgomery
4471 multiplication procedure that handles input lengths divisible by, but
4472 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4473 and DH private keys are impossible. This is because the subroutine in
4474 question is not used in operations with the private key itself and an input
4475 of the attacker's direct choice. Otherwise the bug can manifest itself as
4476 transient authentication and key negotiation failures or reproducible
4477 erroneous outcome of public-key operations with specially crafted input.
4478 Among EC algorithms only Brainpool P-512 curves are affected and one
4479 presumably can attack ECDH key negotiation. Impact was not analyzed in
4480 detail, because pre-requisites for attack are considered unlikely. Namely
4481 multiple clients have to choose the curve in question and the server has to
4482 share the private key among them, neither of which is default behaviour.
4483 Even then only clients that chose the curve will be affected.
4485 This issue was publicly reported as transient failures and was not
4486 initially recognized as a security issue. Thanks to Richard Morgan for
4487 providing reproducible case.
4492 * Removed automatic addition of RPATH in shared libraries and executables,
4493 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4497 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4499 * Fix Use After Free for large message sizes
4501 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4502 message larger than approx 16k is received then the underlying buffer to
4503 store the incoming message is reallocated and moved. Unfortunately a
4504 dangling pointer to the old location is left which results in an attempt to
4505 write to the previously freed location. This is likely to result in a
4506 crash, however it could potentially lead to execution of arbitrary code.
4508 This issue only affects OpenSSL 1.1.0a.
4510 This issue was reported to OpenSSL by Robert Święcki.
4515 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4517 * OCSP Status Request extension unbounded memory growth
4519 A malicious client can send an excessively large OCSP Status Request
4520 extension. If that client continually requests renegotiation, sending a
4521 large OCSP Status Request extension each time, then there will be unbounded
4522 memory growth on the server. This will eventually lead to a Denial Of
4523 Service attack through memory exhaustion. Servers with a default
4524 configuration are vulnerable even if they do not support OCSP. Builds using
4525 the "no-ocsp" build time option are not affected.
4527 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4532 * SSL_peek() hang on empty record
4534 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4535 sends an empty record. This could be exploited by a malicious peer in a
4536 Denial Of Service attack.
4538 This issue was reported to OpenSSL by Alex Gaynor.
4543 * Excessive allocation of memory in tls_get_message_header() and
4544 dtls1_preprocess_fragment()
4546 A (D)TLS message includes 3 bytes for its length in the header for the
4547 message. This would allow for messages up to 16Mb in length. Messages of
4548 this length are excessive and OpenSSL includes a check to ensure that a
4549 peer is sending reasonably sized messages in order to avoid too much memory
4550 being consumed to service a connection. A flaw in the logic of version
4551 1.1.0 means that memory for the message is allocated too early, prior to
4552 the excessive message length check. Due to way memory is allocated in
4553 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4554 to service a connection. This could lead to a Denial of Service through
4555 memory exhaustion. However, the excessive message length check still takes
4556 place, and this would cause the connection to immediately fail. Assuming
4557 that the application calls SSL_free() on the failed connection in a timely
4558 manner then the 21Mb of allocated memory will then be immediately freed
4559 again. Therefore, the excessive memory allocation will be transitory in
4560 nature. This then means that there is only a security impact if:
4562 1) The application does not call SSL_free() in a timely manner in the event
4563 that the connection fails
4565 2) The application is working in a constrained environment where there is
4566 very little free memory
4568 3) The attacker initiates multiple connection attempts such that there are
4569 multiple connections in a state where memory has been allocated for the
4570 connection; SSL_free() has not yet been called; and there is insufficient
4571 memory to service the multiple requests.
4573 Except in the instance of (1) above any Denial Of Service is likely to be
4574 transitory because as soon as the connection fails the memory is
4575 subsequently freed again in the SSL_free() call. However there is an
4576 increased risk during this period of application crashes due to the lack of
4577 memory - which would then mean a more serious Denial of Service.
4579 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4580 (CVE-2016-6307 and CVE-2016-6308)
4584 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4585 had to be removed. Primary reason is that vendor assembler can't
4586 assemble our modules with -KPIC flag. As result it, assembly
4587 support, was not even available as option. But its lack means
4588 lack of side-channel resistant code, which is incompatible with
4589 security by todays standards. Fortunately gcc is readily available
4590 prepackaged option, which we firmly point at...
4594 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4596 * Windows command-line tool supports UTF-8 opt-in option for arguments
4597 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4598 (to any value) allows Windows user to access PKCS#12 file generated
4599 with Windows CryptoAPI and protected with non-ASCII password, as well
4600 as files generated under UTF-8 locale on Linux also protected with
4605 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4606 have been disabled by default and removed from DEFAULT, just like RC4.
4607 See the RC4 item below to re-enable both.
4611 * The method for finding the storage location for the Windows RAND seed file
4612 has changed. First we check %RANDFILE%. If that is not set then we check
4613 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4614 all else fails we fall back to C:\.
4618 * The EVP_EncryptUpdate() function has had its return type changed from void
4619 to int. A return of 0 indicates and error while a return of 1 indicates
4624 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4625 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4626 off the constant time implementation for RSA, DSA and DH have been made
4627 no-ops and deprecated.
4631 * Windows RAND implementation was simplified to only get entropy by
4632 calling CryptGenRandom(). Various other RAND-related tickets
4635 *Joseph Wylie Yandle, Rich Salz*
4637 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4638 and `OPENSSL_LH_`, respectively. The old names are available
4639 with API compatibility. They new names are now completely documented.
4643 * Unify TYPE_up_ref(obj) methods signature.
4644 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4645 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4646 int (instead of void) like all others TYPE_up_ref() methods.
4647 So now these methods also check the return value of CRYPTO_atomic_add(),
4648 and the validity of object reference counter.
4650 *fdasilvayy@gmail.com*
4652 * With Windows Visual Studio builds, the .pdb files are installed
4653 alongside the installed libraries and executables. For a static
4654 library installation, ossl_static.pdb is the associate compiler
4655 generated .pdb file to be used when linking programs.
4659 * Remove openssl.spec. Packaging files belong with the packagers.
4663 * Automatic Darwin/OSX configuration has had a refresh, it will now
4664 recognise x86_64 architectures automatically. You can still decide
4665 to build for a different bitness with the environment variable
4666 KERNEL_BITS (can be 32 or 64), for example:
4668 KERNEL_BITS=32 ./config
4672 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4673 256 bit AES and HMAC with SHA256.
4677 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4681 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4685 * To enable users to have their own config files and build file templates,
4686 Configure looks in the directory indicated by the environment variable
4687 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4688 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4689 name and is used as is.
4693 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4694 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4695 X509_CERT_FILE_CTX was removed.
4699 * "shared" builds are now the default. To create only static libraries use
4700 the "no-shared" Configure option.
4704 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4705 All of these option have not worked for some while and are fundamental
4710 * Make various cleanup routines no-ops and mark them as deprecated. Most
4711 global cleanup functions are no longer required because they are handled
4712 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4713 Explicitly de-initing can cause problems (e.g. where a library that uses
4714 OpenSSL de-inits, but an application is still using it). The affected
4715 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4716 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4717 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4718 COMP_zlib_cleanup().
4722 * --strict-warnings no longer enables runtime debugging options
4723 such as REF_DEBUG. Instead, debug options are automatically
4724 enabled with '--debug' builds.
4726 *Andy Polyakov, Emilia Käsper*
4728 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4729 have been moved out of the public header files. New functions for managing
4730 these have been added.
4734 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4735 objects have been moved out of the public header files. New
4736 functions for managing these have been added.
4740 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4741 have been moved out of the public header files. New functions for managing
4742 these have been added.
4746 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4747 moved out of the public header files. New functions for managing these
4752 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4756 * Removed the mk1mf build scripts.
4760 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4761 it is always safe to #include a header now.
4765 * Removed the aged BC-32 config and all its supporting scripts
4769 * Removed support for Ultrix, Netware, and OS/2.
4773 * Add support for HKDF.
4775 *Alessandro Ghedini*
4777 * Add support for blake2b and blake2s
4781 * Added support for "pipelining". Ciphers that have the
4782 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4783 encryptions/decryptions simultaneously. There are currently no built-in
4784 ciphers with this property but the expectation is that engines will be able
4785 to offer it to significantly improve throughput. Support has been extended
4786 into libssl so that multiple records for a single connection can be
4787 processed in one go (for >=TLS 1.1).
4791 * Added the AFALG engine. This is an async capable engine which is able to
4792 offload work to the Linux kernel. In this initial version it only supports
4793 AES128-CBC. The kernel must be version 4.1.0 or greater.
4797 * OpenSSL now uses a new threading API. It is no longer necessary to
4798 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4799 are two supported threading models: pthreads and windows threads. It is
4800 also possible to configure OpenSSL at compile time for "no-threads". The
4801 old threading API should no longer be used. The functions have been
4802 replaced with "no-op" compatibility macros.
4804 *Alessandro Ghedini, Matt Caswell*
4806 * Modify behavior of ALPN to invoke callback after SNI/servername
4807 callback, such that updates to the SSL_CTX affect ALPN.
4811 * Add SSL_CIPHER queries for authentication and key-exchange.
4815 * Changes to the DEFAULT cipherlist:
4816 - Prefer (EC)DHE handshakes over plain RSA.
4817 - Prefer AEAD ciphers over legacy ciphers.
4818 - Prefer ECDSA over RSA when both certificates are available.
4819 - Prefer TLSv1.2 ciphers/PRF.
4820 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4825 * Change the ECC default curve list to be this, in order: x25519,
4826 secp256r1, secp521r1, secp384r1.
4830 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4831 disabled by default. They can be re-enabled using the
4832 enable-weak-ssl-ciphers option to Configure.
4836 * If the server has ALPN configured, but supports no protocols that the
4837 client advertises, send a fatal "no_application_protocol" alert.
4838 This behaviour is SHALL in RFC 7301, though it isn't universally
4839 implemented by other servers.
4843 * Add X25519 support.
4844 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4845 for public and private key encoding using the format documented in
4846 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4847 key generation and key derivation.
4849 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4854 * Deprecate SRP_VBASE_get_by_user.
4855 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4856 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4857 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4858 seed, even if the seed is configured.
4860 Users should use SRP_VBASE_get1_by_user instead. Note that in
4861 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4862 also that even though configuring the SRP seed attempts to hide
4863 invalid usernames by continuing the handshake with fake
4864 credentials, this behaviour is not constant time and no strong
4865 guarantees are made that the handshake is indistinguishable from
4866 that of a valid user.
4870 * Configuration change; it's now possible to build dynamic engines
4871 without having to build shared libraries and vice versa. This
4872 only applies to the engines in `engines/`, those in `crypto/engine/`
4873 will always be built into libcrypto (i.e. "static").
4875 Building dynamic engines is enabled by default; to disable, use
4876 the configuration option "disable-dynamic-engine".
4878 The only requirements for building dynamic engines are the
4879 presence of the DSO module and building with position independent
4880 code, so they will also automatically be disabled if configuring
4881 with "disable-dso" or "disable-pic".
4883 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4884 are also taken away from openssl/opensslconf.h, as they are
4889 * Configuration change; if there is a known flag to compile
4890 position independent code, it will always be applied on the
4891 libcrypto and libssl object files, and never on the application
4892 object files. This means other libraries that use routines from
4893 libcrypto / libssl can be made into shared libraries regardless
4894 of how OpenSSL was configured.
4896 If this isn't desirable, the configuration options "disable-pic"
4897 or "no-pic" can be used to disable the use of PIC. This will
4898 also disable building shared libraries and dynamic engines.
4902 * Removed JPAKE code. It was experimental and has no wide use.
4906 * The INSTALL_PREFIX Makefile variable has been renamed to
4907 DESTDIR. That makes for less confusion on what this variable
4908 is for. Also, the configuration option --install_prefix is
4913 * Heartbeat for TLS has been removed and is disabled by default
4914 for DTLS; configure with enable-heartbeats. Code that uses the
4915 old #define's might need to be updated.
4917 *Emilia Käsper, Rich Salz*
4919 * Rename REF_CHECK to REF_DEBUG.
4923 * New "unified" build system
4925 The "unified" build system is aimed to be a common system for all
4926 platforms we support. With it comes new support for VMS.
4928 This system builds supports building in a different directory tree
4929 than the source tree. It produces one Makefile (for unix family
4930 or lookalikes), or one descrip.mms (for VMS).
4932 The source of information to make the Makefile / descrip.mms is
4933 small files called 'build.info', holding the necessary
4934 information for each directory with source to compile, and a
4935 template in Configurations, like unix-Makefile.tmpl or
4938 With this change, the library names were also renamed on Windows
4939 and on VMS. They now have names that are closer to the standard
4940 on Unix, and include the major version number, and in certain
4941 cases, the architecture they are built for. See "Notes on shared
4942 libraries" in INSTALL.
4944 We rely heavily on the perl module Text::Template.
4948 * Added support for auto-initialisation and de-initialisation of the library.
4949 OpenSSL no longer requires explicit init or deinit routines to be called,
4950 except in certain circumstances. See the OPENSSL_init_crypto() and
4951 OPENSSL_init_ssl() man pages for further information.
4955 * The arguments to the DTLSv1_listen function have changed. Specifically the
4956 "peer" argument is now expected to be a BIO_ADDR object.
4958 * Rewrite of BIO networking library. The BIO library lacked consistent
4959 support of IPv6, and adding it required some more extensive
4960 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4961 which hold all types of addresses and chains of address information.
4962 It also introduces a new API, with functions like BIO_socket,
4963 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4964 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4965 have been adapted accordingly.
4969 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4974 * CRIME protection: disable compression by default, even if OpenSSL is
4975 compiled with zlib enabled. Applications can still enable compression
4976 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4977 using the SSL_CONF library to configure compression.
4981 * The signature of the session callback configured with
4982 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4983 was explicitly marked as `const unsigned char*` instead of
4988 * Always DPURIFY. Remove the use of uninitialized memory in the
4989 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4993 * Removed many obsolete configuration items, including
4994 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4995 MD2_CHAR, MD2_INT, MD2_LONG
4997 IDEA_SHORT, IDEA_LONG
4998 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5000 *Rich Salz, with advice from Andy Polyakov*
5002 * Many BN internals have been moved to an internal header file.
5004 *Rich Salz with help from Andy Polyakov*
5006 * Configuration and writing out the results from it has changed.
5007 Files such as Makefile include/openssl/opensslconf.h and are now
5008 produced through general templates, such as Makefile.in and
5009 crypto/opensslconf.h.in and some help from the perl module
5012 Also, the center of configuration information is no longer
5013 Makefile. Instead, Configure produces a perl module in
5014 configdata.pm which holds most of the config data (in the hash
5015 table %config), the target data that comes from the target
5016 configuration in one of the `Configurations/*.conf` files (in
5021 * To clarify their intended purposes, the Configure options
5022 --prefix and --openssldir change their semantics, and become more
5023 straightforward and less interdependent.
5025 --prefix shall be used exclusively to give the location INSTALLTOP
5026 where programs, scripts, libraries, include files and manuals are
5027 going to be installed. The default is now /usr/local.
5029 --openssldir shall be used exclusively to give the default
5030 location OPENSSLDIR where certificates, private keys, CRLs are
5031 managed. This is also where the default openssl.cnf gets
5033 If the directory given with this option is a relative path, the
5034 values of both the --prefix value and the --openssldir value will
5035 be combined to become OPENSSLDIR.
5036 The default for --openssldir is INSTALLTOP/ssl.
5038 Anyone who uses --openssldir to specify where OpenSSL is to be
5039 installed MUST change to use --prefix instead.
5043 * The GOST engine was out of date and therefore it has been removed. An up
5044 to date GOST engine is now being maintained in an external repository.
5045 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5046 support for GOST ciphersuites (these are only activated if a GOST engine
5051 * EGD is no longer supported by default; use enable-egd when
5054 *Ben Kaduk and Rich Salz*
5056 * The distribution now has Makefile.in files, which are used to
5057 create Makefile's when Configure is run. *Configure must be run
5058 before trying to build now.*
5062 * The return value for SSL_CIPHER_description() for error conditions
5067 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5069 Obtaining and performing DNSSEC validation of TLSA records is
5070 the application's responsibility. The application provides
5071 the TLSA records of its choice to OpenSSL, and these are then
5072 used to authenticate the peer.
5074 The TLSA records need not even come from DNS. They can, for
5075 example, be used to implement local end-entity certificate or
5076 trust-anchor "pinning", where the "pin" data takes the form
5077 of TLSA records, which can augment or replace verification
5078 based on the usual WebPKI public certification authorities.
5082 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5083 continues to support deprecated interfaces in default builds.
5084 However, applications are strongly advised to compile their
5085 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5086 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5087 or the 1.1.0 releases.
5089 In environments in which all applications have been ported to
5090 not use any deprecated interfaces OpenSSL's Configure script
5091 should be used with the --api=1.1.0 option to entirely remove
5092 support for the deprecated features from the library and
5093 unconditionally disable them in the installed headers.
5094 Essentially the same effect can be achieved with the "no-deprecated"
5095 argument to Configure, except that this will always restrict
5096 the build to just the latest API, rather than a fixed API
5099 As applications are ported to future revisions of the API,
5100 they should update their compile-time OPENSSL_API_COMPAT define
5101 accordingly, but in most cases should be able to continue to
5102 compile with later releases.
5104 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5105 0x10000000L and 0x00908000L, respectively. However those
5106 versions did not support the OPENSSL_API_COMPAT feature, and
5107 so applications are not typically tested for explicit support
5108 of just the undeprecated features of either release.
5112 * Add support for setting the minimum and maximum supported protocol.
5113 It can bet set via the SSL_set_min_proto_version() and
5114 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5115 MaxProtocol. It's recommended to use the new APIs to disable
5116 protocols instead of disabling individual protocols using
5117 SSL_set_options() or SSL_CONF's Protocol. This change also
5118 removes support for disabling TLS 1.2 in the OpenSSL TLS
5119 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5123 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5127 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5128 and integrates ECDSA and ECDH functionality into EC. Implementations can
5129 now redirect key generation and no longer need to convert to or from
5132 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5133 include the ec.h header file instead.
5137 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5138 ciphers who are no longer supported and drops support the ephemeral RSA key
5139 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5143 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5144 opaque. For HMAC_CTX, the following constructors and destructors
5147 HMAC_CTX *HMAC_CTX_new(void);
5148 void HMAC_CTX_free(HMAC_CTX *ctx);
5150 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5151 destroy such methods has been added. See EVP_MD_meth_new(3) and
5152 EVP_CIPHER_meth_new(3) for documentation.
5155 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5156 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5157 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5158 an already created structure.
5159 2) For consistency with the majority of our object creators and
5160 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5161 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5162 for deprecated builds.
5166 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5167 cryptographic operations to be performed asynchronously as long as an
5168 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5169 further details. Libssl has also had this capability integrated with the
5170 introduction of the new mode SSL_MODE_ASYNC and associated error
5171 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5172 pages. This work was developed in partnership with Intel Corp.
5176 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5177 always enabled now. If you want to disable the support you should
5178 exclude it using the list of supported ciphers. This also means that the
5179 "-no_ecdhe" option has been removed from s_server.
5183 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5184 SSL_{CTX_}set1_curves() which can set a list.
5188 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5189 curve you want to support using SSL_{CTX_}set1_curves().
5193 * State machine rewrite. The state machine code has been significantly
5194 refactored in order to remove much duplication of code and solve issues
5195 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5196 further details). This change does have some associated API changes.
5197 Notably the SSL_state() function has been removed and replaced by
5198 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5199 SSL_set_state() has been removed altogether. The previous handshake states
5200 defined in ssl.h and ssl3.h have also been removed.
5204 * All instances of the string "ssleay" in the public API were replaced
5205 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5206 Some error codes related to internal RSA_eay API's were renamed.
5210 * The demo files in crypto/threads were moved to demo/threads.
5214 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5217 *Matt Caswell, Rich Salz*
5219 * New ASN.1 embed macro.
5221 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5222 structure is not allocated: it is part of the parent. That is instead of
5230 This reduces memory fragmentation and make it impossible to accidentally
5231 set a mandatory field to NULL.
5233 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5234 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5235 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5240 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5244 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5245 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5246 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5247 DES and RC4 ciphersuites.
5251 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5252 This changes the decoding behaviour for some invalid messages,
5253 though the change is mostly in the more lenient direction, and
5254 legacy behaviour is preserved as much as possible.
5258 * Fix no-stdio build.
5259 *David Woodhouse <David.Woodhouse@intel.com> and also*
5260 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5262 * New testing framework
5263 The testing framework has been largely rewritten and is now using
5264 perl and the perl modules Test::Harness and an extended variant of
5265 Test::More called OpenSSL::Test to do its work. All test scripts in
5266 test/ have been rewritten into test recipes, and all direct calls to
5267 executables in test/Makefile have become individual recipes using the
5268 simplified testing OpenSSL::Test::Simple.
5270 For documentation on our testing modules, do:
5272 perldoc test/testlib/OpenSSL/Test/Simple.pm
5273 perldoc test/testlib/OpenSSL/Test.pm
5277 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5278 are used; the latter aborts on memory leaks (usually checked on exit).
5279 Some undocumented "set malloc, etc., hooks" functions were removed
5280 and others were changed. All are now documented.
5284 * In DSA_generate_parameters_ex, if the provided seed is too short,
5287 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5289 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5290 from RFC4279, RFC4785, RFC5487, RFC5489.
5292 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5293 original RSA_PSK patch.
5297 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5298 era flag was never set throughout the codebase (only read). Also removed
5299 SSL3_FLAGS_POP_BUFFER which was only used if
5300 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5304 * Changed the default name options in the "ca", "crl", "req" and "x509"
5305 to be "oneline" instead of "compat".
5309 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5310 not aware of clients that still exhibit this bug, and the workaround
5311 hasn't been working properly for a while.
5315 * The return type of BIO_number_read() and BIO_number_written() as well as
5316 the corresponding num_read and num_write members in the BIO structure has
5317 changed from unsigned long to uint64_t. On platforms where an unsigned
5318 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5323 * Given the pervasive nature of TLS extensions it is inadvisable to run
5324 OpenSSL without support for them. It also means that maintaining
5325 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5326 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5330 * Removed support for the two export grade static DH ciphersuites
5331 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5332 were newly added (along with a number of other static DH ciphersuites) to
5333 1.0.2. However the two export ones have *never* worked since they were
5334 introduced. It seems strange in any case to be adding new export
5335 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5339 * Version negotiation has been rewritten. In particular SSLv23_method(),
5340 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5341 and turned into macros which simply call the new preferred function names
5342 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5343 should use the new names instead. Also as part of this change the ssl23.h
5344 header file has been removed.
5348 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5349 code and the associated standard is no longer considered fit-for-purpose.
5353 * RT2547 was closed. When generating a private key, try to make the
5354 output file readable only by the owner. This behavior change might
5355 be noticeable when interacting with other software.
5357 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5362 * Added HTTP GET support to the ocsp command.
5366 * Changed default digest for the dgst and enc commands from MD5 to
5371 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5375 * Added support for TLS extended master secret from
5376 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5377 initial patch which was a great help during development.
5381 * All libssl internal structures have been removed from the public header
5382 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5383 now redundant). Users should not attempt to access internal structures
5384 directly. Instead they should use the provided API functions.
5388 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5389 Access to deprecated functions can be re-enabled by running config with
5390 "enable-deprecated". In addition applications wishing to use deprecated
5391 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5392 will, by default, disable some transitive includes that previously existed
5393 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5397 * Added support for OCB mode. OpenSSL has been granted a patent license
5398 compatible with the OpenSSL license for use of OCB. Details are available
5399 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5400 for OCB can be removed by calling config with no-ocb.
5404 * SSLv2 support has been removed. It still supports receiving an SSLv2
5405 compatible client hello.
5409 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5410 done while fixing the error code for the key-too-small case.
5412 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5414 * CA.sh has been removed; use CA.pl instead.
5418 * Removed old DES API.
5422 * Remove various unsupported platforms:
5428 Sinix/ReliantUNIX RM400
5433 16-bit platforms such as WIN16
5437 * Clean up OPENSSL_NO_xxx #define's
5438 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5439 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5440 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5441 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5442 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5443 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5444 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5445 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5446 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5447 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5451 * Cleaned up dead code
5452 Remove all but one '#ifdef undef' which is to be looked at.
5456 * Clean up calling of xxx_free routines.
5457 Just like free(), fix most of the xxx_free routines to accept
5458 NULL. Remove the non-null checks from callers. Save much code.
5462 * Add secure heap for storage of private keys (when possible).
5463 Add BIO_s_secmem(), CBIGNUM, etc.
5464 Contributed by Akamai Technologies under our Corporate CLA.
5468 * Experimental support for a new, fast, unbiased prime candidate generator,
5469 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5471 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5473 * New output format NSS in the sess_id command line tool. This allows
5474 exporting the session id and the master key in NSS keylog format.
5476 *Martin Kaiser <martin@kaiser.cx>*
5478 * Harmonize version and its documentation. -f flag is used to display
5481 *mancha <mancha1@zoho.com>*
5483 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5484 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5486 *mancha <mancha1@zoho.com>*
5488 * Fix some double frees. These are not thought to be exploitable.
5490 *mancha <mancha1@zoho.com>*
5492 * A missing bounds check in the handling of the TLS heartbeat extension
5493 can be used to reveal up to 64k of memory to a connected client or
5496 Thanks for Neel Mehta of Google Security for discovering this bug and to
5497 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5498 preparing the fix ([CVE-2014-0160])
5500 *Adam Langley, Bodo Moeller*
5502 * Fix for the attack described in the paper "Recovering OpenSSL
5503 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5504 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5505 <http://eprint.iacr.org/2014/140>
5507 Thanks to Yuval Yarom and Naomi Benger for discovering this
5508 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5510 *Yuval Yarom and Naomi Benger*
5512 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5513 this fixes a limitation in previous versions of OpenSSL.
5517 * Experimental encrypt-then-mac support.
5519 Experimental support for encrypt then mac from
5520 draft-gutmann-tls-encrypt-then-mac-02.txt
5522 To enable it set the appropriate extension number (0x42 for the test
5523 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5525 For non-compliant peers (i.e. just about everything) this should have no
5528 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5532 * Add EVP support for key wrapping algorithms, to avoid problems with
5533 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5534 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5535 algorithms and include tests cases.
5539 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5544 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5545 MGF1 digest and OAEP label.
5549 * Make openssl verify return errors.
5551 *Chris Palmer <palmer@google.com> and Ben Laurie*
5553 * New function ASN1_TIME_diff to calculate the difference between two
5554 ASN1_TIME structures or one structure and the current time.
5558 * Update fips_test_suite to support multiple command line options. New
5559 test to induce all self test errors in sequence and check expected
5564 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5565 sign or verify all in one operation.
5569 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5570 test programs and fips_test_suite. Includes functionality to parse
5571 the minimal script output of fipsalgest.pl directly.
5575 * Add authorisation parameter to FIPS_module_mode_set().
5579 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5583 * Use separate DRBG fields for internal and external flags. New function
5584 FIPS_drbg_health_check() to perform on demand health checking. Add
5585 generation tests to fips_test_suite with reduced health check interval to
5586 demonstrate periodic health checking. Add "nodh" option to
5587 fips_test_suite to skip very slow DH test.
5591 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5596 * More extensive health check for DRBG checking many more failure modes.
5597 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5598 combination: call this in fips_test_suite.
5602 * Add support for canonical generation of DSA parameter 'g'. See
5605 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5606 POST to handle HMAC cases.
5610 * Add functions FIPS_module_version() and FIPS_module_version_text()
5611 to return numerical and string versions of the FIPS module number.
5615 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5616 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5617 outside the validated module in the FIPS capable OpenSSL.
5621 * Minor change to DRBG entropy callback semantics. In some cases
5622 there is no multiple of the block length between min_len and
5623 max_len. Allow the callback to return more than max_len bytes
5624 of entropy but discard any extra: it is the callback's responsibility
5625 to ensure that the extra data discarded does not impact the
5626 requested amount of entropy.
5630 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5631 information in FIPS186-3, SP800-57 and SP800-131A.
5635 * CCM support via EVP. Interface is very similar to GCM case except we
5636 must supply all data in one chunk (i.e. no update, final) and the
5637 message length must be supplied if AAD is used. Add algorithm test
5642 * Initial version of POST overhaul. Add POST callback to allow the status
5643 of POST to be monitored and/or failures induced. Modify fips_test_suite
5644 to use callback. Always run all selftests even if one fails.
5648 * XTS support including algorithm test driver in the fips_gcmtest program.
5649 Note: this does increase the maximum key length from 32 to 64 bytes but
5650 there should be no binary compatibility issues as existing applications
5651 will never use XTS mode.
5655 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5656 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5657 performs algorithm blocking for unapproved PRNG types. Also do not
5658 set PRNG type in FIPS_mode_set(): leave this to the application.
5659 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5660 the standard OpenSSL PRNG: set additional data to a date time vector.
5664 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5665 This shouldn't present any incompatibility problems because applications
5666 shouldn't be using these directly and any that are will need to rethink
5667 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5671 * Extensive self tests and health checking required by SP800-90 DRBG.
5672 Remove strength parameter from FIPS_drbg_instantiate and always
5673 instantiate at maximum supported strength.
5677 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5681 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5685 * New function DH_compute_key_padded() to compute a DH key and pad with
5686 leading zeroes if needed: this complies with SP800-56A et al.
5690 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5691 anything, incomplete, subject to change and largely untested at present.
5695 * Modify fipscanisteronly build option to only build the necessary object
5696 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5700 * Add experimental option FIPSSYMS to give all symbols in
5701 fipscanister.o and FIPS or fips prefix. This will avoid
5702 conflicts with future versions of OpenSSL. Add perl script
5703 util/fipsas.pl to preprocess assembly language source files
5704 and rename any affected symbols.
5708 * Add selftest checks and algorithm block of non-fips algorithms in
5709 FIPS mode. Remove DES2 from selftests.
5713 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5714 return internal method without any ENGINE dependencies. Add new
5715 tiny fips sign and verify functions.
5719 * New build option no-ec2m to disable characteristic 2 code.
5723 * New build option "fipscanisteronly". This only builds fipscanister.o
5724 and (currently) associated fips utilities. Uses the file Makefile.fips
5725 instead of Makefile.org as the prototype.
5729 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5730 Update fips_gcmtest to use IV generator.
5734 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5735 setting output buffer to NULL. The `*Final` function must be
5736 called although it will not retrieve any additional data. The tag
5737 can be set or retrieved with a ctrl. The IV length is by default 12
5738 bytes (96 bits) but can be set to an alternative value. If the IV
5739 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5744 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5745 underlying do_cipher function handles all cipher semantics itself
5746 including padding and finalisation. This is useful if (for example)
5747 an ENGINE cipher handles block padding itself. The behaviour of
5748 do_cipher is subtly changed if this flag is set: the return value
5749 is the number of characters written to the output buffer (zero is
5750 no longer an error code) or a negative error code. Also if the
5751 input buffer is NULL and length 0 finalisation should be performed.
5755 * If a candidate issuer certificate is already part of the constructed
5756 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5760 * Improve forward-security support: add functions
5762 void SSL_CTX_set_not_resumable_session_callback(
5763 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5764 void SSL_set_not_resumable_session_callback(
5765 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5767 for use by SSL/TLS servers; the callback function will be called whenever a
5768 new session is created, and gets to decide whether the session may be
5769 cached to make it resumable (return 0) or not (return 1). (As by the
5770 SSL/TLS protocol specifications, the session_id sent by the server will be
5771 empty to indicate that the session is not resumable; also, the server will
5772 not generate RFC 4507 (RFC 5077) session tickets.)
5774 A simple reasonable callback implementation is to return is_forward_secure.
5775 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5776 by the SSL/TLS server library, indicating whether it can provide forward
5779 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5781 * New -verify_name option in command line utilities to set verification
5786 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5787 Add CMAC pkey methods.
5791 * Experimental renegotiation in s_server -www mode. If the client
5792 browses /reneg connection is renegotiated. If /renegcert it is
5793 renegotiated requesting a certificate.
5797 * Add an "external" session cache for debugging purposes to s_server. This
5798 should help trace issues which normally are only apparent in deployed
5799 multi-process servers.
5803 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5804 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5805 BIO_set_cipher() and some obscure PEM functions were changed so they
5806 can now return an error. The RAND changes required a change to the
5807 RAND_METHOD structure.
5811 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5812 a gcc attribute to warn if the result of a function is ignored. This
5813 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5814 whose return value is often ignored.
5818 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5819 These allow SCTs (signed certificate timestamps) to be requested and
5820 validated when establishing a connection.
5822 *Rob Percival <robpercival@google.com>*
5827 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5829 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5830 used even when parsing explicit parameters, when loading a encoded key
5831 or calling `EC_GROUP_new_from_ecpkparameters()`/
5832 `EC_GROUP_new_from_ecparameters()`.
5833 This prevents bypass of security hardening and performance gains,
5834 especially for curves with specialized EC_METHODs.
5835 By default, if a key encoded with explicit parameters is loaded and later
5836 encoded, the output is still encoded with explicit parameters, even if
5837 internally a "named" EC_GROUP is used for computation.
5841 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5842 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5843 NULL. After this change, only the cofactor parameter can be NULL. It also
5844 does some minimal sanity checks on the passed order.
5849 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5850 An attack is simple, if the first CMS_recipientInfo is valid but the
5851 second CMS_recipientInfo is chosen ciphertext. If the second
5852 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5853 encryption key will be replaced by garbage, and the message cannot be
5854 decoded, but if the RSA decryption fails, the correct encryption key is
5855 used and the recipient will not notice the attack.
5856 As a work around for this potential attack the length of the decrypted
5857 key must be equal to the cipher default key length, in case the
5858 certificate is not given and all recipientInfo are tried out.
5859 The old behaviour can be re-enabled in the CMS code by setting the
5860 CMS_DEBUG_DECRYPT flag.
5865 * Document issue with installation paths in diverse Windows builds
5867 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5868 binaries and run-time config file.
5873 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5875 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5876 This changes the size when using the `genpkey` command when no size is given.
5877 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5878 generation commands to use 2048 bits by default.
5882 * Add FIPS support for Android Arm 64-bit
5884 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5885 Module in Version 2.0.10. For some reason, the corresponding target
5886 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5887 built with FIPS support on Android Arm 64-bit. This omission has been
5890 *Matthias St. Pierre*
5892 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5894 * 0-byte record padding oracle
5896 If an application encounters a fatal protocol error and then calls
5897 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5898 then OpenSSL can respond differently to the calling application if a 0 byte
5899 record is received with invalid padding compared to if a 0 byte record is
5900 received with an invalid MAC. If the application then behaves differently
5901 based on that in a way that is detectable to the remote peer, then this
5902 amounts to a padding oracle that could be used to decrypt data.
5904 In order for this to be exploitable "non-stitched" ciphersuites must be in
5905 use. Stitched ciphersuites are optimised implementations of certain
5906 commonly used ciphersuites. Also the application must call SSL_shutdown()
5907 twice even if a protocol error has occurred (applications should not do
5908 this but some do anyway).
5910 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5911 Aviram, with additional investigation by Steven Collison and Andrew
5912 Hourselt. It was reported to OpenSSL on 10th December 2018.
5917 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5921 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5923 * Microarchitecture timing vulnerability in ECC scalar multiplication
5925 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5926 shown to be vulnerable to a microarchitecture timing side channel attack.
5927 An attacker with sufficient access to mount local timing attacks during
5928 ECDSA signature generation could recover the private key.
5930 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5931 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5937 * Timing vulnerability in DSA signature generation
5939 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5940 timing side channel attack. An attacker could use variations in the signing
5941 algorithm to recover the private key.
5943 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5948 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5949 Module, accidentally introduced while backporting security fixes from the
5950 development branch and hindering the use of ECC in FIPS mode.
5954 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5956 * Client DoS due to large DH parameter
5958 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5959 malicious server can send a very large prime value to the client. This will
5960 cause the client to spend an unreasonably long period of time generating a
5961 key for this prime resulting in a hang until the client has finished. This
5962 could be exploited in a Denial Of Service attack.
5964 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5969 * Cache timing vulnerability in RSA Key Generation
5971 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5972 a cache timing side channel attack. An attacker with sufficient access to
5973 mount cache timing attacks during the RSA key generation process could
5974 recover the private key.
5976 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5977 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5982 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5983 parameter is no longer accepted, as it leads to a corrupt table. NULL
5984 pem_str is reserved for alias entries only.
5988 * Revert blinding in ECDSA sign and instead make problematic addition
5989 length-invariant. Switch even to fixed-length Montgomery multiplication.
5993 * Change generating and checking of primes so that the error rate of not
5994 being prime depends on the intended use based on the size of the input.
5995 For larger primes this will result in more rounds of Miller-Rabin.
5996 The maximal error rate for primes with more than 1080 bits is lowered
5999 *Kurt Roeckx, Annie Yousar*
6001 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6005 * Add blinding to ECDSA and DSA signatures to protect against side channel
6006 attacks discovered by Keegan Ryan (NCC Group).
6010 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6011 now allow empty (zero character) pass phrases.
6015 * Certificate time validation (X509_cmp_time) enforces stricter
6016 compliance with RFC 5280. Fractional seconds and timezone offsets
6017 are no longer allowed.
6021 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6023 * Constructed ASN.1 types with a recursive definition could exceed the stack
6025 Constructed ASN.1 types with a recursive definition (such as can be found
6026 in PKCS7) could eventually exceed the stack given malicious input with
6027 excessive recursion. This could result in a Denial Of Service attack. There
6028 are no such structures used within SSL/TLS that come from untrusted sources
6029 so this is considered safe.
6031 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6037 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6039 * Read/write after SSL object in error state
6041 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6042 mechanism. The intent was that if a fatal error occurred during a handshake
6043 then OpenSSL would move into the error state and would immediately fail if
6044 you attempted to continue the handshake. This works as designed for the
6045 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6046 SSL_connect()), however due to a bug it does not work correctly if
6047 SSL_read() or SSL_write() is called directly. In that scenario, if the
6048 handshake fails then a fatal error will be returned in the initial function
6049 call. If SSL_read()/SSL_write() is subsequently called by the application
6050 for the same SSL object then it will succeed and the data is passed without
6051 being decrypted/encrypted directly from the SSL/TLS record layer.
6053 In order to exploit this issue an application bug would have to be present
6054 that resulted in a call to SSL_read()/SSL_write() being issued after having
6055 already received a fatal error.
6057 This issue was reported to OpenSSL by David Benjamin (Google).
6062 * rsaz_1024_mul_avx2 overflow bug on x86_64
6064 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6065 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6066 Analysis suggests that attacks against RSA and DSA as a result of this
6067 defect would be very difficult to perform and are not believed likely.
6068 Attacks against DH1024 are considered just feasible, because most of the
6069 work necessary to deduce information about a private key may be performed
6070 offline. The amount of resources required for such an attack would be
6071 significant. However, for an attack on TLS to be meaningful, the server
6072 would have to share the DH1024 private key among multiple clients, which is
6073 no longer an option since CVE-2016-0701.
6075 This only affects processors that support the AVX2 but not ADX extensions
6076 like Intel Haswell (4th generation).
6078 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6079 was originally found via the OSS-Fuzz project.
6084 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6086 * bn_sqrx8x_internal carry bug on x86_64
6088 There is a carry propagating bug in the x86_64 Montgomery squaring
6089 procedure. No EC algorithms are affected. Analysis suggests that attacks
6090 against RSA and DSA as a result of this defect would be very difficult to
6091 perform and are not believed likely. Attacks against DH are considered just
6092 feasible (although very difficult) because most of the work necessary to
6093 deduce information about a private key may be performed offline. The amount
6094 of resources required for such an attack would be very significant and
6095 likely only accessible to a limited number of attackers. An attacker would
6096 additionally need online access to an unpatched system using the target
6097 private key in a scenario with persistent DH parameters and a private
6098 key that is shared between multiple clients.
6100 This only affects processors that support the BMI1, BMI2 and ADX extensions
6101 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6103 This issue was reported to OpenSSL by the OSS-Fuzz project.
6108 * Malformed X.509 IPAddressFamily could cause OOB read
6110 If an X.509 certificate has a malformed IPAddressFamily extension,
6111 OpenSSL could do a one-byte buffer overread. The most likely result
6112 would be an erroneous display of the certificate in text format.
6114 This issue was reported to OpenSSL by the OSS-Fuzz project.
6118 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6120 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6121 platform rather than 'mingw'.
6125 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6127 * Truncated packet could crash via OOB read
6129 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6130 cipher is being used, then a truncated packet can cause that host to
6131 perform an out-of-bounds read, usually resulting in a crash.
6133 This issue was reported to OpenSSL by Robert Święcki of Google.
6138 * BN_mod_exp may produce incorrect results on x86_64
6140 There is a carry propagating bug in the x86_64 Montgomery squaring
6141 procedure. No EC algorithms are affected. Analysis suggests that attacks
6142 against RSA and DSA as a result of this defect would be very difficult to
6143 perform and are not believed likely. Attacks against DH are considered just
6144 feasible (although very difficult) because most of the work necessary to
6145 deduce information about a private key may be performed offline. The amount
6146 of resources required for such an attack would be very significant and
6147 likely only accessible to a limited number of attackers. An attacker would
6148 additionally need online access to an unpatched system using the target
6149 private key in a scenario with persistent DH parameters and a private
6150 key that is shared between multiple clients. For example this can occur by
6151 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6152 similar to CVE-2015-3193 but must be treated as a separate problem.
6154 This issue was reported to OpenSSL by the OSS-Fuzz project.
6159 * Montgomery multiplication may produce incorrect results
6161 There is a carry propagating bug in the Broadwell-specific Montgomery
6162 multiplication procedure that handles input lengths divisible by, but
6163 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6164 and DH private keys are impossible. This is because the subroutine in
6165 question is not used in operations with the private key itself and an input
6166 of the attacker's direct choice. Otherwise the bug can manifest itself as
6167 transient authentication and key negotiation failures or reproducible
6168 erroneous outcome of public-key operations with specially crafted input.
6169 Among EC algorithms only Brainpool P-512 curves are affected and one
6170 presumably can attack ECDH key negotiation. Impact was not analyzed in
6171 detail, because pre-requisites for attack are considered unlikely. Namely
6172 multiple clients have to choose the curve in question and the server has to
6173 share the private key among them, neither of which is default behaviour.
6174 Even then only clients that chose the curve will be affected.
6176 This issue was publicly reported as transient failures and was not
6177 initially recognized as a security issue. Thanks to Richard Morgan for
6178 providing reproducible case.
6183 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6184 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6185 prevent issues where no progress is being made and the peer continually
6186 sends unrecognised record types, using up resources processing them.
6190 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6192 * Missing CRL sanity check
6194 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6195 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6196 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6198 This issue only affects the OpenSSL 1.0.2i
6203 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6205 * OCSP Status Request extension unbounded memory growth
6207 A malicious client can send an excessively large OCSP Status Request
6208 extension. If that client continually requests renegotiation, sending a
6209 large OCSP Status Request extension each time, then there will be unbounded
6210 memory growth on the server. This will eventually lead to a Denial Of
6211 Service attack through memory exhaustion. Servers with a default
6212 configuration are vulnerable even if they do not support OCSP. Builds using
6213 the "no-ocsp" build time option are not affected.
6215 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6220 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6223 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6229 * OOB write in MDC2_Update()
6231 An overflow can occur in MDC2_Update() either if called directly or
6232 through the EVP_DigestUpdate() function using MDC2. If an attacker
6233 is able to supply very large amounts of input data after a previous
6234 call to EVP_EncryptUpdate() with a partial block then a length check
6235 can overflow resulting in a heap corruption.
6237 The amount of data needed is comparable to SIZE_MAX which is impractical
6240 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6245 * Malformed SHA512 ticket DoS
6247 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6248 DoS attack where a malformed ticket will result in an OOB read which will
6251 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6252 a custom server callback and ticket lookup mechanism.
6254 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6259 * OOB write in BN_bn2dec()
6261 The function BN_bn2dec() does not check the return value of BN_div_word().
6262 This can cause an OOB write if an application uses this function with an
6263 overly large BIGNUM. This could be a problem if an overly large certificate
6264 or CRL is printed out from an untrusted source. TLS is not affected because
6265 record limits will reject an oversized certificate before it is parsed.
6267 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6272 * OOB read in TS_OBJ_print_bio()
6274 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6275 the total length the OID text representation would use and not the amount
6276 of data written. This will result in OOB reads when large OIDs are
6279 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6284 * Pointer arithmetic undefined behaviour
6286 Avoid some undefined pointer arithmetic
6288 A common idiom in the codebase is to check limits in the following manner:
6291 Where "p" points to some malloc'd data of SIZE bytes and
6294 "len" here could be from some externally supplied data (e.g. from a TLS
6297 The rules of C pointer arithmetic are such that "p + len" is only well
6298 defined where len <= SIZE. Therefore the above idiom is actually
6299 undefined behaviour.
6301 For example this could cause problems if some malloc implementation
6302 provides an address for "p" such that "p + len" actually overflows for
6303 values of len that are too big and therefore p + len < limit.
6305 This issue was reported to OpenSSL by Guido Vranken
6310 * Constant time flag not preserved in DSA signing
6312 Operations in the DSA signing algorithm should run in constant time in
6313 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6314 implementation means that a non-constant time codepath is followed for
6315 certain operations. This has been demonstrated through a cache-timing
6316 attack to be sufficient for an attacker to recover the private DSA key.
6318 This issue was reported by César Pereida (Aalto University), Billy Brumley
6319 (Tampere University of Technology), and Yuval Yarom (The University of
6320 Adelaide and NICTA).
6325 * DTLS buffered message DoS
6327 In a DTLS connection where handshake messages are delivered out-of-order
6328 those messages that OpenSSL is not yet ready to process will be buffered
6329 for later use. Under certain circumstances, a flaw in the logic means that
6330 those messages do not get removed from the buffer even though the handshake
6331 has been completed. An attacker could force up to approx. 15 messages to
6332 remain in the buffer when they are no longer required. These messages will
6333 be cleared when the DTLS connection is closed. The default maximum size for
6334 a message is 100k. Therefore, the attacker could force an additional 1500k
6335 to be consumed per connection. By opening many simultaneous connections an
6336 attacker could cause a DoS attack through memory exhaustion.
6338 This issue was reported to OpenSSL by Quan Luo.
6343 * DTLS replay protection DoS
6345 A flaw in the DTLS replay attack protection mechanism means that records
6346 that arrive for future epochs update the replay protection "window" before
6347 the MAC for the record has been validated. This could be exploited by an
6348 attacker by sending a record for the next epoch (which does not have to
6349 decrypt or have a valid MAC), with a very large sequence number. This means
6350 that all subsequent legitimate packets are dropped causing a denial of
6351 service for a specific DTLS connection.
6353 This issue was reported to OpenSSL by the OCAP audit team.
6358 * Certificate message OOB reads
6360 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6361 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6362 theoretical DoS risk but this has not been observed in practice on common
6365 The messages affected are client certificate, client certificate request
6366 and server certificate. As a result the attack can only be performed
6367 against a client or a server which enables client authentication.
6369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6374 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6376 * Prevent padding oracle in AES-NI CBC MAC check
6378 A MITM attacker can use a padding oracle attack to decrypt traffic
6379 when the connection uses an AES CBC cipher and the server support
6382 This issue was introduced as part of the fix for Lucky 13 padding
6383 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6384 constant time by making sure that always the same bytes are read and
6385 compared against either the MAC or padding bytes. But it no longer
6386 checked that there was enough data to have both the MAC and padding
6389 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6393 * Fix EVP_EncodeUpdate overflow
6395 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6396 Base64 encoding of binary data. If an attacker is able to supply very large
6397 amounts of input data then a length check can overflow resulting in a heap
6400 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6401 the `PEM_write_bio*` family of functions. These are mainly used within the
6402 OpenSSL command line applications, so any application which processes data
6403 from an untrusted source and outputs it as a PEM file should be considered
6404 vulnerable to this issue. User applications that call these APIs directly
6405 with large amounts of untrusted data may also be vulnerable.
6407 This issue was reported by Guido Vranken.
6412 * Fix EVP_EncryptUpdate overflow
6414 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6415 is able to supply very large amounts of input data after a previous call to
6416 EVP_EncryptUpdate() with a partial block then a length check can overflow
6417 resulting in a heap corruption. Following an analysis of all OpenSSL
6418 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6419 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6420 the first called function after an EVP_EncryptInit(), and therefore that
6421 specific call must be safe. The second form is where the length passed to
6422 EVP_EncryptUpdate() can be seen from the code to be some small value and
6423 therefore there is no possibility of an overflow. Since all instances are
6424 one of these two forms, it is believed that there can be no overflows in
6425 internal code due to this problem. It should be noted that
6426 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6427 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6428 of these calls have also been analysed too and it is believed there are no
6429 instances in internal usage where an overflow could occur.
6431 This issue was reported by Guido Vranken.
6436 * Prevent ASN.1 BIO excessive memory allocation
6438 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6439 a short invalid encoding can cause allocation of large amounts of memory
6440 potentially consuming excessive resources or exhausting memory.
6442 Any application parsing untrusted data through d2i BIO functions is
6443 affected. The memory based functions such as d2i_X509() are *not* affected.
6444 Since the memory based functions are used by the TLS library, TLS
6445 applications are not affected.
6447 This issue was reported by Brian Carpenter.
6454 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6455 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6456 in arbitrary stack data being returned in the buffer.
6458 This issue was reported by Guido Vranken.
6463 * Modify behavior of ALPN to invoke callback after SNI/servername
6464 callback, such that updates to the SSL_CTX affect ALPN.
6468 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6473 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6474 methods are enabled and ssl2 is disabled the methods return NULL.
6478 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6480 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6481 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6482 provide any "EXPORT" or "LOW" strength ciphers.
6486 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6487 is by default disabled at build-time. Builds that are not configured with
6488 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6489 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6490 will need to explicitly call either of:
6492 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6494 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6496 as appropriate. Even if either of those is used, or the application
6497 explicitly uses the version-specific SSLv2_method() or its client and
6498 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6499 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6500 ciphers, and SSLv2 56-bit DES are no longer available.
6505 * Fix a double-free in DSA code
6507 A double free bug was discovered when OpenSSL parses malformed DSA private
6508 keys and could lead to a DoS attack or memory corruption for applications
6509 that receive DSA private keys from untrusted sources. This scenario is
6512 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6518 * Disable SRP fake user seed to address a server memory leak.
6520 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6522 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6523 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6524 was changed to ignore the "fake user" SRP seed, even if the seed
6527 Users should use SRP_VBASE_get1_by_user instead. Note that in
6528 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6529 also that even though configuring the SRP seed attempts to hide
6530 invalid usernames by continuing the handshake with fake
6531 credentials, this behaviour is not constant time and no strong
6532 guarantees are made that the handshake is indistinguishable from
6533 that of a valid user.
6538 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6540 In the BN_hex2bn function the number of hex digits is calculated using an
6541 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6542 large values of `i` this can result in `bn_expand` not allocating any
6543 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6544 field as NULL leading to a subsequent NULL ptr deref. For very large values
6545 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6546 In this case memory is allocated to the internal BIGNUM data field, but it
6547 is insufficiently sized leading to heap corruption. A similar issue exists
6548 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6549 is ever called by user applications with very large untrusted hex/dec data.
6550 This is anticipated to be a rare occurrence.
6552 All OpenSSL internal usage of these functions use data that is not expected
6553 to be untrusted, e.g. config file data or application command line
6554 arguments. If user developed applications generate config file data based
6555 on untrusted data then it is possible that this could also lead to security
6556 consequences. This is also anticipated to be rare.
6558 This issue was reported to OpenSSL by Guido Vranken.
6563 * Fix memory issues in `BIO_*printf` functions
6565 The internal `fmtstr` function used in processing a "%s" format string in
6566 the `BIO_*printf` functions could overflow while calculating the length of a
6567 string and cause an OOB read when printing very long strings.
6569 Additionally the internal `doapr_outch` function can attempt to write to an
6570 OOB memory location (at an offset from the NULL pointer) in the event of a
6571 memory allocation failure. In 1.0.2 and below this could be caused where
6572 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6573 could be in processing a very long "%s" format string. Memory leaks can
6576 The first issue may mask the second issue dependent on compiler behaviour.
6577 These problems could enable attacks where large amounts of untrusted data
6578 is passed to the `BIO_*printf` functions. If applications use these functions
6579 in this way then they could be vulnerable. OpenSSL itself uses these
6580 functions when printing out human-readable dumps of ASN.1 data. Therefore
6581 applications that print this data could be vulnerable if the data is from
6582 untrusted sources. OpenSSL command line applications could also be
6583 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6584 as command line arguments.
6586 Libssl is not considered directly vulnerable. Additionally certificates etc
6587 received via remote connections via libssl are also unlikely to be able to
6588 trigger these issues because of message size limits enforced within libssl.
6590 This issue was reported to OpenSSL Guido Vranken.
6595 * Side channel attack on modular exponentiation
6597 A side-channel attack was found which makes use of cache-bank conflicts on
6598 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6599 of RSA keys. The ability to exploit this issue is limited as it relies on
6600 an attacker who has control of code in a thread running on the same
6601 hyper-threaded core as the victim thread which is performing decryptions.
6603 This issue was reported to OpenSSL by Yuval Yarom, The University of
6604 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6605 Nadia Heninger, University of Pennsylvania with more information at
6606 <http://cachebleed.info>.
6611 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6612 if no keysize is specified with default_bits. This fixes an
6613 omission in an earlier change that changed all RSA/DSA key generation
6614 commands to use 2048 bits by default.
6618 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6620 * DH small subgroups
6622 Historically OpenSSL only ever generated DH parameters based on "safe"
6623 primes. More recently (in version 1.0.2) support was provided for
6624 generating X9.42 style parameter files such as those required for RFC 5114
6625 support. The primes used in such files may not be "safe". Where an
6626 application is using DH configured with parameters based on primes that are
6627 not "safe" then an attacker could use this fact to find a peer's private
6628 DH exponent. This attack requires that the attacker complete multiple
6629 handshakes in which the peer uses the same private DH exponent. For example
6630 this could be used to discover a TLS server's private DH exponent if it's
6631 reusing the private DH exponent or it's using a static DH ciphersuite.
6633 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6634 TLS. It is not on by default. If the option is not set then the server
6635 reuses the same private DH exponent for the life of the server process and
6636 would be vulnerable to this attack. It is believed that many popular
6637 applications do set this option and would therefore not be at risk.
6639 The fix for this issue adds an additional check where a "q" parameter is
6640 available (as is the case in X9.42 based parameters). This detects the
6641 only known attack, and is the only possible defense for static DH
6642 ciphersuites. This could have some performance impact.
6644 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6645 default and cannot be disabled. This could have some performance impact.
6647 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6652 * SSLv2 doesn't block disabled ciphers
6654 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6655 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6656 been disabled, provided that the SSLv2 protocol was not also disabled via
6659 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6660 and Sebastian Schinzel.
6665 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6667 * BN_mod_exp may produce incorrect results on x86_64
6669 There is a carry propagating bug in the x86_64 Montgomery squaring
6670 procedure. No EC algorithms are affected. Analysis suggests that attacks
6671 against RSA and DSA as a result of this defect would be very difficult to
6672 perform and are not believed likely. Attacks against DH are considered just
6673 feasible (although very difficult) because most of the work necessary to
6674 deduce information about a private key may be performed offline. The amount
6675 of resources required for such an attack would be very significant and
6676 likely only accessible to a limited number of attackers. An attacker would
6677 additionally need online access to an unpatched system using the target
6678 private key in a scenario with persistent DH parameters and a private
6679 key that is shared between multiple clients. For example this can occur by
6680 default in OpenSSL DHE based SSL/TLS ciphersuites.
6682 This issue was reported to OpenSSL by Hanno Böck.
6687 * Certificate verify crash with missing PSS parameter
6689 The signature verification routines will crash with a NULL pointer
6690 dereference if presented with an ASN.1 signature using the RSA PSS
6691 algorithm and absent mask generation function parameter. Since these
6692 routines are used to verify certificate signature algorithms this can be
6693 used to crash any certificate verification operation and exploited in a
6694 DoS attack. Any application which performs certificate verification is
6695 vulnerable including OpenSSL clients and servers which enable client
6698 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6703 * X509_ATTRIBUTE memory leak
6705 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6706 memory. This structure is used by the PKCS#7 and CMS routines so any
6707 application which reads PKCS#7 or CMS data from untrusted sources is
6708 affected. SSL/TLS is not affected.
6710 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6716 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6717 This changes the decoding behaviour for some invalid messages,
6718 though the change is mostly in the more lenient direction, and
6719 legacy behaviour is preserved as much as possible.
6723 * In DSA_generate_parameters_ex, if the provided seed is too short,
6726 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6728 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6730 * Alternate chains certificate forgery
6732 During certificate verification, OpenSSL will attempt to find an
6733 alternative certificate chain if the first attempt to build such a chain
6734 fails. An error in the implementation of this logic can mean that an
6735 attacker could cause certain checks on untrusted certificates to be
6736 bypassed, such as the CA flag, enabling them to use a valid leaf
6737 certificate to act as a CA and "issue" an invalid certificate.
6739 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6744 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6746 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6747 incompatibility in the handling of HMAC. The previous ABI has now been
6752 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6754 * Malformed ECParameters causes infinite loop
6756 When processing an ECParameters structure OpenSSL enters an infinite loop
6757 if the curve specified is over a specially malformed binary polynomial
6760 This can be used to perform denial of service against any
6761 system which processes public keys, certificate requests or
6762 certificates. This includes TLS clients and TLS servers with
6763 client authentication enabled.
6765 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6770 * Exploitable out-of-bounds read in X509_cmp_time
6772 X509_cmp_time does not properly check the length of the ASN1_TIME
6773 string and can read a few bytes out of bounds. In addition,
6774 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6777 An attacker can use this to craft malformed certificates and CRLs of
6778 various sizes and potentially cause a segmentation fault, resulting in
6779 a DoS on applications that verify certificates or CRLs. TLS clients
6780 that verify CRLs are affected. TLS clients and servers with client
6781 authentication enabled may be affected if they use custom verification
6784 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6785 independently by Hanno Böck.
6790 * PKCS7 crash with missing EnvelopedContent
6792 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6793 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6794 with missing content and trigger a NULL pointer dereference on parsing.
6796 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6797 structures from untrusted sources are affected. OpenSSL clients and
6798 servers are not affected.
6800 This issue was reported to OpenSSL by Michal Zalewski (Google).
6805 * CMS verify infinite loop with unknown hash function
6807 When verifying a signedData message the CMS code can enter an infinite loop
6808 if presented with an unknown hash function OID. This can be used to perform
6809 denial of service against any system which verifies signedData messages using
6811 This issue was reported to OpenSSL by Johannes Bauer.
6816 * Race condition handling NewSessionTicket
6818 If a NewSessionTicket is received by a multi-threaded client when attempting to
6819 reuse a previous ticket then a race condition can occur potentially leading to
6820 a double free of the ticket data.
6825 * Only support 256-bit or stronger elliptic curves with the
6826 'ecdh_auto' setting (server) or by default (client). Of supported
6827 curves, prefer P-256 (both).
6831 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6833 * ClientHello sigalgs DoS fix
6835 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6836 invalid signature algorithms extension a NULL pointer dereference will
6837 occur. This can be exploited in a DoS attack against the server.
6839 This issue was was reported to OpenSSL by David Ramos of Stanford
6843 *Stephen Henson and Matt Caswell*
6845 * Multiblock corrupted pointer fix
6847 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6848 feature only applies on 64 bit x86 architecture platforms that support AES
6849 NI instructions. A defect in the implementation of "multiblock" can cause
6850 OpenSSL's internal write buffer to become incorrectly set to NULL when
6851 using non-blocking IO. Typically, when the user application is using a
6852 socket BIO for writing, this will only result in a failed connection.
6853 However if some other BIO is used then it is likely that a segmentation
6854 fault will be triggered, thus enabling a potential DoS attack.
6856 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6861 * Segmentation fault in DTLSv1_listen fix
6863 The DTLSv1_listen function is intended to be stateless and processes the
6864 initial ClientHello from many peers. It is common for user code to loop
6865 over the call to DTLSv1_listen until a valid ClientHello is received with
6866 an associated cookie. A defect in the implementation of DTLSv1_listen means
6867 that state is preserved in the SSL object from one invocation to the next
6868 that can lead to a segmentation fault. Errors processing the initial
6869 ClientHello can trigger this scenario. An example of such an error could be
6870 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6873 This issue was reported to OpenSSL by Per Allansson.
6878 * Segmentation fault in ASN1_TYPE_cmp fix
6880 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6881 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6882 certificate signature algorithm consistency this can be used to crash any
6883 certificate verification operation and exploited in a DoS attack. Any
6884 application which performs certificate verification is vulnerable including
6885 OpenSSL clients and servers which enable client authentication.
6890 * Segmentation fault for invalid PSS parameters fix
6892 The signature verification routines will crash with a NULL pointer
6893 dereference if presented with an ASN.1 signature using the RSA PSS
6894 algorithm and invalid parameters. Since these routines are used to verify
6895 certificate signature algorithms this can be used to crash any
6896 certificate verification operation and exploited in a DoS attack. Any
6897 application which performs certificate verification is vulnerable including
6898 OpenSSL clients and servers which enable client authentication.
6900 This issue was was reported to OpenSSL by Brian Carpenter.
6905 * ASN.1 structure reuse memory corruption fix
6907 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6908 memory corruption via an invalid write. Such reuse is and has been
6909 strongly discouraged and is believed to be rare.
6911 Applications that parse structures containing CHOICE or ANY DEFINED BY
6912 components may be affected. Certificate parsing (d2i_X509 and related
6913 functions) are however not affected. OpenSSL clients and servers are
6919 * PKCS7 NULL pointer dereferences fix
6921 The PKCS#7 parsing code does not handle missing outer ContentInfo
6922 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6923 missing content and trigger a NULL pointer dereference on parsing.
6925 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6926 otherwise parse PKCS#7 structures from untrusted sources are
6927 affected. OpenSSL clients and servers are not affected.
6929 This issue was reported to OpenSSL by Michal Zalewski (Google).
6934 * DoS via reachable assert in SSLv2 servers fix
6936 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6937 servers that both support SSLv2 and enable export cipher suites by sending
6938 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6940 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6941 (OpenSSL development team).
6946 * Empty CKE with client auth and DHE fix
6948 If client auth is used then a server can seg fault in the event of a DHE
6949 ciphersuite being selected and a zero length ClientKeyExchange message
6950 being sent by the client. This could be exploited in a DoS attack.
6955 * Handshake with unseeded PRNG fix
6957 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6958 with an unseeded PRNG. The conditions are:
6959 - The client is on a platform where the PRNG has not been seeded
6960 automatically, and the user has not seeded manually
6961 - A protocol specific client method version has been used (i.e. not
6962 SSL_client_methodv23)
6963 - A ciphersuite is used that does not require additional random data from
6964 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6966 If the handshake succeeds then the client random that has been used will
6967 have been generated from a PRNG with insufficient entropy and therefore the
6968 output may be predictable.
6970 For example using the following command with an unseeded openssl will
6971 succeed on an unpatched platform:
6973 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6978 * Use After Free following d2i_ECPrivatekey error fix
6980 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6981 could cause a use after free condition. This, in turn, could cause a double
6982 free in several private key parsing functions (such as d2i_PrivateKey
6983 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6984 for applications that receive EC private keys from untrusted
6985 sources. This scenario is considered rare.
6987 This issue was discovered by the BoringSSL project and fixed in their
6993 * X509_to_X509_REQ NULL pointer deref fix
6995 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6996 the certificate key is invalid. This function is rarely used in practice.
6998 This issue was discovered by Brian Carpenter.
7003 * Removed the export ciphers from the DEFAULT ciphers
7007 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7009 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7010 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7011 So far those who have to target multiple platforms would compromise
7012 and argue that binary targeting say ARMv5 would still execute on
7013 ARMv8. "Universal" build resolves this compromise by providing
7014 near-optimal performance even on newer platforms.
7018 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7019 (other platforms pending).
7021 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7023 * Add support for the SignedCertificateTimestampList certificate and
7024 OCSP response extensions from RFC6962.
7028 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7029 for corner cases. (Certain input points at infinity could lead to
7030 bogus results, with non-infinity inputs mapped to infinity too.)
7034 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7035 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7036 common cases are optimized and there still is room for further
7037 improvements. Vector Permutation AES for Altivec is also added.
7041 * Add support for little-endian ppc64 Linux target.
7043 *Marcelo Cerri (IBM)*
7045 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7046 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7047 are optimized and there still is room for further improvements.
7048 Both 32- and 64-bit modes are supported.
7050 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7052 * Improved ARMv7 NEON support.
7056 * Support for SPARC Architecture 2011 crypto extensions, first
7057 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7058 SHA256/512, MD5, GHASH and modular exponentiation.
7060 *Andy Polyakov, David Miller*
7062 * Accelerated modular exponentiation for Intel processors, a.k.a.
7065 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7067 * Support for new and upcoming Intel processors, including AVX2,
7068 BMI and SHA ISA extensions. This includes additional "stitched"
7069 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7072 This work was sponsored by Intel Corp.
7076 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7077 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7078 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7082 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7083 this fixes a limitation in previous versions of OpenSSL.
7087 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7088 MGF1 digest and OAEP label.
7092 * Add EVP support for key wrapping algorithms, to avoid problems with
7093 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7094 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7095 algorithms and include tests cases.
7099 * Add functions to allocate and set the fields of an ECDSA_METHOD
7102 *Douglas E. Engert, Steve Henson*
7104 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7105 difference in days and seconds between two tm or ASN1_TIME structures.
7109 * Add -rev test option to s_server to just reverse order of characters
7110 received by client and send back to server. Also prints an abbreviated
7111 summary of the connection parameters.
7115 * New option -brief for s_client and s_server to print out a brief summary
7116 of connection parameters.
7120 * Add callbacks for arbitrary TLS extensions.
7122 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7124 * New option -crl_download in several openssl utilities to download CRLs
7125 from CRLDP extension in certificates.
7129 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7133 * New function X509_CRL_diff to generate a delta CRL from the difference
7134 of two full CRLs. Add support to "crl" utility.
7138 * New functions to set lookup_crls function and to retrieve
7139 X509_STORE from X509_STORE_CTX.
7143 * Print out deprecated issuer and subject unique ID fields in
7148 * Extend OCSP I/O functions so they can be used for simple general purpose
7149 HTTP as well as OCSP. New wrapper function which can be used to download
7150 CRLs using the OCSP API.
7154 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7158 * `SSL_CONF*` functions. These provide a common framework for application
7159 configuration using configuration files or command lines.
7163 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7164 message callback and prints the results. Needs compile time option
7165 "enable-ssl-trace". New options to s_client and s_server to enable
7170 * New ctrl and macro to retrieve supported points extensions.
7171 Print out extension in s_server and s_client.
7175 * New functions to retrieve certificate signature and signature
7180 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7185 * New Suite B modes for TLS code. These use and enforce the requirements
7186 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7187 only use Suite B curves. The Suite B modes can be set by using the
7188 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7192 * New chain verification flags for Suite B levels of security. Check
7193 algorithms are acceptable when flags are set in X509_verify_cert.
7197 * Make tls1_check_chain return a set of flags indicating checks passed
7198 by a certificate chain. Add additional tests to handle client
7199 certificates: checks for matching certificate type and issuer name
7204 * If an attempt is made to use a signature algorithm not in the peer
7205 preference list abort the handshake. If client has no suitable
7206 signature algorithms in response to a certificate request do not
7207 use the certificate.
7211 * If server EC tmp key is not in client preference list abort handshake.
7215 * Add support for certificate stores in CERT structure. This makes it
7216 possible to have different stores per SSL structure or one store in
7217 the parent SSL_CTX. Include distinct stores for certificate chain
7218 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7219 to build and store a certificate chain in CERT structure: returning
7220 an error if the chain cannot be built: this will allow applications
7221 to test if a chain is correctly configured.
7223 Note: if the CERT based stores are not set then the parent SSL_CTX
7224 store is used to retain compatibility with existing behaviour.
7228 * New function ssl_set_client_disabled to set a ciphersuite disabled
7229 mask based on the current session, check mask when sending client
7230 hello and checking the requested ciphersuite.
7234 * New ctrls to retrieve and set certificate types in a certificate
7235 request message. Print out received values in s_client. If certificate
7236 types is not set with custom values set sensible values based on
7237 supported signature algorithms.
7241 * Support for distinct client and server supported signature algorithms.
7245 * Add certificate callback. If set this is called whenever a certificate
7246 is required by client or server. An application can decide which
7247 certificate chain to present based on arbitrary criteria: for example
7248 supported signature algorithms. Add very simple example to s_server.
7249 This fixes many of the problems and restrictions of the existing client
7250 certificate callback: for example you can now clear an existing
7251 certificate and specify the whole chain.
7255 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7256 the certificate can be used for (if anything). Set valid_flags field
7257 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7258 to have similar checks in it.
7260 Add new "cert_flags" field to CERT structure and include a "strict mode".
7261 This enforces some TLS certificate requirements (such as only permitting
7262 certificate signature algorithms contained in the supported algorithms
7263 extension) which some implementations ignore: this option should be used
7264 with caution as it could cause interoperability issues.
7268 * Update and tidy signature algorithm extension processing. Work out
7269 shared signature algorithms based on preferences and peer algorithms
7270 and print them out in s_client and s_server. Abort handshake if no
7271 shared signature algorithms.
7275 * Add new functions to allow customised supported signature algorithms
7276 for SSL and SSL_CTX structures. Add options to s_client and s_server
7281 * New function SSL_certs_clear() to delete all references to certificates
7282 from an SSL structure. Before this once a certificate had been added
7283 it couldn't be removed.
7287 * Integrate hostname, email address and IP address checking with certificate
7288 verification. New verify options supporting checking in openssl utility.
7292 * Fixes and wildcard matching support to hostname and email checking
7293 functions. Add manual page.
7295 *Florian Weimer (Red Hat Product Security Team)*
7297 * New functions to check a hostname email or IP address against a
7298 certificate. Add options x509 utility to print results of checks against
7303 * Fix OCSP checking.
7305 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7307 * Initial experimental support for explicitly trusted non-root CAs.
7308 OpenSSL still tries to build a complete chain to a root but if an
7309 intermediate CA has a trust setting included that is used. The first
7310 setting is used: whether to trust (e.g., -addtrust option to the x509
7315 * Add -trusted_first option which attempts to find certificates in the
7316 trusted store even if an untrusted chain is also supplied.
7320 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7321 platform support for Linux and Android.
7325 * Support for linux-x32, ILP32 environment in x86_64 framework.
7329 * Experimental multi-implementation support for FIPS capable OpenSSL.
7330 When in FIPS mode the approved implementations are used as normal,
7331 when not in FIPS mode the internal unapproved versions are used instead.
7332 This means that the FIPS capable OpenSSL isn't forced to use the
7333 (often lower performance) FIPS implementations outside FIPS mode.
7337 * Transparently support X9.42 DH parameters when calling
7338 PEM_read_bio_DHparameters. This means existing applications can handle
7339 the new parameter format automatically.
7343 * Initial experimental support for X9.42 DH parameter format: mainly
7344 to support use of 'q' parameter for RFC5114 parameters.
7348 * Add DH parameters from RFC5114 including test data to dhtest.
7352 * Support for automatic EC temporary key parameter selection. If enabled
7353 the most preferred EC parameters are automatically used instead of
7354 hardcoded fixed parameters. Now a server just has to call:
7355 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7356 support ECDH and use the most appropriate parameters.
7360 * Enhance and tidy EC curve and point format TLS extension code. Use
7361 static structures instead of allocation if default values are used.
7362 New ctrls to set curves we wish to support and to retrieve shared curves.
7363 Print out shared curves in s_server. New options to s_server and s_client
7364 to set list of supported curves.
7368 * New ctrls to retrieve supported signature algorithms and
7369 supported curve values as an array of NIDs. Extend openssl utility
7370 to print out received values.
7374 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7375 between NIDs and the more common NIST names such as "P-256". Enhance
7376 ecparam utility and ECC method to recognise the NIST names for curves.
7380 * Enhance SSL/TLS certificate chain handling to support different
7381 chains for each certificate instead of one chain in the parent SSL_CTX.
7385 * Support for fixed DH ciphersuite client authentication: where both
7386 server and client use DH certificates with common parameters.
7390 * Support for fixed DH ciphersuites: those requiring DH server
7395 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7397 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7398 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7399 X509_CINF_get_signature were reverted post internal team review.
7404 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7406 * OCSP Status Request extension unbounded memory growth
7408 A malicious client can send an excessively large OCSP Status Request
7409 extension. If that client continually requests renegotiation, sending a
7410 large OCSP Status Request extension each time, then there will be unbounded
7411 memory growth on the server. This will eventually lead to a Denial Of
7412 Service attack through memory exhaustion. Servers with a default
7413 configuration are vulnerable even if they do not support OCSP. Builds using
7414 the "no-ocsp" build time option are not affected.
7416 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7421 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7424 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7430 * OOB write in MDC2_Update()
7432 An overflow can occur in MDC2_Update() either if called directly or
7433 through the EVP_DigestUpdate() function using MDC2. If an attacker
7434 is able to supply very large amounts of input data after a previous
7435 call to EVP_EncryptUpdate() with a partial block then a length check
7436 can overflow resulting in a heap corruption.
7438 The amount of data needed is comparable to SIZE_MAX which is impractical
7441 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7446 * Malformed SHA512 ticket DoS
7448 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7449 DoS attack where a malformed ticket will result in an OOB read which will
7452 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7453 a custom server callback and ticket lookup mechanism.
7455 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7460 * OOB write in BN_bn2dec()
7462 The function BN_bn2dec() does not check the return value of BN_div_word().
7463 This can cause an OOB write if an application uses this function with an
7464 overly large BIGNUM. This could be a problem if an overly large certificate
7465 or CRL is printed out from an untrusted source. TLS is not affected because
7466 record limits will reject an oversized certificate before it is parsed.
7468 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7473 * OOB read in TS_OBJ_print_bio()
7475 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7476 the total length the OID text representation would use and not the amount
7477 of data written. This will result in OOB reads when large OIDs are
7480 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7485 * Pointer arithmetic undefined behaviour
7487 Avoid some undefined pointer arithmetic
7489 A common idiom in the codebase is to check limits in the following manner:
7492 Where "p" points to some malloc'd data of SIZE bytes and
7495 "len" here could be from some externally supplied data (e.g. from a TLS
7498 The rules of C pointer arithmetic are such that "p + len" is only well
7499 defined where len <= SIZE. Therefore, the above idiom is actually
7500 undefined behaviour.
7502 For example this could cause problems if some malloc implementation
7503 provides an address for "p" such that "p + len" actually overflows for
7504 values of len that are too big and therefore p + len < limit.
7506 This issue was reported to OpenSSL by Guido Vranken
7511 * Constant time flag not preserved in DSA signing
7513 Operations in the DSA signing algorithm should run in constant time in
7514 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7515 implementation means that a non-constant time codepath is followed for
7516 certain operations. This has been demonstrated through a cache-timing
7517 attack to be sufficient for an attacker to recover the private DSA key.
7519 This issue was reported by César Pereida (Aalto University), Billy Brumley
7520 (Tampere University of Technology), and Yuval Yarom (The University of
7521 Adelaide and NICTA).
7526 * DTLS buffered message DoS
7528 In a DTLS connection where handshake messages are delivered out-of-order
7529 those messages that OpenSSL is not yet ready to process will be buffered
7530 for later use. Under certain circumstances, a flaw in the logic means that
7531 those messages do not get removed from the buffer even though the handshake
7532 has been completed. An attacker could force up to approx. 15 messages to
7533 remain in the buffer when they are no longer required. These messages will
7534 be cleared when the DTLS connection is closed. The default maximum size for
7535 a message is 100k. Therefore, the attacker could force an additional 1500k
7536 to be consumed per connection. By opening many simultaneous connections an
7537 attacker could cause a DoS attack through memory exhaustion.
7539 This issue was reported to OpenSSL by Quan Luo.
7544 * DTLS replay protection DoS
7546 A flaw in the DTLS replay attack protection mechanism means that records
7547 that arrive for future epochs update the replay protection "window" before
7548 the MAC for the record has been validated. This could be exploited by an
7549 attacker by sending a record for the next epoch (which does not have to
7550 decrypt or have a valid MAC), with a very large sequence number. This means
7551 that all subsequent legitimate packets are dropped causing a denial of
7552 service for a specific DTLS connection.
7554 This issue was reported to OpenSSL by the OCAP audit team.
7559 * Certificate message OOB reads
7561 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7562 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7563 theoretical DoS risk but this has not been observed in practice on common
7566 The messages affected are client certificate, client certificate request
7567 and server certificate. As a result the attack can only be performed
7568 against a client or a server which enables client authentication.
7570 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7575 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7577 * Prevent padding oracle in AES-NI CBC MAC check
7579 A MITM attacker can use a padding oracle attack to decrypt traffic
7580 when the connection uses an AES CBC cipher and the server support
7583 This issue was introduced as part of the fix for Lucky 13 padding
7584 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7585 constant time by making sure that always the same bytes are read and
7586 compared against either the MAC or padding bytes. But it no longer
7587 checked that there was enough data to have both the MAC and padding
7590 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7595 * Fix EVP_EncodeUpdate overflow
7597 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7598 Base64 encoding of binary data. If an attacker is able to supply very large
7599 amounts of input data then a length check can overflow resulting in a heap
7602 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7603 the `PEM_write_bio*` family of functions. These are mainly used within the
7604 OpenSSL command line applications, so any application which processes data
7605 from an untrusted source and outputs it as a PEM file should be considered
7606 vulnerable to this issue. User applications that call these APIs directly
7607 with large amounts of untrusted data may also be vulnerable.
7609 This issue was reported by Guido Vranken.
7614 * Fix EVP_EncryptUpdate overflow
7616 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7617 is able to supply very large amounts of input data after a previous call to
7618 EVP_EncryptUpdate() with a partial block then a length check can overflow
7619 resulting in a heap corruption. Following an analysis of all OpenSSL
7620 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7621 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7622 the first called function after an EVP_EncryptInit(), and therefore that
7623 specific call must be safe. The second form is where the length passed to
7624 EVP_EncryptUpdate() can be seen from the code to be some small value and
7625 therefore there is no possibility of an overflow. Since all instances are
7626 one of these two forms, it is believed that there can be no overflows in
7627 internal code due to this problem. It should be noted that
7628 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7629 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7630 of these calls have also been analysed too and it is believed there are no
7631 instances in internal usage where an overflow could occur.
7633 This issue was reported by Guido Vranken.
7638 * Prevent ASN.1 BIO excessive memory allocation
7640 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7641 a short invalid encoding can casuse allocation of large amounts of memory
7642 potentially consuming excessive resources or exhausting memory.
7644 Any application parsing untrusted data through d2i BIO functions is
7645 affected. The memory based functions such as d2i_X509() are *not* affected.
7646 Since the memory based functions are used by the TLS library, TLS
7647 applications are not affected.
7649 This issue was reported by Brian Carpenter.
7656 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7657 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7658 in arbitrary stack data being returned in the buffer.
7660 This issue was reported by Guido Vranken.
7665 * Modify behavior of ALPN to invoke callback after SNI/servername
7666 callback, such that updates to the SSL_CTX affect ALPN.
7670 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7675 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7676 methods are enabled and ssl2 is disabled the methods return NULL.
7680 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7682 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7683 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7684 provide any "EXPORT" or "LOW" strength ciphers.
7688 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7689 is by default disabled at build-time. Builds that are not configured with
7690 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7691 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7692 will need to explicitly call either of:
7694 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7696 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7698 as appropriate. Even if either of those is used, or the application
7699 explicitly uses the version-specific SSLv2_method() or its client and
7700 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7701 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7702 ciphers, and SSLv2 56-bit DES are no longer available.
7707 * Fix a double-free in DSA code
7709 A double free bug was discovered when OpenSSL parses malformed DSA private
7710 keys and could lead to a DoS attack or memory corruption for applications
7711 that receive DSA private keys from untrusted sources. This scenario is
7714 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7720 * Disable SRP fake user seed to address a server memory leak.
7722 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7724 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7725 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7726 was changed to ignore the "fake user" SRP seed, even if the seed
7729 Users should use SRP_VBASE_get1_by_user instead. Note that in
7730 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7731 also that even though configuring the SRP seed attempts to hide
7732 invalid usernames by continuing the handshake with fake
7733 credentials, this behaviour is not constant time and no strong
7734 guarantees are made that the handshake is indistinguishable from
7735 that of a valid user.
7740 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7742 In the BN_hex2bn function the number of hex digits is calculated using an
7743 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7744 large values of `i` this can result in `bn_expand` not allocating any
7745 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7746 field as NULL leading to a subsequent NULL ptr deref. For very large values
7747 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7748 In this case memory is allocated to the internal BIGNUM data field, but it
7749 is insufficiently sized leading to heap corruption. A similar issue exists
7750 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7751 is ever called by user applications with very large untrusted hex/dec data.
7752 This is anticipated to be a rare occurrence.
7754 All OpenSSL internal usage of these functions use data that is not expected
7755 to be untrusted, e.g. config file data or application command line
7756 arguments. If user developed applications generate config file data based
7757 on untrusted data then it is possible that this could also lead to security
7758 consequences. This is also anticipated to be rare.
7760 This issue was reported to OpenSSL by Guido Vranken.
7765 * Fix memory issues in `BIO_*printf` functions
7767 The internal `fmtstr` function used in processing a "%s" format string in
7768 the `BIO_*printf` functions could overflow while calculating the length of a
7769 string and cause an OOB read when printing very long strings.
7771 Additionally the internal `doapr_outch` function can attempt to write to an
7772 OOB memory location (at an offset from the NULL pointer) in the event of a
7773 memory allocation failure. In 1.0.2 and below this could be caused where
7774 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7775 could be in processing a very long "%s" format string. Memory leaks can
7778 The first issue may mask the second issue dependent on compiler behaviour.
7779 These problems could enable attacks where large amounts of untrusted data
7780 is passed to the `BIO_*printf` functions. If applications use these functions
7781 in this way then they could be vulnerable. OpenSSL itself uses these
7782 functions when printing out human-readable dumps of ASN.1 data. Therefore
7783 applications that print this data could be vulnerable if the data is from
7784 untrusted sources. OpenSSL command line applications could also be
7785 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7786 as command line arguments.
7788 Libssl is not considered directly vulnerable. Additionally certificates etc
7789 received via remote connections via libssl are also unlikely to be able to
7790 trigger these issues because of message size limits enforced within libssl.
7792 This issue was reported to OpenSSL Guido Vranken.
7797 * Side channel attack on modular exponentiation
7799 A side-channel attack was found which makes use of cache-bank conflicts on
7800 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7801 of RSA keys. The ability to exploit this issue is limited as it relies on
7802 an attacker who has control of code in a thread running on the same
7803 hyper-threaded core as the victim thread which is performing decryptions.
7805 This issue was reported to OpenSSL by Yuval Yarom, The University of
7806 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7807 Nadia Heninger, University of Pennsylvania with more information at
7808 <http://cachebleed.info>.
7813 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7814 if no keysize is specified with default_bits. This fixes an
7815 omission in an earlier change that changed all RSA/DSA key generation
7816 commands to use 2048 bits by default.
7820 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7822 * Protection for DH small subgroup attacks
7824 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7825 switched on by default and cannot be disabled. This could have some
7830 * SSLv2 doesn't block disabled ciphers
7832 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7833 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7834 been disabled, provided that the SSLv2 protocol was not also disabled via
7837 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7838 and Sebastian Schinzel.
7843 * Reject DH handshakes with parameters shorter than 1024 bits.
7847 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7849 * Certificate verify crash with missing PSS parameter
7851 The signature verification routines will crash with a NULL pointer
7852 dereference if presented with an ASN.1 signature using the RSA PSS
7853 algorithm and absent mask generation function parameter. Since these
7854 routines are used to verify certificate signature algorithms this can be
7855 used to crash any certificate verification operation and exploited in a
7856 DoS attack. Any application which performs certificate verification is
7857 vulnerable including OpenSSL clients and servers which enable client
7860 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7865 * X509_ATTRIBUTE memory leak
7867 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7868 memory. This structure is used by the PKCS#7 and CMS routines so any
7869 application which reads PKCS#7 or CMS data from untrusted sources is
7870 affected. SSL/TLS is not affected.
7872 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7878 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7879 This changes the decoding behaviour for some invalid messages,
7880 though the change is mostly in the more lenient direction, and
7881 legacy behaviour is preserved as much as possible.
7885 * In DSA_generate_parameters_ex, if the provided seed is too short,
7886 use a random seed, as already documented.
7888 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7890 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7892 * Alternate chains certificate forgery
7894 During certificate verification, OpenSSL will attempt to find an
7895 alternative certificate chain if the first attempt to build such a chain
7896 fails. An error in the implementation of this logic can mean that an
7897 attacker could cause certain checks on untrusted certificates to be
7898 bypassed, such as the CA flag, enabling them to use a valid leaf
7899 certificate to act as a CA and "issue" an invalid certificate.
7901 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7907 * Race condition handling PSK identify hint
7909 If PSK identity hints are received by a multi-threaded client then
7910 the values are wrongly updated in the parent SSL_CTX structure. This can
7911 result in a race condition potentially leading to a double free of the
7917 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7919 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7920 incompatibility in the handling of HMAC. The previous ABI has now been
7923 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7925 * Malformed ECParameters causes infinite loop
7927 When processing an ECParameters structure OpenSSL enters an infinite loop
7928 if the curve specified is over a specially malformed binary polynomial
7931 This can be used to perform denial of service against any
7932 system which processes public keys, certificate requests or
7933 certificates. This includes TLS clients and TLS servers with
7934 client authentication enabled.
7936 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7941 * Exploitable out-of-bounds read in X509_cmp_time
7943 X509_cmp_time does not properly check the length of the ASN1_TIME
7944 string and can read a few bytes out of bounds. In addition,
7945 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7948 An attacker can use this to craft malformed certificates and CRLs of
7949 various sizes and potentially cause a segmentation fault, resulting in
7950 a DoS on applications that verify certificates or CRLs. TLS clients
7951 that verify CRLs are affected. TLS clients and servers with client
7952 authentication enabled may be affected if they use custom verification
7955 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7956 independently by Hanno Böck.
7961 * PKCS7 crash with missing EnvelopedContent
7963 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7964 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7965 with missing content and trigger a NULL pointer dereference on parsing.
7967 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7968 structures from untrusted sources are affected. OpenSSL clients and
7969 servers are not affected.
7971 This issue was reported to OpenSSL by Michal Zalewski (Google).
7976 * CMS verify infinite loop with unknown hash function
7978 When verifying a signedData message the CMS code can enter an infinite loop
7979 if presented with an unknown hash function OID. This can be used to perform
7980 denial of service against any system which verifies signedData messages using
7982 This issue was reported to OpenSSL by Johannes Bauer.
7987 * Race condition handling NewSessionTicket
7989 If a NewSessionTicket is received by a multi-threaded client when attempting to
7990 reuse a previous ticket then a race condition can occur potentially leading to
7991 a double free of the ticket data.
7996 * Reject DH handshakes with parameters shorter than 768 bits.
7998 *Kurt Roeckx and Emilia Kasper*
8000 * dhparam: generate 2048-bit parameters by default.
8002 *Kurt Roeckx and Emilia Kasper*
8004 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8006 * Segmentation fault in ASN1_TYPE_cmp fix
8008 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8009 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8010 certificate signature algorithm consistency this can be used to crash any
8011 certificate verification operation and exploited in a DoS attack. Any
8012 application which performs certificate verification is vulnerable including
8013 OpenSSL clients and servers which enable client authentication.
8018 * ASN.1 structure reuse memory corruption fix
8020 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8021 memory corruption via an invalid write. Such reuse is and has been
8022 strongly discouraged and is believed to be rare.
8024 Applications that parse structures containing CHOICE or ANY DEFINED BY
8025 components may be affected. Certificate parsing (d2i_X509 and related
8026 functions) are however not affected. OpenSSL clients and servers are
8032 * PKCS7 NULL pointer dereferences fix
8034 The PKCS#7 parsing code does not handle missing outer ContentInfo
8035 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8036 missing content and trigger a NULL pointer dereference on parsing.
8038 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8039 otherwise parse PKCS#7 structures from untrusted sources are
8040 affected. OpenSSL clients and servers are not affected.
8042 This issue was reported to OpenSSL by Michal Zalewski (Google).
8047 * DoS via reachable assert in SSLv2 servers fix
8049 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8050 servers that both support SSLv2 and enable export cipher suites by sending
8051 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8053 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8054 (OpenSSL development team).
8059 * Use After Free following d2i_ECPrivatekey error fix
8061 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8062 could cause a use after free condition. This, in turn, could cause a double
8063 free in several private key parsing functions (such as d2i_PrivateKey
8064 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8065 for applications that receive EC private keys from untrusted
8066 sources. This scenario is considered rare.
8068 This issue was discovered by the BoringSSL project and fixed in their
8074 * X509_to_X509_REQ NULL pointer deref fix
8076 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8077 the certificate key is invalid. This function is rarely used in practice.
8079 This issue was discovered by Brian Carpenter.
8084 * Removed the export ciphers from the DEFAULT ciphers
8088 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8090 * Build fixes for the Windows and OpenVMS platforms
8092 *Matt Caswell and Richard Levitte*
8094 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8096 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8097 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8098 dereference. This could lead to a Denial Of Service attack. Thanks to
8099 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8104 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8105 dtls1_buffer_record function under certain conditions. In particular this
8106 could occur if an attacker sent repeated DTLS records with the same
8107 sequence number but for the next epoch. The memory leak could be exploited
8108 by an attacker in a Denial of Service attack through memory exhaustion.
8109 Thanks to Chris Mueller for reporting this issue.
8114 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8115 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8116 method would be set to NULL which could later result in a NULL pointer
8117 dereference. Thanks to Frank Schmirler for reporting this issue.
8122 * Abort handshake if server key exchange message is omitted for ephemeral
8125 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8126 reporting this issue.
8131 * Remove non-export ephemeral RSA code on client and server. This code
8132 violated the TLS standard by allowing the use of temporary RSA keys in
8133 non-export ciphersuites and could be used by a server to effectively
8134 downgrade the RSA key length used to a value smaller than the server
8135 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8136 INRIA or reporting this issue.
8141 * Fixed issue where DH client certificates are accepted without verification.
8142 An OpenSSL server will accept a DH certificate for client authentication
8143 without the certificate verify message. This effectively allows a client to
8144 authenticate without the use of a private key. This only affects servers
8145 which trust a client certificate authority which issues certificates
8146 containing DH keys: these are extremely rare and hardly ever encountered.
8147 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8153 * Ensure that the session ID context of an SSL is updated when its
8154 SSL_CTX is updated via SSL_set_SSL_CTX.
8156 The session ID context is typically set from the parent SSL_CTX,
8157 and can vary with the CTX.
8161 * Fix various certificate fingerprint issues.
8163 By using non-DER or invalid encodings outside the signed portion of a
8164 certificate the fingerprint can be changed without breaking the signature.
8165 Although no details of the signed portion of the certificate can be changed
8166 this can cause problems with some applications: e.g. those using the
8167 certificate fingerprint for blacklists.
8169 1. Reject signatures with non zero unused bits.
8171 If the BIT STRING containing the signature has non zero unused bits reject
8172 the signature. All current signature algorithms require zero unused bits.
8174 2. Check certificate algorithm consistency.
8176 Check the AlgorithmIdentifier inside TBS matches the one in the
8177 certificate signature. NB: this will result in signature failure
8178 errors for some broken certificates.
8180 Thanks to Konrad Kraszewski from Google for reporting this issue.
8182 3. Check DSA/ECDSA signatures use DER.
8184 Re-encode DSA/ECDSA signatures and compare with the original received
8185 signature. Return an error if there is a mismatch.
8187 This will reject various cases including garbage after signature
8188 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8189 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8190 (negative or with leading zeroes).
8192 Further analysis was conducted and fixes were developed by Stephen Henson
8193 of the OpenSSL core team.
8199 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8200 results on some platforms, including x86_64. This bug occurs at random
8201 with a very low probability, and is not known to be exploitable in any
8202 way, though its exact impact is difficult to determine. Thanks to Pieter
8203 Wuille (Blockstream) who reported this issue and also suggested an initial
8204 fix. Further analysis was conducted by the OpenSSL development team and
8205 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8206 the OpenSSL core team.
8211 * Do not resume sessions on the server if the negotiated protocol
8212 version does not match the session's version. Resuming with a different
8213 version, while not strictly forbidden by the RFC, is of questionable
8214 sanity and breaks all known clients.
8216 *David Benjamin, Emilia Käsper*
8218 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8219 early CCS messages during renegotiation. (Note that because
8220 renegotiation is encrypted, this early CCS was not exploitable.)
8224 * Tighten client-side session ticket handling during renegotiation:
8225 ensure that the client only accepts a session ticket if the server sends
8226 the extension anew in the ServerHello. Previously, a TLS client would
8227 reuse the old extension state and thus accept a session ticket if one was
8228 announced in the initial ServerHello.
8230 Similarly, ensure that the client requires a session ticket if one
8231 was advertised in the ServerHello. Previously, a TLS client would
8232 ignore a missing NewSessionTicket message.
8236 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8240 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8241 sends a carefully crafted handshake message, to cause OpenSSL to fail
8242 to free up to 64k of memory causing a memory leak. This could be
8243 exploited in a Denial Of Service attack. This issue affects OpenSSL
8244 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8245 whether SRTP is used or configured. Implementations of OpenSSL that
8246 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8248 The fix was developed by the OpenSSL team.
8253 * Session Ticket Memory Leak.
8255 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8256 integrity of that ticket is first verified. In the event of a session
8257 ticket integrity check failing, OpenSSL will fail to free memory
8258 causing a memory leak. By sending a large number of invalid session
8259 tickets an attacker could exploit this issue in a Denial Of Service
8265 * Build option no-ssl3 is incomplete.
8267 When OpenSSL is configured with "no-ssl3" as a build option, servers
8268 could accept and complete an SSL 3.0 handshake, and clients could be
8269 configured to send them.
8272 *Akamai and the OpenSSL team*
8274 * Add support for TLS_FALLBACK_SCSV.
8275 Client applications doing fallback retries should call
8276 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8279 *Adam Langley, Bodo Moeller*
8281 * Add additional DigestInfo checks.
8283 Re-encode DigestInto in DER and check against the original when
8284 verifying RSA signature: this will reject any improperly encoded
8285 DigestInfo structures.
8287 Note: this is a precautionary measure and no attacks are currently known.
8291 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8293 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8294 SRP code can be overrun an internal buffer. Add sanity check that
8295 g, A, B < N to SRP code.
8297 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8298 Group for discovering this issue.
8303 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8304 TLS 1.0 instead of higher protocol versions when the ClientHello message
8305 is badly fragmented. This allows a man-in-the-middle attacker to force a
8306 downgrade to TLS 1.0 even if both the server and the client support a
8307 higher protocol version, by modifying the client's TLS records.
8309 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8310 researching this issue.
8315 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8316 to a denial of service attack. A malicious server can crash the client
8317 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8318 ciphersuite and sending carefully crafted handshake messages.
8320 Thanks to Felix Gröbert (Google) for discovering and researching this
8326 * By sending carefully crafted DTLS packets an attacker could cause openssl
8327 to leak memory. This can be exploited through a Denial of Service attack.
8328 Thanks to Adam Langley for discovering and researching this issue.
8333 * An attacker can force openssl to consume large amounts of memory whilst
8334 processing DTLS handshake messages. This can be exploited through a
8335 Denial of Service attack.
8336 Thanks to Adam Langley for discovering and researching this issue.
8341 * An attacker can force an error condition which causes openssl to crash
8342 whilst processing DTLS packets due to memory being freed twice. This
8343 can be exploited through a Denial of Service attack.
8344 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8350 * If a multithreaded client connects to a malicious server using a resumed
8351 session and the server sends an ec point format extension it could write
8352 up to 255 bytes to freed memory.
8354 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8360 * A malicious server can crash an OpenSSL client with a null pointer
8361 dereference (read) by specifying an SRP ciphersuite even though it was not
8362 properly negotiated with the client. This can be exploited through a
8363 Denial of Service attack.
8365 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8366 discovering and researching this issue.
8371 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8372 X509_name_oneline, X509_name_print_ex et al. to leak some information
8373 from the stack. Applications may be affected if they echo pretty printing
8374 output to the attacker.
8376 Thanks to Ivan Fratric (Google) for discovering this issue.
8379 *Emilia Käsper, and Steve Henson*
8381 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8382 for corner cases. (Certain input points at infinity could lead to
8383 bogus results, with non-infinity inputs mapped to infinity too.)
8387 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8389 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8390 handshake can force the use of weak keying material in OpenSSL
8391 SSL/TLS clients and servers.
8393 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8394 researching this issue. ([CVE-2014-0224])
8396 *KIKUCHI Masashi, Steve Henson*
8398 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8399 OpenSSL DTLS client the code can be made to recurse eventually crashing
8402 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8405 *Imre Rad, Steve Henson*
8407 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8408 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8409 client or server. This is potentially exploitable to run arbitrary
8410 code on a vulnerable client or server.
8412 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8414 *Jüri Aedla, Steve Henson*
8416 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8417 are subject to a denial of service attack.
8419 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8420 this issue. ([CVE-2014-3470])
8422 *Felix Gröbert, Ivan Fratric, Steve Henson*
8424 * Harmonize version and its documentation. -f flag is used to display
8427 *mancha <mancha1@zoho.com>*
8429 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8430 in i2d_ECPrivateKey.
8432 *mancha <mancha1@zoho.com>*
8434 * Fix some double frees. These are not thought to be exploitable.
8436 *mancha <mancha1@zoho.com>*
8438 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8440 * A missing bounds check in the handling of the TLS heartbeat extension
8441 can be used to reveal up to 64k of memory to a connected client or
8444 Thanks for Neel Mehta of Google Security for discovering this bug and to
8445 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8446 preparing the fix ([CVE-2014-0160])
8448 *Adam Langley, Bodo Moeller*
8450 * Fix for the attack described in the paper "Recovering OpenSSL
8451 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8452 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8453 <http://eprint.iacr.org/2014/140>
8455 Thanks to Yuval Yarom and Naomi Benger for discovering this
8456 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8458 *Yuval Yarom and Naomi Benger*
8460 * TLS pad extension: draft-agl-tls-padding-03
8462 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8463 TLS client Hello record length value would otherwise be > 255 and
8464 less that 512 pad with a dummy extension containing zeroes so it
8465 is at least 512 bytes long.
8467 *Adam Langley, Steve Henson*
8469 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8471 * Fix for TLS record tampering bug. A carefully crafted invalid
8472 handshake could crash OpenSSL with a NULL pointer exception.
8473 Thanks to Anton Johansson for reporting this issues.
8476 * Keep original DTLS digest and encryption contexts in retransmission
8477 structures so we can use the previous session parameters if they need
8478 to be resent. ([CVE-2013-6450])
8482 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8483 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8484 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8485 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8486 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8487 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8489 *Rob Stradling, Adam Langley*
8491 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8493 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8494 supporting platforms or when small records were transferred.
8496 *Andy Polyakov, Steve Henson*
8498 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8500 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8502 This addresses the flaw in CBC record processing discovered by
8503 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8504 at: <http://www.isg.rhul.ac.uk/tls/>
8506 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8507 Security Group at Royal Holloway, University of London
8508 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8509 Emilia Käsper for the initial patch.
8512 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8514 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8515 ciphersuites which can be exploited in a denial of service attack.
8516 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8517 and detecting this bug and to Wolfgang Ettlinger
8518 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8523 * Return an error when checking OCSP signatures when key is NULL.
8524 This fixes a DoS attack. ([CVE-2013-0166])
8528 * Make openssl verify return errors.
8530 *Chris Palmer <palmer@google.com> and Ben Laurie*
8532 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8533 the right response is stapled. Also change SSL_get_certificate()
8534 so it returns the certificate actually sent.
8535 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8537 *Rob Stradling <rob.stradling@comodo.com>*
8539 * Fix possible deadlock when decoding public keys.
8543 * Don't use TLS 1.0 record version number in initial client hello
8548 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8550 * Sanity check record length before skipping explicit IV in TLS
8551 1.2, 1.1 and DTLS to fix DoS attack.
8553 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8554 fuzzing as a service testing platform.
8559 * Initialise tkeylen properly when encrypting CMS messages.
8560 Thanks to Solar Designer of Openwall for reporting this issue.
8564 * In FIPS mode don't try to use composite ciphers as they are not
8569 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8571 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8572 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8573 mean any application compiled against OpenSSL 1.0.0 headers setting
8574 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8575 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8576 0x10000000L Any application which was previously compiled against
8577 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8578 will need to be recompiled as a result. Letting be results in
8579 inability to disable specifically TLS 1.1 and in client context,
8580 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8584 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8585 disable just protocol X, but all protocols above X *if* there are
8586 protocols *below* X still enabled. In more practical terms it means
8587 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8588 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8589 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8594 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8596 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8597 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8598 in CRYPTO_realloc_clean.
8600 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8601 issue and to Adam Langley <agl@chromium.org> for fixing it.
8604 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8606 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8610 * Workarounds for some broken servers that "hang" if a client hello
8611 record length exceeds 255 bytes.
8613 1. Do not use record version number > TLS 1.0 in initial client
8614 hello: some (but not all) hanging servers will now work.
8615 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8616 the number of ciphers sent in the client hello. This should be
8617 set to an even number, such as 50, for example by passing:
8618 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8619 Most broken servers should now work.
8620 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8621 TLS 1.2 client support entirely.
8625 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8629 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8631 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8632 STRING form instead of a DigestInfo.
8636 * The format used for MDC2 RSA signatures is inconsistent between EVP
8637 and the RSA_sign/RSA_verify functions. This was made more apparent when
8638 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8639 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8640 the correct format in RSA_verify so both forms transparently work.
8644 * Some servers which support TLS 1.0 can choke if we initially indicate
8645 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8646 encrypted premaster secret. As a workaround use the maximum permitted
8647 client version in client hello, this should keep such servers happy
8648 and still work with previous versions of OpenSSL.
8652 * Add support for TLS/DTLS heartbeats.
8654 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8656 * Add support for SCTP.
8658 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8660 * Improved PRNG seeding for VOS.
8662 *Paul Green <Paul.Green@stratus.com>*
8664 * Extensive assembler packs updates, most notably:
8666 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8667 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8668 - x86_64: bit-sliced AES implementation;
8669 - ARM: NEON support, contemporary platforms optimizations;
8670 - s390x: z196 support;
8671 - `*`: GHASH and GF(2^m) multiplication implementations;
8675 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8676 (removal of unnecessary code)
8678 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8680 * Add TLS key material exporter from RFC 5705.
8684 * Add DTLS-SRTP negotiation from RFC 5764.
8688 * Add Next Protocol Negotiation,
8689 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8690 disabled with a no-npn flag to config or Configure. Code donated
8693 *Adam Langley <agl@google.com> and Ben Laurie*
8695 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8696 NIST-P256, NIST-P521, with constant-time single point multiplication on
8697 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8698 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8699 Code made available under Apache License version 2.0.
8701 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8702 line to include this in your build of OpenSSL, and run "make depend" (or
8703 "make update"). This enables the following EC_METHODs:
8705 EC_GFp_nistp224_method()
8706 EC_GFp_nistp256_method()
8707 EC_GFp_nistp521_method()
8709 EC_GROUP_new_by_curve_name() will automatically use these (while
8710 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8713 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8715 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8716 all platforms. Move ssize_t definition from e_os.h to the public
8717 header file e_os2.h as it now appears in public header file cms.h
8721 * New -sigopt option to the ca, req and x509 utilities. Additional
8722 signature parameters can be passed using this option and in
8727 * Add RSA PSS signing function. This will generate and set the
8728 appropriate AlgorithmIdentifiers for PSS based on those in the
8729 corresponding EVP_MD_CTX structure. No application support yet.
8733 * Support for companion algorithm specific ASN1 signing routines.
8734 New function ASN1_item_sign_ctx() signs a pre-initialised
8735 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8736 the appropriate parameters.
8740 * Add new algorithm specific ASN1 verification initialisation function
8741 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8742 handling will be the same no matter what EVP_PKEY_METHOD is used.
8743 Add a PSS handler to support verification of PSS signatures: checked
8744 against a number of sample certificates.
8748 * Add signature printing for PSS. Add PSS OIDs.
8750 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8752 * Add algorithm specific signature printing. An individual ASN1 method
8753 can now print out signatures instead of the standard hex dump.
8755 More complex signatures (e.g. PSS) can print out more meaningful
8756 information. Include DSA version that prints out the signature
8761 * Password based recipient info support for CMS library: implementing
8766 * Split password based encryption into PBES2 and PBKDF2 functions. This
8767 neatly separates the code into cipher and PBE sections and is required
8768 for some algorithms that split PBES2 into separate pieces (such as
8769 password based CMS).
8773 * Session-handling fixes:
8774 - Fix handling of connections that are resuming with a session ID,
8775 but also support Session Tickets.
8776 - Fix a bug that suppressed issuing of a new ticket if the client
8777 presented a ticket with an expired session.
8778 - Try to set the ticket lifetime hint to something reasonable.
8779 - Make tickets shorter by excluding irrelevant information.
8780 - On the client side, don't ignore renewed tickets.
8782 *Adam Langley, Bodo Moeller (Google)*
8784 * Fix PSK session representation.
8788 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8790 This work was sponsored by Intel.
8794 * Add GCM support to TLS library. Some custom code is needed to split
8795 the IV between the fixed (from PRF) and explicit (from TLS record)
8796 portions. This adds all GCM ciphersuites supported by RFC5288 and
8797 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8798 add a special AESGCM string for GCM only.
8802 * Expand range of ctrls for AES GCM. Permit setting invocation
8803 field on decrypt and retrieval of invocation field only on encrypt.
8807 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8808 As required by RFC5289 these ciphersuites cannot be used if for
8809 versions of TLS earlier than 1.2.
8813 * For FIPS capable OpenSSL interpret a NULL default public key method
8814 as unset and return the appropriate default but do *not* set the default.
8815 This means we can return the appropriate method in applications that
8816 switch between FIPS and non-FIPS modes.
8820 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8821 ENGINE is used then we cannot handle that in the FIPS module so we
8822 keep original code iff non-FIPS operations are allowed.
8826 * Add -attime option to openssl utilities.
8828 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8830 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8834 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8835 FIPS EC methods unconditionally for now.
8839 * New build option no-ec2m to disable characteristic 2 code.
8843 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8844 all cases can be covered as some introduce binary incompatibilities.
8848 * Redirect RSA operations to FIPS module including keygen,
8849 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8853 * Add similar low-level API blocking to ciphers.
8857 * low-level digest APIs are not approved in FIPS mode: any attempt
8858 to use these will cause a fatal error. Applications that *really* want
8859 to use them can use the `private_*` version instead.
8863 * Redirect cipher operations to FIPS module for FIPS builds.
8867 * Redirect digest operations to FIPS module for FIPS builds.
8871 * Update build system to add "fips" flag which will link in fipscanister.o
8872 for static and shared library builds embedding a signature if needed.
8876 * Output TLS supported curves in preference order instead of numerical
8877 order. This is currently hardcoded for the highest order curves first.
8878 This should be configurable so applications can judge speed vs strength.
8882 * Add TLS v1.2 server support for client authentication.
8886 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8891 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8892 FIPS modules versions.
8896 * Add TLS v1.2 client side support for client authentication. Keep cache
8897 of handshake records longer as we don't know the hash algorithm to use
8898 until after the certificate request message is received.
8902 * Initial TLS v1.2 client support. Add a default signature algorithms
8903 extension including all the algorithms we support. Parse new signature
8904 format in client key exchange. Relax some ECC signing restrictions for
8905 TLS v1.2 as indicated in RFC5246.
8909 * Add server support for TLS v1.2 signature algorithms extension. Switch
8910 to new signature format when needed using client digest preference.
8911 All server ciphersuites should now work correctly in TLS v1.2. No client
8912 support yet and no support for client certificates.
8916 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8917 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8918 ciphersuites. At present only RSA key exchange ciphersuites work with
8919 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8920 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8921 and version checking.
8925 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8926 with this defined it will not be affected by any changes to ssl internal
8927 structures. Add several utility functions to allow openssl application
8928 to work with OPENSSL_NO_SSL_INTERN defined.
8932 * A long standing patch to add support for SRP from EdelWeb (Peter
8933 Sylvester and Christophe Renou) was integrated.
8934 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8935 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8938 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8942 * Permit abbreviated handshakes when renegotiating using the function
8943 SSL_renegotiate_abbreviated().
8945 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8947 * Add call to ENGINE_register_all_complete() to
8948 ENGINE_load_builtin_engines(), so some implementations get used
8949 automatically instead of needing explicit application support.
8953 * Add support for TLS key exporter as described in RFC5705.
8955 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8957 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8958 a few changes are required:
8960 Add SSL_OP_NO_TLSv1_1 flag.
8961 Add TLSv1_1 methods.
8962 Update version checking logic to handle version 1.1.
8963 Add explicit IV handling (ported from DTLS code).
8964 Add command line options to s_client/s_server.
8971 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8973 * X509_ATTRIBUTE memory leak
8975 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8976 memory. This structure is used by the PKCS#7 and CMS routines so any
8977 application which reads PKCS#7 or CMS data from untrusted sources is
8978 affected. SSL/TLS is not affected.
8980 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8986 * Race condition handling PSK identify hint
8988 If PSK identity hints are received by a multi-threaded client then
8989 the values are wrongly updated in the parent SSL_CTX structure. This can
8990 result in a race condition potentially leading to a double free of the
8996 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8998 * Malformed ECParameters causes infinite loop
9000 When processing an ECParameters structure OpenSSL enters an infinite loop
9001 if the curve specified is over a specially malformed binary polynomial
9004 This can be used to perform denial of service against any
9005 system which processes public keys, certificate requests or
9006 certificates. This includes TLS clients and TLS servers with
9007 client authentication enabled.
9009 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9014 * Exploitable out-of-bounds read in X509_cmp_time
9016 X509_cmp_time does not properly check the length of the ASN1_TIME
9017 string and can read a few bytes out of bounds. In addition,
9018 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9021 An attacker can use this to craft malformed certificates and CRLs of
9022 various sizes and potentially cause a segmentation fault, resulting in
9023 a DoS on applications that verify certificates or CRLs. TLS clients
9024 that verify CRLs are affected. TLS clients and servers with client
9025 authentication enabled may be affected if they use custom verification
9028 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9029 independently by Hanno Böck.
9034 * PKCS7 crash with missing EnvelopedContent
9036 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9037 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9038 with missing content and trigger a NULL pointer dereference on parsing.
9040 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9041 structures from untrusted sources are affected. OpenSSL clients and
9042 servers are not affected.
9044 This issue was reported to OpenSSL by Michal Zalewski (Google).
9049 * CMS verify infinite loop with unknown hash function
9051 When verifying a signedData message the CMS code can enter an infinite loop
9052 if presented with an unknown hash function OID. This can be used to perform
9053 denial of service against any system which verifies signedData messages using
9055 This issue was reported to OpenSSL by Johannes Bauer.
9060 * Race condition handling NewSessionTicket
9062 If a NewSessionTicket is received by a multi-threaded client when attempting to
9063 reuse a previous ticket then a race condition can occur potentially leading to
9064 a double free of the ticket data.
9069 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9071 * Segmentation fault in ASN1_TYPE_cmp fix
9073 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9074 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9075 certificate signature algorithm consistency this can be used to crash any
9076 certificate verification operation and exploited in a DoS attack. Any
9077 application which performs certificate verification is vulnerable including
9078 OpenSSL clients and servers which enable client authentication.
9083 * ASN.1 structure reuse memory corruption fix
9085 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9086 memory corruption via an invalid write. Such reuse is and has been
9087 strongly discouraged and is believed to be rare.
9089 Applications that parse structures containing CHOICE or ANY DEFINED BY
9090 components may be affected. Certificate parsing (d2i_X509 and related
9091 functions) are however not affected. OpenSSL clients and servers are
9097 * PKCS7 NULL pointer dereferences fix
9099 The PKCS#7 parsing code does not handle missing outer ContentInfo
9100 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9101 missing content and trigger a NULL pointer dereference on parsing.
9103 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9104 otherwise parse PKCS#7 structures from untrusted sources are
9105 affected. OpenSSL clients and servers are not affected.
9107 This issue was reported to OpenSSL by Michal Zalewski (Google).
9112 * DoS via reachable assert in SSLv2 servers fix
9114 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9115 servers that both support SSLv2 and enable export cipher suites by sending
9116 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9118 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9119 (OpenSSL development team).
9124 * Use After Free following d2i_ECPrivatekey error fix
9126 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9127 could cause a use after free condition. This, in turn, could cause a double
9128 free in several private key parsing functions (such as d2i_PrivateKey
9129 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9130 for applications that receive EC private keys from untrusted
9131 sources. This scenario is considered rare.
9133 This issue was discovered by the BoringSSL project and fixed in their
9139 * X509_to_X509_REQ NULL pointer deref fix
9141 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9142 the certificate key is invalid. This function is rarely used in practice.
9144 This issue was discovered by Brian Carpenter.
9149 * Removed the export ciphers from the DEFAULT ciphers
9153 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9155 * Build fixes for the Windows and OpenVMS platforms
9157 *Matt Caswell and Richard Levitte*
9159 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9161 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9162 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9163 dereference. This could lead to a Denial Of Service attack. Thanks to
9164 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9169 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9170 dtls1_buffer_record function under certain conditions. In particular this
9171 could occur if an attacker sent repeated DTLS records with the same
9172 sequence number but for the next epoch. The memory leak could be exploited
9173 by an attacker in a Denial of Service attack through memory exhaustion.
9174 Thanks to Chris Mueller for reporting this issue.
9179 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9180 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9181 method would be set to NULL which could later result in a NULL pointer
9182 dereference. Thanks to Frank Schmirler for reporting this issue.
9187 * Abort handshake if server key exchange message is omitted for ephemeral
9190 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9191 reporting this issue.
9196 * Remove non-export ephemeral RSA code on client and server. This code
9197 violated the TLS standard by allowing the use of temporary RSA keys in
9198 non-export ciphersuites and could be used by a server to effectively
9199 downgrade the RSA key length used to a value smaller than the server
9200 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9201 INRIA or reporting this issue.
9206 * Fixed issue where DH client certificates are accepted without verification.
9207 An OpenSSL server will accept a DH certificate for client authentication
9208 without the certificate verify message. This effectively allows a client to
9209 authenticate without the use of a private key. This only affects servers
9210 which trust a client certificate authority which issues certificates
9211 containing DH keys: these are extremely rare and hardly ever encountered.
9212 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9218 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9219 results on some platforms, including x86_64. This bug occurs at random
9220 with a very low probability, and is not known to be exploitable in any
9221 way, though its exact impact is difficult to determine. Thanks to Pieter
9222 Wuille (Blockstream) who reported this issue and also suggested an initial
9223 fix. Further analysis was conducted by the OpenSSL development team and
9224 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9225 the OpenSSL core team.
9230 * Fix various certificate fingerprint issues.
9232 By using non-DER or invalid encodings outside the signed portion of a
9233 certificate the fingerprint can be changed without breaking the signature.
9234 Although no details of the signed portion of the certificate can be changed
9235 this can cause problems with some applications: e.g. those using the
9236 certificate fingerprint for blacklists.
9238 1. Reject signatures with non zero unused bits.
9240 If the BIT STRING containing the signature has non zero unused bits reject
9241 the signature. All current signature algorithms require zero unused bits.
9243 2. Check certificate algorithm consistency.
9245 Check the AlgorithmIdentifier inside TBS matches the one in the
9246 certificate signature. NB: this will result in signature failure
9247 errors for some broken certificates.
9249 Thanks to Konrad Kraszewski from Google for reporting this issue.
9251 3. Check DSA/ECDSA signatures use DER.
9253 Re-encode DSA/ECDSA signatures and compare with the original received
9254 signature. Return an error if there is a mismatch.
9256 This will reject various cases including garbage after signature
9257 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9258 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9259 (negative or with leading zeroes).
9261 Further analysis was conducted and fixes were developed by Stephen Henson
9262 of the OpenSSL core team.
9268 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9270 * Session Ticket Memory Leak.
9272 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9273 integrity of that ticket is first verified. In the event of a session
9274 ticket integrity check failing, OpenSSL will fail to free memory
9275 causing a memory leak. By sending a large number of invalid session
9276 tickets an attacker could exploit this issue in a Denial Of Service
9282 * Build option no-ssl3 is incomplete.
9284 When OpenSSL is configured with "no-ssl3" as a build option, servers
9285 could accept and complete an SSL 3.0 handshake, and clients could be
9286 configured to send them.
9289 *Akamai and the OpenSSL team*
9291 * Add support for TLS_FALLBACK_SCSV.
9292 Client applications doing fallback retries should call
9293 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9296 *Adam Langley, Bodo Moeller*
9298 * Add additional DigestInfo checks.
9300 Re-encode DigestInto in DER and check against the original when
9301 verifying RSA signature: this will reject any improperly encoded
9302 DigestInfo structures.
9304 Note: this is a precautionary measure and no attacks are currently known.
9308 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9310 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9311 to a denial of service attack. A malicious server can crash the client
9312 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9313 ciphersuite and sending carefully crafted handshake messages.
9315 Thanks to Felix Gröbert (Google) for discovering and researching this
9321 * By sending carefully crafted DTLS packets an attacker could cause openssl
9322 to leak memory. This can be exploited through a Denial of Service attack.
9323 Thanks to Adam Langley for discovering and researching this issue.
9328 * An attacker can force openssl to consume large amounts of memory whilst
9329 processing DTLS handshake messages. This can be exploited through a
9330 Denial of Service attack.
9331 Thanks to Adam Langley for discovering and researching this issue.
9336 * An attacker can force an error condition which causes openssl to crash
9337 whilst processing DTLS packets due to memory being freed twice. This
9338 can be exploited through a Denial of Service attack.
9339 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9345 * If a multithreaded client connects to a malicious server using a resumed
9346 session and the server sends an ec point format extension it could write
9347 up to 255 bytes to freed memory.
9349 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9355 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9356 X509_name_oneline, X509_name_print_ex et al. to leak some information
9357 from the stack. Applications may be affected if they echo pretty printing
9358 output to the attacker.
9360 Thanks to Ivan Fratric (Google) for discovering this issue.
9363 *Emilia Käsper, and Steve Henson*
9365 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9366 for corner cases. (Certain input points at infinity could lead to
9367 bogus results, with non-infinity inputs mapped to infinity too.)
9371 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9373 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9374 handshake can force the use of weak keying material in OpenSSL
9375 SSL/TLS clients and servers.
9377 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9378 researching this issue. ([CVE-2014-0224])
9380 *KIKUCHI Masashi, Steve Henson*
9382 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9383 OpenSSL DTLS client the code can be made to recurse eventually crashing
9386 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9389 *Imre Rad, Steve Henson*
9391 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9392 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9393 client or server. This is potentially exploitable to run arbitrary
9394 code on a vulnerable client or server.
9396 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9398 *Jüri Aedla, Steve Henson*
9400 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9401 are subject to a denial of service attack.
9403 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9404 this issue. ([CVE-2014-3470])
9406 *Felix Gröbert, Ivan Fratric, Steve Henson*
9408 * Harmonize version and its documentation. -f flag is used to display
9411 *mancha <mancha1@zoho.com>*
9413 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9414 in i2d_ECPrivateKey.
9416 *mancha <mancha1@zoho.com>*
9418 * Fix some double frees. These are not thought to be exploitable.
9420 *mancha <mancha1@zoho.com>*
9422 * Fix for the attack described in the paper "Recovering OpenSSL
9423 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9424 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9425 <http://eprint.iacr.org/2014/140>
9427 Thanks to Yuval Yarom and Naomi Benger for discovering this
9428 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9430 *Yuval Yarom and Naomi Benger*
9432 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9434 * Keep original DTLS digest and encryption contexts in retransmission
9435 structures so we can use the previous session parameters if they need
9436 to be resent. ([CVE-2013-6450])
9440 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9441 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9442 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9443 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9444 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9445 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9447 *Rob Stradling, Adam Langley*
9449 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9451 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9453 This addresses the flaw in CBC record processing discovered by
9454 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9455 at: <http://www.isg.rhul.ac.uk/tls/>
9457 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9458 Security Group at Royal Holloway, University of London
9459 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9460 Emilia Käsper for the initial patch.
9463 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9465 * Return an error when checking OCSP signatures when key is NULL.
9466 This fixes a DoS attack. ([CVE-2013-0166])
9470 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9471 the right response is stapled. Also change SSL_get_certificate()
9472 so it returns the certificate actually sent.
9473 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9474 (This is a backport)
9476 *Rob Stradling <rob.stradling@comodo.com>*
9478 * Fix possible deadlock when decoding public keys.
9482 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9484 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9487 * Sanity check record length before skipping explicit IV in DTLS
9490 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9491 fuzzing as a service testing platform.
9496 * Initialise tkeylen properly when encrypting CMS messages.
9497 Thanks to Solar Designer of Openwall for reporting this issue.
9501 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9503 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9504 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9505 in CRYPTO_realloc_clean.
9507 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9508 issue and to Adam Langley <agl@chromium.org> for fixing it.
9511 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9513 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9515 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9516 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9517 content decryption and always return the same error. Note: this attack
9518 needs on average 2^20 messages so it only affects automated senders. The
9519 old behaviour can be re-enabled in the CMS code by setting the
9520 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9521 an MMA defence is not necessary.
9522 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9523 this issue. ([CVE-2012-0884])
9527 * Fix CVE-2011-4619: make sure we really are receiving a
9528 client hello before rejecting multiple SGC restarts. Thanks to
9529 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9533 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9535 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9536 Thanks to Antonio Martin, Enterprise Secure Access Research and
9537 Development, Cisco Systems, Inc. for discovering this bug and
9538 preparing a fix. ([CVE-2012-0050])
9542 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9544 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9545 of the Vaudenay padding oracle attack on CBC mode encryption
9546 which enables an efficient plaintext recovery attack against
9547 the OpenSSL implementation of DTLS. Their attack exploits timing
9548 differences arising during decryption processing. A research
9549 paper describing this attack can be found at:
9550 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9551 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9552 Security Group at Royal Holloway, University of London
9553 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9554 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9555 for preparing the fix. ([CVE-2011-4108])
9557 *Robin Seggelmann, Michael Tuexen*
9559 * Clear bytes used for block padding of SSL 3.0 records.
9562 *Adam Langley (Google)*
9564 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9565 Kadianakis <desnacked@gmail.com> for discovering this issue and
9566 Adam Langley for preparing the fix. ([CVE-2011-4619])
9568 *Adam Langley (Google)*
9570 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9572 *Andrey Kulikov <amdeich@gmail.com>*
9574 * Prevent malformed RFC3779 data triggering an assertion failure.
9575 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9576 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9578 *Rob Austein <sra@hactrn.net>*
9580 * Improved PRNG seeding for VOS.
9582 *Paul Green <Paul.Green@stratus.com>*
9584 * Fix ssl_ciph.c set-up race.
9586 *Adam Langley (Google)*
9588 * Fix spurious failures in ecdsatest.c.
9590 *Emilia Käsper (Google)*
9592 * Fix the BIO_f_buffer() implementation (which was mixing different
9593 interpretations of the `..._len` fields).
9595 *Adam Langley (Google)*
9597 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9598 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9599 threads won't reuse the same blinding coefficients.
9601 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9602 lock to call BN_BLINDING_invert_ex, and avoids one use of
9603 BN_BLINDING_update for each BN_BLINDING structure (previously,
9604 the last update always remained unused).
9606 *Emilia Käsper (Google)*
9608 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9610 *Bob Buckholz (Google)*
9612 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9614 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9615 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9617 *Kaspar Brand <ossl@velox.ch>*
9619 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9620 for multi-threaded use of ECDH. ([CVE-2011-3210])
9622 *Adam Langley (Google)*
9624 * Fix x509_name_ex_d2i memory leak on bad inputs.
9628 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9629 signature public key algorithm by using OID xref utilities instead.
9630 Before this you could only use some ECC ciphersuites with SHA1 only.
9634 * Add protection against ECDSA timing attacks as mentioned in the paper
9635 by Billy Bob Brumley and Nicola Tuveri, see:
9636 <http://eprint.iacr.org/2011/232.pdf>
9638 *Billy Bob Brumley and Nicola Tuveri*
9640 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9642 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9644 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9646 * Fix bug in string printing code: if *any* escaping is enabled we must
9647 escape the escape character (backslash) or the resulting string is
9652 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9654 * Disable code workaround for ancient and obsolete Netscape browsers
9655 and servers: an attacker can use it in a ciphersuite downgrade attack.
9656 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9660 * Fixed J-PAKE implementation error, originally discovered by
9661 Sebastien Martini, further info and confirmation from Stefan
9662 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9666 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9668 * Fix extension code to avoid race conditions which can result in a buffer
9669 overrun vulnerability: resumed sessions must not be modified as they can
9670 be shared by multiple threads. CVE-2010-3864
9674 * Fix WIN32 build system to correctly link an ENGINE directory into
9679 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9681 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9684 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9686 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9688 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9689 context. The operation can be customised via the ctrl mechanism in
9690 case ENGINEs want to include additional functionality.
9694 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9698 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9699 output hashes compatible with older versions of OpenSSL.
9701 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9703 * Fix compression algorithm handling: if resuming a session use the
9704 compression algorithm of the resumed session instead of determining
9705 it from client hello again. Don't allow server to change algorithm.
9709 * Add load_crls() function to commands tidying load_certs() too. Add option
9710 to verify utility to allow additional CRLs to be included.
9714 * Update OCSP request code to permit adding custom headers to the request:
9715 some responders need this.
9719 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9722 *Julia Lawall <julia@diku.dk>*
9724 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9725 needlessly dereferenced structures, used obsolete functions and
9726 didn't handle all updated verify codes correctly.
9730 * Disable MD2 in the default configuration.
9734 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9735 indicate the initial BIO being pushed or popped. This makes it possible
9736 to determine whether the BIO is the one explicitly called or as a result
9737 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9738 it handles reference counts correctly and doesn't zero out the I/O bio
9739 when it is not being explicitly popped. WARNING: applications which
9740 included workarounds for the old buggy behaviour will need to be modified
9741 or they could free up already freed BIOs.
9745 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9746 renaming to all platforms (within the 0.9.8 branch, this was
9747 done conditionally on Netware platforms to avoid a name clash).
9749 *Guenter <lists@gknw.net>*
9751 * Add ECDHE and PSK support to DTLS.
9753 *Michael Tuexen <tuexen@fh-muenster.de>*
9755 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9760 * Add "missing" function EVP_MD_flags() (without this the only way to
9761 retrieve a digest flags is by accessing the structure directly. Update
9762 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9763 or cipher is registered as in the "from" argument. Print out all
9764 registered digests in the dgst usage message instead of manually
9765 attempting to work them out.
9769 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9770 this allows the use of compression and extensions. Change default cipher
9771 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9772 by default unless an application cipher string requests it.
9776 * Alter match criteria in PKCS12_parse(). It used to try to use local
9777 key ids to find matching certificates and keys but some PKCS#12 files
9778 don't follow the (somewhat unwritten) rules and this strategy fails.
9779 Now just gather all certificates together and the first private key
9780 then look for the first certificate that matches the key.
9784 * Support use of registered digest and cipher names for dgst and cipher
9785 commands instead of having to add each one as a special case. So now
9792 openssl dgst -sha256 foo
9794 and this works for ENGINE based algorithms too.
9798 * Update Gost ENGINE to support parameter files.
9800 *Victor B. Wagner <vitus@cryptocom.ru>*
9802 * Support GeneralizedTime in ca utility.
9804 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9806 * Enhance the hash format used for certificate directory links. The new
9807 form uses the canonical encoding (meaning equivalent names will work
9808 even if they aren't identical) and uses SHA1 instead of MD5. This form
9809 is incompatible with the older format and as a result c_rehash should
9810 be used to rebuild symbolic links.
9814 * Make PKCS#8 the default write format for private keys, replacing the
9815 traditional format. This form is standardised, more secure and doesn't
9816 include an implicit MD5 dependency.
9820 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9821 committed to OpenSSL should pass this lot as a minimum.
9825 * Add session ticket override functionality for use by EAP-FAST.
9827 *Jouni Malinen <j@w1.fi>*
9829 * Modify HMAC functions to return a value. Since these can be implemented
9830 in an ENGINE errors can occur.
9834 * Type-checked OBJ_bsearch_ex.
9838 * Type-checked OBJ_bsearch. Also some constification necessitated
9839 by type-checking. Still to come: TXT_DB, bsearch(?),
9840 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9845 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9846 seconds to a tm structure directly, instead of going through OS
9847 specific date routines. This avoids any issues with OS routines such
9848 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9849 and X509_time_adj_ex() to cover the extended range. The existing
9850 X509_time_adj() is still usable and will no longer have any date issues.
9854 * Delta CRL support. New use deltas option which will attempt to locate
9855 and search any appropriate delta CRLs available.
9857 This work was sponsored by Google.
9861 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9862 code and add additional score elements. Validate alternate CRL paths
9863 as part of the CRL checking and indicate a new error "CRL path validation
9864 error" in this case. Applications wanting additional details can use
9865 the verify callback and check the new "parent" field. If this is not
9866 NULL CRL path validation is taking place. Existing applications won't
9867 see this because it requires extended CRL support which is off by
9870 This work was sponsored by Google.
9874 * Support for freshest CRL extension.
9876 This work was sponsored by Google.
9880 * Initial indirect CRL support. Currently only supported in the CRLs
9881 passed directly and not via lookup. Process certificate issuer
9882 CRL entry extension and lookup CRL entries by bother issuer name
9883 and serial number. Check and process CRL issuer entry in IDP extension.
9885 This work was sponsored by Google.
9889 * Add support for distinct certificate and CRL paths. The CRL issuer
9890 certificate is validated separately in this case. Only enabled if
9891 an extended CRL support flag is set: this flag will enable additional
9892 CRL functionality in future.
9894 This work was sponsored by Google.
9898 * Add support for policy mappings extension.
9900 This work was sponsored by Google.
9904 * Fixes to pathlength constraint, self issued certificate handling,
9905 policy processing to align with RFC3280 and PKITS tests.
9907 This work was sponsored by Google.
9911 * Support for name constraints certificate extension. DN, email, DNS
9912 and URI types are currently supported.
9914 This work was sponsored by Google.
9918 * To cater for systems that provide a pointer-based thread ID rather
9919 than numeric, deprecate the current numeric thread ID mechanism and
9920 replace it with a structure and associated callback type. This
9921 mechanism allows a numeric "hash" to be extracted from a thread ID in
9922 either case, and on platforms where pointers are larger than 'long',
9923 mixing is done to help ensure the numeric 'hash' is usable even if it
9924 can't be guaranteed unique. The default mechanism is to use "&errno"
9925 as a pointer-based thread ID to distinguish between threads.
9927 Applications that want to provide their own thread IDs should now use
9928 CRYPTO_THREADID_set_callback() to register a callback that will call
9929 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9931 Note that ERR_remove_state() is now deprecated, because it is tied
9932 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9933 to free the current thread's error state should be replaced by
9934 ERR_remove_thread_state(NULL).
9936 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9937 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9938 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9939 application was previously providing a numeric thread callback that
9940 was inappropriate for distinguishing threads, then uniqueness might
9941 have been obtained with &errno that happened immediately in the
9942 intermediate development versions of OpenSSL; this is no longer the
9943 case, the numeric thread callback will now override the automatic use
9946 *Geoff Thorpe, with help from Bodo Moeller*
9948 * Initial support for different CRL issuing certificates. This covers a
9949 simple case where the self issued certificates in the chain exist and
9950 the real CRL issuer is higher in the existing chain.
9952 This work was sponsored by Google.
9956 * Removed effectively defunct crypto/store from the build.
9960 * Revamp of STACK to provide stronger type-checking. Still to come:
9961 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9962 ASN1_STRING, CONF_VALUE.
9966 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9967 RAM on SSL connections. This option can save about 34k per idle SSL.
9971 * Revamp of LHASH to provide stronger type-checking. Still to come:
9972 STACK, TXT_DB, bsearch, qsort.
9976 * Initial support for Cryptographic Message Syntax (aka CMS) based
9977 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9978 support for data, signedData, compressedData, digestedData and
9979 encryptedData, envelopedData types included. Scripts to check against
9980 RFC4134 examples draft and interop and consistency checks of many
9981 content types and variants.
9985 * Add options to enc utility to support use of zlib compression BIO.
9989 * Extend mk1mf to support importing of options and assembly language
9990 files from Configure script, currently only included in VC-WIN32.
9991 The assembly language rules can now optionally generate the source
9992 files from the associated perl scripts.
9996 * Implement remaining functionality needed to support GOST ciphersuites.
9997 Interop testing has been performed using CryptoPro implementations.
9999 *Victor B. Wagner <vitus@cryptocom.ru>*
10001 * s390x assembler pack.
10005 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10010 * Implement Opaque PRF Input TLS extension as specified in
10011 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10012 official specification yet and no extension type assignment by
10013 IANA exists, this extension (for now) will have to be explicitly
10014 enabled when building OpenSSL by providing the extension number
10015 to use. For example, specify an option
10017 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10019 to the "config" or "Configure" script to enable the extension,
10020 assuming extension number 0x9527 (which is a completely arbitrary
10021 and unofficial assignment based on the MD5 hash of the Internet
10022 Draft). Note that by doing so, you potentially lose
10023 interoperability with other TLS implementations since these might
10024 be using the same extension number for other purposes.
10026 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10027 opaque PRF input value to use in the handshake. This will create
10028 an internal copy of the length-'len' string at 'src', and will
10029 return non-zero for success.
10031 To get more control and flexibility, provide a callback function
10034 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10035 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10039 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10042 Callback function 'cb' will be called in handshakes, and is
10043 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10044 Argument 'arg' is for application purposes (the value as given to
10045 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10046 be provided to the callback function). The callback function
10047 has to return non-zero to report success: usually 1 to use opaque
10048 PRF input just if possible, or 2 to enforce use of the opaque PRF
10049 input. In the latter case, the library will abort the handshake
10050 if opaque PRF input is not successfully negotiated.
10052 Arguments 'peerinput' and 'len' given to the callback function
10053 will always be NULL and 0 in the case of a client. A server will
10054 see the client's opaque PRF input through these variables if
10055 available (NULL and 0 otherwise). Note that if the server
10056 provides an opaque PRF input, the length must be the same as the
10057 length of the client's opaque PRF input.
10059 Note that the callback function will only be called when creating
10060 a new session (session resumption can resume whatever was
10061 previously negotiated), and will not be called in SSL 2.0
10062 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10063 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10064 for applications that need to enforce opaque PRF input.
10068 * Update ssl code to support digests other than SHA1+MD5 for handshake
10071 *Victor B. Wagner <vitus@cryptocom.ru>*
10073 * Add RFC4507 support to OpenSSL. This includes the corrections in
10074 RFC4507bis. The encrypted ticket format is an encrypted encoded
10075 SSL_SESSION structure, that way new session features are automatically
10078 If a client application caches session in an SSL_SESSION structure
10079 support is transparent because tickets are now stored in the encoded
10082 The SSL_CTX structure automatically generates keys for ticket
10083 protection in servers so again support should be possible
10084 with no application modification.
10086 If a client or server wishes to disable RFC4507 support then the option
10087 SSL_OP_NO_TICKET can be set.
10089 Add a TLS extension debugging callback to allow the contents of any client
10090 or server extensions to be examined.
10092 This work was sponsored by Google.
10096 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10097 OpenSSL should now compile cleanly on gcc 4.2
10099 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10101 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10102 support including streaming MAC support: this is required for GOST
10103 ciphersuite support.
10105 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10107 * Add option -stream to use PKCS#7 streaming in smime utility. New
10108 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10109 to output in BER and PEM format.
10113 * Experimental support for use of HMAC via EVP_PKEY interface. This
10114 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10115 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10116 ENGINE support for HMAC keys which are unextractable. New -mac and
10117 -macopt options to dgst utility.
10121 * New option -sigopt to dgst utility. Update dgst to use
10122 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10123 alternative signing parameters such as X9.31 or PSS in the dgst
10128 * Change ssl_cipher_apply_rule(), the internal function that does
10129 the work each time a ciphersuite string requests enabling
10130 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10131 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10132 the order of disabled ciphersuites such that those ciphersuites
10133 that most recently went from enabled to disabled not only stay
10134 in order with respect to each other, but also have higher priority
10135 than other disabled ciphersuites the next time ciphersuites are
10138 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10139 the same ciphersuites as with "HIGH" alone, but in a specific
10140 order where the PSK ciphersuites come first (since they are the
10141 most recently disabled ciphersuites when "HIGH" is parsed).
10143 Also, change ssl_create_cipher_list() (using this new
10144 functionality) such that between otherwise identical
10145 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10150 * Change ssl_create_cipher_list() so that it automatically
10151 arranges the ciphersuites in reasonable order before starting
10152 to process the rule string. Thus, the definition for "DEFAULT"
10153 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10154 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10155 This makes it much easier to arrive at a reasonable default order
10156 in applications for which anonymous ciphers are OK (meaning
10157 that you can't actually use DEFAULT).
10159 *Bodo Moeller; suggested by Victor Duchovni*
10161 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10162 processing) into multiple integers instead of setting
10163 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10164 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10165 (These masks as well as the individual bit definitions are hidden
10166 away into the non-exported interface ssl/ssl_locl.h, so this
10167 change to the definition of the SSL_CIPHER structure shouldn't
10168 affect applications.) This give us more bits for each of these
10169 categories, so there is no longer a need to coagulate AES128 and
10170 AES256 into a single algorithm bit, and to coagulate Camellia128
10171 and Camellia256 into a single algorithm bit, which has led to all
10174 Thus, among other things, the kludge introduced in 0.9.7m and
10175 0.9.8e for masking out AES256 independently of AES128 or masking
10176 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10178 With the change, we also introduce new ciphersuite aliases that
10179 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10184 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10185 Use the leftmost N bytes of the signature input if the input is
10186 larger than the prime q (with N being the size in bytes of q).
10190 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10191 it yet and it is largely untested.
10195 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10199 * Initial incomplete changes to avoid need for function casts in OpenSSL
10200 some compilers (gcc 4.2 and later) reject their use. Safestack is
10201 reimplemented. Update ASN1 to avoid use of legacy functions.
10205 * Win32/64 targets are linked with Winsock2.
10209 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10210 to external functions. This can be used to increase CRL handling
10211 efficiency especially when CRLs are very large by (for example) storing
10212 the CRL revoked certificates in a database.
10216 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10217 new CRLs added to a directory can be used. New command line option
10218 -verify_return_error to s_client and s_server. This causes real errors
10219 to be returned by the verify callback instead of carrying on no matter
10220 what. This reflects the way a "real world" verify callback would behave.
10224 * GOST engine, supporting several GOST algorithms and public key formats.
10225 Kindly donated by Cryptocom.
10229 * Partial support for Issuing Distribution Point CRL extension. CRLs
10230 partitioned by DP are handled but no indirect CRL or reason partitioning
10231 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10232 selected via a scoring technique which handles IDP and AKID in CRLs.
10236 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10237 will ultimately be used for all verify operations: this will remove the
10238 X509_STORE dependency on certificate verification and allow alternative
10239 lookup methods. X509_STORE based implementations of these two callbacks.
10243 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10244 Modify get_crl() to find a valid (unexpired) CRL if possible.
10248 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10249 this would be called X509_CRL_cmp() but that name is already used by
10250 a function that just compares CRL issuer names. Cache several CRL
10251 extensions in X509_CRL structure and cache CRLDP in X509.
10255 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10256 this maps equivalent X509_NAME structures into a consistent structure.
10257 Name comparison can then be performed rapidly using memcmp().
10261 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10266 * Allow digests to supply their own micalg string for S/MIME type using
10267 the ctrl EVP_MD_CTRL_MICALG.
10271 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10272 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10273 ctrl. It can then customise the structure before and/or after signing
10278 * New function OBJ_add_sigid() to allow application defined signature OIDs
10279 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10280 to free up any added signature OIDs.
10284 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10285 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10286 digest and cipher tables. New options added to openssl utility:
10287 list-message-digest-algorithms and list-cipher-algorithms.
10291 * Change the array representation of binary polynomials: the list
10292 of degrees of non-zero coefficients is now terminated with -1.
10293 Previously it was terminated with 0, which was also part of the
10294 value; thus, the array representation was not applicable to
10295 polynomials where t^0 has coefficient zero. This change makes
10296 the array representation useful in a more general context.
10300 * Various modifications and fixes to SSL/TLS cipher string
10301 handling. For ECC, the code now distinguishes between fixed ECDH
10302 with RSA certificates on the one hand and with ECDSA certificates
10303 on the other hand, since these are separate ciphersuites. The
10304 unused code for Fortezza ciphersuites has been removed.
10306 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10307 (not "ECDHE"). For consistency with the code for DH
10308 certificates, use of ECDH certificates is now considered ECDH
10309 authentication, not RSA or ECDSA authentication (the latter is
10310 merely the CA's signing algorithm and not actively used in the
10313 The temporary ciphersuite alias "ECCdraft" is no longer
10314 available, and ECC ciphersuites are no longer excluded from "ALL"
10315 and "DEFAULT". The following aliases now exist for RFC 4492
10316 ciphersuites, most of these by analogy with the DH case:
10318 kECDHr - ECDH cert, signed with RSA
10319 kECDHe - ECDH cert, signed with ECDSA
10320 kECDH - ECDH cert (signed with either RSA or ECDSA)
10321 kEECDH - ephemeral ECDH
10322 ECDH - ECDH cert or ephemeral ECDH
10325 aECDSA - ECDSA cert
10328 AECDH - anonymous ECDH
10329 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10333 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10334 Use correct micalg parameters depending on digest(s) in signed message.
10338 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10339 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10343 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10344 an engine to register a method. Add ENGINE lookups for methods and
10345 functional reference processing.
10349 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10350 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10355 * New -resign option to smime utility. This adds one or more signers
10356 to an existing PKCS#7 signedData structure. Also -md option to use an
10357 alternative message digest algorithm for signing.
10361 * Tidy up PKCS#7 routines and add new functions to make it easier to
10362 create PKCS7 structures containing multiple signers. Update smime
10363 application to support multiple signers.
10367 * New -macalg option to pkcs12 utility to allow setting of an alternative
10372 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10373 Reorganize PBE internals to lookup from a static table using NIDs,
10374 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10375 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10376 PRF which will be automatically used with PBES2.
10380 * Replace the algorithm specific calls to generate keys in "req" with the
10385 * Update PKCS#7 enveloped data routines to use new API. This is now
10386 supported by any public key method supporting the encrypt operation. A
10387 ctrl is added to allow the public key algorithm to examine or modify
10388 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10393 * Add a ctrl to asn1 method to allow a public key algorithm to express
10394 a default digest type to use. In most cases this will be SHA1 but some
10395 algorithms (such as GOST) need to specify an alternative digest. The
10396 return value indicates how strong the preference is 1 means optional and
10397 2 is mandatory (that is it is the only supported type). Modify
10398 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10399 use the default md. Update openssl utilities to use the default digest
10400 type for signing if it is not explicitly indicated.
10404 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10405 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10406 signing method from the key type. This effectively removes the link
10407 between digests and public key types.
10411 * Add an OID cross reference table and utility functions. Its purpose is to
10412 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10413 rsaEncryption. This will allow some of the algorithm specific hackery
10414 needed to use the correct OID to be removed.
10418 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10419 structures for PKCS7_sign(). They are now set up by the relevant public
10424 * Add provisional EC pkey method with support for ECDSA and ECDH.
10428 * Add support for key derivation (agreement) in the API, DH method and
10433 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10434 public and private key formats. As a side effect these add additional
10435 command line functionality not previously available: DSA signatures can be
10436 generated and verified using pkeyutl and DH key support and generation in
10443 *Oliver Tappe <zooey@hirschkaefer.de>*
10445 * New make target "install_html_docs" installs HTML renditions of the
10448 *Oliver Tappe <zooey@hirschkaefer.de>*
10450 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10451 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10452 support key and parameter generation and add initial key generation
10453 functionality for RSA.
10457 * Add functions for main EVP_PKEY_method operations. The undocumented
10458 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10459 `EVP_PKEY_{encrypt,decrypt}_old`.
10463 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10464 key API, doesn't do much yet.
10468 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10469 public key algorithms. New option to openssl utility:
10470 "list-public-key-algorithms" to print out info.
10474 * Implement the Supported Elliptic Curves Extension for
10475 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10479 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10480 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10484 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10485 utilities such as rsa, dsa, dsaparam etc except they process any key
10490 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10491 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10492 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10497 * Initial support for pluggable public key ASN1.
10498 De-spaghettify the public key ASN1 handling. Move public and private
10499 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10500 algorithm specific handling to a single module within the relevant
10501 algorithm directory. Add functions to allow (near) opaque processing
10502 of public and private key structures.
10506 * Implement the Supported Point Formats Extension for
10507 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10511 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10512 for the psk identity [hint] and the psk callback functions to the
10513 SSL_SESSION, SSL and SSL_CTX structure.
10516 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10520 SSL_CTX_use_psk_identity_hint
10521 SSL_get_psk_identity_hint
10522 SSL_get_psk_identity
10523 SSL_use_psk_identity_hint
10525 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10527 * Add RFC 3161 compliant time stamp request creation, response generation
10528 and response verification functionality.
10530 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10532 * Add initial support for TLS extensions, specifically for the server_name
10533 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10534 have new members for a hostname. The SSL data structure has an
10535 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10536 stored in that context to allow for session resumption, even after the
10537 SSL has been switched to a new SSL_CTX in reaction to a client's
10538 server_name extension.
10540 New functions (subject to change):
10542 SSL_get_servername()
10543 SSL_get_servername_type()
10546 New CTRL codes and macros (subject to change):
10548 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10549 - SSL_CTX_set_tlsext_servername_callback()
10550 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10551 - SSL_CTX_set_tlsext_servername_arg()
10552 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10554 openssl s_client has a new '-servername ...' option.
10556 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10557 '-key2 ...', '-servername_fatal' (subject to change). This allows
10558 testing the HostName extension for a specific single hostname ('-cert'
10559 and '-key' remain fallbacks for handshakes without HostName
10560 negotiation). If the unrecognized_name alert has to be sent, this by
10561 default is a warning; it becomes fatal with the '-servername_fatal'
10564 *Peter Sylvester, Remy Allais, Christophe Renou*
10566 * Whirlpool hash implementation is added.
10570 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10571 bn(64,32). Because of instruction set limitations it doesn't have
10572 any negative impact on performance. This was done mostly in order
10573 to make it possible to share assembler modules, such as bn_mul_mont
10574 implementations, between 32- and 64-bit builds without hassle.
10578 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10579 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10584 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10585 dedicated Montgomery multiplication procedure, is introduced.
10586 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10587 "64-bit" performance on certain 32-bit targets.
10591 * New option SSL_OP_NO_COMP to disable use of compression selectively
10592 in SSL structures. New SSL ctrl to set maximum send fragment size.
10593 Save memory by setting the I/O buffer sizes dynamically instead of
10594 using the maximum available value.
10598 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10599 in addition to the text details.
10603 * Very, very preliminary EXPERIMENTAL support for printing of general
10604 ASN1 structures. This currently produces rather ugly output and doesn't
10605 handle several customised structures at all.
10609 * Integrated support for PVK file format and some related formats such
10610 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10611 these in the 'rsa' and 'dsa' utilities.
10615 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10619 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10620 place for the (very old) "NETSCAPE" format certificates which are now
10621 handled using new ASN1 code equivalents.
10625 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10626 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10627 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10631 * Modify CRL distribution points extension code to print out previously
10632 unsupported fields. Enhance extension setting code to allow setting of
10637 * Add print and set support for Issuing Distribution Point CRL extension.
10641 * Change 'Configure' script to enable Camellia by default.
10648 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10650 * When rejecting SSL/TLS records due to an incorrect version number, never
10651 update s->server with a new major version number. As of
10652 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10653 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10654 the previous behavior could result in a read attempt at NULL when
10655 receiving specific incorrect SSL/TLS records once record payload
10656 protection is active. ([CVE-2010-0740])
10658 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10660 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10661 could be crashed if the relevant tables were not present (e.g. chrooted).
10663 *Tomas Hoger <thoger@redhat.com>*
10665 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10667 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10669 *Martin Olsson, Neel Mehta*
10671 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10672 accommodate for stack sorting, always a write lock!).
10676 * On some versions of WIN32 Heap32Next is very slow. This can cause
10677 excessive delays in the RAND_poll(): over a minute. As a workaround
10678 include a time check in the inner Heap32Next loop too.
10682 * The code that handled flushing of data in SSL/TLS originally used the
10683 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10684 the problem outlined in PR#1949. The fix suggested there however can
10685 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10686 of Apache). So instead simplify the code to flush unconditionally.
10687 This should be fine since flushing with no data to flush is a no op.
10691 * Handle TLS versions 2.0 and later properly and correctly use the
10692 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10693 off ancient servers have a habit of sticking around for a while...
10697 * Modify compression code so it frees up structures without using the
10698 ex_data callbacks. This works around a problem where some applications
10699 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10700 restarting) then use compression (e.g. SSL with compression) later.
10701 This results in significant per-connection memory leaks and
10702 has caused some security issues including CVE-2008-1678 and
10707 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10708 change when encrypting or decrypting.
10712 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10713 connect and renegotiate with servers which do not support RI.
10714 Until RI is more widely deployed this option is enabled by default.
10718 * Add "missing" ssl ctrls to clear options and mode.
10722 * If client attempts to renegotiate and doesn't support RI respond with
10723 a no_renegotiation alert as required by RFC5746. Some renegotiating
10724 TLS clients will continue a connection gracefully when they receive
10725 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10726 waiting for a server hello which it will never receive. Now we treat a
10727 received no_renegotiation alert as a fatal error. This is because
10728 applications requesting a renegotiation might well expect it to succeed
10729 and would have no code in place to handle the server denying it so the
10730 only safe thing to do is to terminate the connection.
10734 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10735 peer supports secure renegotiation and 0 otherwise. Print out peer
10736 renegotiation support in s_client/s_server.
10740 * Replace the highly broken and deprecated SPKAC certification method with
10741 the updated NID creation version. This should correctly handle UTF8.
10745 * Implement RFC5746. Re-enable renegotiation but require the extension
10746 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10747 turns out to be a bad idea. It has been replaced by
10748 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10749 SSL_CTX_set_options(). This is really not recommended unless you
10750 know what you are doing.
10752 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10754 * Fixes to stateless session resumption handling. Use initial_ctx when
10755 issuing and attempting to decrypt tickets in case it has changed during
10756 servername handling. Use a non-zero length session ID when attempting
10757 stateless session resumption: this makes it possible to determine if
10758 a resumption has occurred immediately after receiving server hello
10759 (several places in OpenSSL subtly assume this) instead of later in
10764 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10765 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10766 fixes for a few places where the return code is not checked
10769 *Julia Lawall <julia@diku.dk>*
10771 * Add --strict-warnings option to Configure script to include devteam
10772 warnings in other configurations.
10776 * Add support for --libdir option and LIBDIR variable in makefiles. This
10777 makes it possible to install openssl libraries in locations which
10778 have names other than "lib", for example "/usr/lib64" which some
10781 *Steve Henson, based on patch from Jeremy Utley*
10783 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10784 X690 8.9.12 and can produce some misleading textual output of OIDs.
10786 *Steve Henson, reported by Dan Kaminsky*
10788 * Delete MD2 from algorithm tables. This follows the recommendation in
10789 several standards that it is not used in new applications due to
10790 several cryptographic weaknesses. For binary compatibility reasons
10791 the MD2 API is still compiled in by default.
10795 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10800 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10801 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10804 *Guenter <lists@gknw.net>*
10806 * Fix the server certificate chain building code to use X509_verify_cert(),
10807 it used to have an ad-hoc builder which was unable to cope with anything
10808 other than a simple chain.
10810 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10812 * Don't check self signed certificate signatures in X509_verify_cert()
10813 by default (a flag can override this): it just wastes time without
10814 adding any security. As a useful side effect self signed root CAs
10815 with non-FIPS digests are now usable in FIPS mode.
10819 * In dtls1_process_out_of_seq_message() the check if the current message
10820 is already buffered was missing. For every new message was memory
10821 allocated, allowing an attacker to perform an denial of service attack
10822 with sending out of seq handshake messages until there is no memory
10823 left. Additionally every future message was buffered, even if the
10824 sequence number made no sense and would be part of another handshake.
10825 So only messages with sequence numbers less than 10 in advance will be
10826 buffered. ([CVE-2009-1378])
10828 *Robin Seggelmann, discovered by Daniel Mentz*
10830 * Records are buffered if they arrive with a future epoch to be
10831 processed after finishing the corresponding handshake. There is
10832 currently no limitation to this buffer allowing an attacker to perform
10833 a DOS attack with sending records with future epochs until there is no
10834 memory left. This patch adds the pqueue_size() function to determine
10835 the size of a buffer and limits the record buffer to 100 entries.
10838 *Robin Seggelmann, discovered by Daniel Mentz*
10840 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10841 parent structure is freed. ([CVE-2009-1379])
10845 * Handle non-blocking I/O properly in SSL_shutdown() call.
10847 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10849 * Add `2.5.4.*` OIDs
10851 *Ilya O. <vrghost@gmail.com>*
10853 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10855 * Disable renegotiation completely - this fixes a severe security
10856 problem ([CVE-2009-3555]) at the cost of breaking all
10857 renegotiation. Renegotiation can be re-enabled by setting
10858 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10859 run-time. This is really not recommended unless you know what
10864 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10866 * Don't set val to NULL when freeing up structures, it is freed up by
10867 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10868 zeroing past the valid field. ([CVE-2009-0789])
10870 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10872 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10873 checked correctly. This would allow some invalid signed attributes to
10874 appear to verify correctly. ([CVE-2009-0591])
10876 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10878 * Reject UniversalString and BMPString types with invalid lengths. This
10879 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10880 a legal length. ([CVE-2009-0590])
10884 * Set S/MIME signing as the default purpose rather than setting it
10885 unconditionally. This allows applications to override it at the store
10890 * Permit restricted recursion of ASN1 strings. This is needed in practice
10891 to handle some structures.
10895 * Improve efficiency of mem_gets: don't search whole buffer each time
10898 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10900 * New -hex option for openssl rand.
10904 * Print out UTF8String and NumericString when parsing ASN1.
10908 * Support NumericString type for name components.
10912 * Allow CC in the environment to override the automatically chosen
10913 compiler. Note that nothing is done to ensure flags work with the
10918 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10920 * Properly check EVP_VerifyFinal() and similar return values
10923 *Ben Laurie, Bodo Moeller, Google Security Team*
10925 * Enable TLS extensions by default.
10929 * Allow the CHIL engine to be loaded, whether the application is
10930 multithreaded or not. (This does not release the developer from the
10931 obligation to set up the dynamic locking callbacks.)
10933 *Sander Temme <sander@temme.net>*
10935 * Use correct exit code if there is an error in dgst command.
10937 *Steve Henson; problem pointed out by Roland Dirlewanger*
10939 * Tweak Configure so that you need to say "experimental-jpake" to enable
10940 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10944 * Add experimental JPAKE support, including demo authentication in
10945 s_client and s_server.
10949 * Set the comparison function in v3_addr_canonize().
10951 *Rob Austein <sra@hactrn.net>*
10953 * Add support for XMPP STARTTLS in s_client.
10955 *Philip Paeps <philip@freebsd.org>*
10957 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10958 to ensure that even with this option, only ciphersuites in the
10959 server's preference list will be accepted. (Note that the option
10960 applies only when resuming a session, so the earlier behavior was
10961 just about the algorithm choice for symmetric cryptography.)
10965 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10967 * Fix NULL pointer dereference if a DTLS server received
10968 ChangeCipherSpec as first record ([CVE-2009-1386]).
10972 * Fix a state transition in s3_srvr.c and d1_srvr.c
10973 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10975 *Nagendra Modadugu*
10977 * The fix in 0.9.8c that supposedly got rid of unsafe
10978 double-checked locking was incomplete for RSA blinding,
10979 addressing just one layer of what turns out to have been
10980 doubly unsafe triple-checked locking.
10982 So now fix this for real by retiring the MONT_HELPER macro
10983 in crypto/rsa/rsa_eay.c.
10985 *Bodo Moeller; problem pointed out by Marius Schilder*
10987 * Various precautionary measures:
10989 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10991 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10992 (NB: This would require knowledge of the secret session ticket key
10993 to exploit, in which case you'd be SOL either way.)
10995 - Change bn_nist.c so that it will properly handle input BIGNUMs
10996 outside the expected range.
10998 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11001 *Neel Mehta, Bodo Moeller*
11003 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11004 the load fails. Useful for distros.
11006 *Ben Laurie and the FreeBSD team*
11008 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11012 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11016 * Expand ENGINE to support engine supplied SSL client certificate functions.
11018 This work was sponsored by Logica.
11022 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11023 keystores. Support for SSL/TLS client authentication too.
11024 Not compiled unless enable-capieng specified to Configure.
11026 This work was sponsored by Logica.
11030 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11031 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11032 attribute creation routines such as certificate requests and PKCS#12
11037 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11039 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11040 handshake which could lead to a client crash as found using the
11041 Codenomicon TLS test suite ([CVE-2008-1672])
11043 *Steve Henson, Mark Cox*
11045 * Fix double free in TLS server name extensions which could lead to
11046 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11050 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11052 Clear the error queue to ensure that error entries left from
11053 older function calls do not interfere with the correct operation.
11055 *Lutz Jaenicke, Erik de Castro Lopo*
11057 * Remove root CA certificates of commercial CAs:
11059 The OpenSSL project does not recommend any specific CA and does not
11060 have any policy with respect to including or excluding any CA.
11061 Therefore, it does not make any sense to ship an arbitrary selection
11062 of root CA certificates with the OpenSSL software.
11066 * RSA OAEP patches to fix two separate invalid memory reads.
11067 The first one involves inputs when 'lzero' is greater than
11068 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11069 before the beginning of from). The second one involves inputs where
11070 the 'db' section contains nothing but zeroes (there is a one-byte
11071 invalid read after the end of 'db').
11073 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11075 * Partial backport from 0.9.9-dev:
11077 Introduce bn_mul_mont (dedicated Montgomery multiplication
11078 procedure) as a candidate for BIGNUM assembler implementation.
11079 While 0.9.9-dev uses assembler for various architectures, only
11080 x86_64 is available by default here in the 0.9.8 branch, and
11081 32-bit x86 is available through a compile-time setting.
11083 To try the 32-bit x86 assembler implementation, use Configure
11084 option "enable-montasm" (which exists only for this backport).
11086 As "enable-montasm" for 32-bit x86 disclaims code stability
11087 anyway, in this constellation we activate additional code
11088 backported from 0.9.9-dev for further performance improvements,
11089 namely BN_from_montgomery_word. (To enable this otherwise,
11090 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11092 *Andy Polyakov (backport partially by Bodo Moeller)*
11094 * Add TLS session ticket callback. This allows an application to set
11095 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11096 values. This is useful for key rollover for example where several key
11097 sets may exist with different names.
11101 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11102 This was broken until now in 0.9.8 releases, such that the only way
11103 a registered ENGINE could be used (assuming it initialises
11104 successfully on the host) was to explicitly set it as the default
11105 for the relevant algorithms. This is in contradiction with 0.9.7
11106 behaviour and the documentation. With this fix, when an ENGINE is
11107 registered into a given algorithm's table of implementations, the
11108 'uptodate' flag is reset so that auto-discovery will be used next
11109 time a new context for that algorithm attempts to select an
11112 *Ian Lister (tweaked by Geoff Thorpe)*
11114 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11115 implementation in the following ways:
11117 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11120 Lack of BER streaming support means one pass streaming processing is
11121 only supported if data is detached: setting the streaming flag is
11122 ignored for embedded content.
11124 CMS support is disabled by default and must be explicitly enabled
11125 with the enable-cms configuration option.
11129 * Update the GMP engine glue to do direct copies between BIGNUM and
11130 mpz_t when openssl and GMP use the same limb size. Otherwise the
11131 existing "conversion via a text string export" trick is still used.
11133 *Paul Sheer <paulsheer@gmail.com>*
11135 * Zlib compression BIO. This is a filter BIO which compressed and
11136 uncompresses any data passed through it.
11140 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11141 RFC3394 compatible AES key wrapping.
11145 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11146 sets string data without copying. X509_ALGOR_set0() and
11147 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11148 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11149 from an X509_ATTRIBUTE structure optionally checking it occurs only
11150 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11155 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11156 to get the expected BN_FLG_CONSTTIME behavior.
11158 *Bodo Moeller (Google)*
11162 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11163 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11164 - added some more tests to do_tests.pl
11165 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11166 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11167 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11168 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11169 - various changes to netware.pl to enable gcc-cross builds on Win32
11171 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11172 - various changes to fix missing prototype warnings
11173 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11174 - added AES, WHIRLPOOL and CPUID assembler code to build files
11175 - added missing AES assembler make rules to mk1mf.pl
11176 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11178 *Guenter Knauf <eflash@gmx.net>*
11180 * Implement certificate status request TLS extension defined in RFC3546.
11181 A client can set the appropriate parameters and receive the encoded
11182 OCSP response via a callback. A server can query the supplied parameters
11183 and set the encoded OCSP response in the callback. Add simplified examples
11184 to s_client and s_server.
11188 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11190 * Fix various bugs:
11191 + Binary incompatibility of ssl_ctx_st structure
11192 + DTLS interoperation with non-compliant servers
11193 + Don't call get_session_cb() without proposed session
11194 + Fix ia64 assembler code
11196 *Andy Polyakov, Steve Henson*
11198 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11200 * DTLS Handshake overhaul. There were longstanding issues with
11201 OpenSSL DTLS implementation, which were making it impossible for
11202 RFC 4347 compliant client to communicate with OpenSSL server.
11203 Unfortunately just fixing these incompatibilities would "cut off"
11204 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11205 server keeps tolerating non RFC compliant syntax. The opposite is
11206 not true, 0.9.8f client can not communicate with earlier server.
11207 This update even addresses CVE-2007-4995.
11211 * Changes to avoid need for function casts in OpenSSL: some compilers
11212 (gcc 4.2 and later) reject their use.
11213 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11216 * Add RFC4507 support to OpenSSL. This includes the corrections in
11217 RFC4507bis. The encrypted ticket format is an encrypted encoded
11218 SSL_SESSION structure, that way new session features are automatically
11221 If a client application caches session in an SSL_SESSION structure
11222 support is transparent because tickets are now stored in the encoded
11225 The SSL_CTX structure automatically generates keys for ticket
11226 protection in servers so again support should be possible
11227 with no application modification.
11229 If a client or server wishes to disable RFC4507 support then the option
11230 SSL_OP_NO_TICKET can be set.
11232 Add a TLS extension debugging callback to allow the contents of any client
11233 or server extensions to be examined.
11235 This work was sponsored by Google.
11239 * Add initial support for TLS extensions, specifically for the server_name
11240 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11241 have new members for a hostname. The SSL data structure has an
11242 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11243 stored in that context to allow for session resumption, even after the
11244 SSL has been switched to a new SSL_CTX in reaction to a client's
11245 server_name extension.
11247 New functions (subject to change):
11249 SSL_get_servername()
11250 SSL_get_servername_type()
11253 New CTRL codes and macros (subject to change):
11255 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11256 - SSL_CTX_set_tlsext_servername_callback()
11257 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11258 - SSL_CTX_set_tlsext_servername_arg()
11259 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11261 openssl s_client has a new '-servername ...' option.
11263 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11264 '-key2 ...', '-servername_fatal' (subject to change). This allows
11265 testing the HostName extension for a specific single hostname ('-cert'
11266 and '-key' remain fallbacks for handshakes without HostName
11267 negotiation). If the unrecognized_name alert has to be sent, this by
11268 default is a warning; it becomes fatal with the '-servername_fatal'
11271 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11273 * Add AES and SSE2 assembly language support to VC++ build.
11277 * Mitigate attack on final subtraction in Montgomery reduction.
11281 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11282 (which previously caused an internal error).
11286 * Squeeze another 10% out of IGE mode when in != out.
11290 * AES IGE mode speedup.
11292 *Dean Gaudet (Google)*
11294 * Add the Korean symmetric 128-bit cipher SEED (see
11295 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11296 add SEED ciphersuites from RFC 4162:
11298 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11299 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11300 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11301 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11303 To minimize changes between patchlevels in the OpenSSL 0.9.8
11304 series, SEED remains excluded from compilation unless OpenSSL
11305 is configured with 'enable-seed'.
11307 *KISA, Bodo Moeller*
11309 * Mitigate branch prediction attacks, which can be practical if a
11310 single processor is shared, allowing a spy process to extract
11311 information. For detailed background information, see
11312 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11313 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11314 and Necessary Software Countermeasures"). The core of the change
11315 are new versions BN_div_no_branch() and
11316 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11317 respectively, which are slower, but avoid the security-relevant
11318 conditional branches. These are automatically called by BN_div()
11319 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11320 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11321 remove a conditional branch.
11323 BN_FLG_CONSTTIME is the new name for the previous
11324 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11325 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11326 in the exponent causes BN_mod_exp_mont() to use the alternative
11327 implementation in BN_mod_exp_mont_consttime().) The old name
11328 remains as a deprecated alias.
11330 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11331 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11332 constant-time implementations for more than just exponentiation.
11333 Here too the old name is kept as a deprecated alias.
11335 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11336 the BN_BLINDING structure gets an independent copy of the
11337 modulus. This means that the previous `BIGNUM *m` argument to
11338 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11339 essentially becomes `const BIGNUM *m`, although we can't actually
11340 change this in the header file before 0.9.9. It allows
11341 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11342 enable BN_FLG_CONSTTIME.
11344 *Matthew D Wood (Intel Corp)*
11346 * In the SSL/TLS server implementation, be strict about session ID
11347 context matching (which matters if an application uses a single
11348 external cache for different purposes). Previously,
11349 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11350 set. This did ensure strict client verification, but meant that,
11351 with applications using a single external cache for quite
11352 different requirements, clients could circumvent ciphersuite
11353 restrictions for a given session ID context by starting a session
11354 in a different context.
11358 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11359 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11360 authentication-only ciphersuites.
11364 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11365 not complete and could lead to a possible single byte overflow
11366 ([CVE-2007-5135]) [Ben Laurie]
11368 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11370 * Since AES128 and AES256 (and similarly Camellia128 and
11371 Camellia256) share a single mask bit in the logic of
11372 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11373 kludge to work properly if AES128 is available and AES256 isn't
11374 (or if Camellia128 is available and Camellia256 isn't).
11378 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11379 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11380 When a point or a seed is encoded in a BIT STRING, we need to
11381 prevent the removal of trailing zero bits to get the proper DER
11382 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11383 of a NamedBitList, for which trailing 0 bits need to be removed.)
11387 * Have SSL/TLS server implementation tolerate "mismatched" record
11388 protocol version while receiving ClientHello even if the
11389 ClientHello is fragmented. (The server can't insist on the
11390 particular protocol version it has chosen before the ServerHello
11391 message has informed the client about his choice.)
11395 * Add RFC 3779 support.
11397 *Rob Austein for ARIN, Ben Laurie*
11399 * Load error codes if they are not already present instead of using a
11400 static variable. This allows them to be cleanly unloaded and reloaded.
11401 Improve header file function name parsing.
11405 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11406 or CAPABILITY handshake as required by RFCs.
11408 *Goetz Babin-Ebell*
11410 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11412 * Introduce limits to prevent malicious keys being able to
11413 cause a denial of service. ([CVE-2006-2940])
11415 *Steve Henson, Bodo Moeller*
11417 * Fix ASN.1 parsing of certain invalid structures that can result
11418 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11420 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11421 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11423 * Fix SSL client code which could crash if connecting to a
11424 malicious SSLv2 server. ([CVE-2006-4343])
11426 *Tavis Ormandy and Will Drewry, Google Security Team*
11428 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11429 match only those. Before that, "AES256-SHA" would be interpreted
11430 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11431 the same strength classification in 0.9.7h) as we currently only
11432 have a single AES bit in the ciphersuite description bitmap.
11433 That change, however, also applied to ciphersuite strings such as
11434 "RC4-MD5" that intentionally matched multiple ciphersuites --
11435 namely, SSL 2.0 ciphersuites in addition to the more common ones
11436 from SSL 3.0/TLS 1.0.
11438 So we change the selection algorithm again: Naming an explicit
11439 ciphersuite selects this one ciphersuite, and any other similar
11440 ciphersuite (same bitmap) from *other* protocol versions.
11441 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11442 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11444 Since SSL 2.0 does not have any ciphersuites for which the
11445 128/256 bit distinction would be relevant, this works for now.
11446 The proper fix will be to use different bits for AES128 and
11447 AES256, which would have avoided the problems from the beginning;
11448 however, bits are scarce, so we can only do this in a new release
11449 (not just a patchlevel) when we can change the SSL_CIPHER
11450 definition to split the single 'unsigned long mask' bitmap into
11451 multiple values to extend the available space.
11455 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11457 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11458 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11460 * Add AES IGE and biIGE modes.
11464 * Change the Unix randomness entropy gathering to use poll() when
11465 possible instead of select(), since the latter has some
11466 undesirable limitations.
11468 *Darryl Miles via Richard Levitte and Bodo Moeller*
11470 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11471 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11472 cannot be implicitly activated as part of, e.g., the "AES" alias.
11473 However, please upgrade to OpenSSL 0.9.9[-dev] for
11474 non-experimental use of the ECC ciphersuites to get TLS extension
11475 support, which is required for curve and point format negotiation
11476 to avoid potential handshake problems.
11480 * Disable rogue ciphersuites:
11482 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11483 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11484 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11486 The latter two were purportedly from
11487 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11490 Also deactivate the remaining ciphersuites from
11491 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11492 unofficial, and the ID has long expired.
11496 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11497 dual-core machines) and other potential thread-safety issues.
11501 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11502 versions), which is now available for royalty-free use
11503 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11504 Also, add Camellia TLS ciphersuites from RFC 4132.
11506 To minimize changes between patchlevels in the OpenSSL 0.9.8
11507 series, Camellia remains excluded from compilation unless OpenSSL
11508 is configured with 'enable-camellia'.
11512 * Disable the padding bug check when compression is in use. The padding
11513 bug check assumes the first packet is of even length, this is not
11514 necessarily true if compression is enabled and can result in false
11515 positives causing handshake failure. The actual bug test is ancient
11516 code so it is hoped that implementations will either have fixed it by
11517 now or any which still have the bug do not support compression.
11521 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11523 * When applying a cipher rule check to see if string match is an explicit
11524 cipher suite and only match that one cipher suite if it is.
11528 * Link in manifests for VC++ if needed.
11530 *Austin Ziegler <halostatue@gmail.com>*
11532 * Update support for ECC-based TLS ciphersuites according to
11533 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11534 TLS extensions, which are supported starting with the 0.9.9
11535 branch, not in the OpenSSL 0.9.8 branch).
11539 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11540 opaque EVP_CIPHER_CTX handling.
11544 * Fixes and enhancements to zlib compression code. We now only use
11545 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11546 to conform with the standards mentioned here:
11547 <http://www.zlib.net/DLL_FAQ.txt>
11548 Static zlib linking now works on Windows and the new --with-zlib-include
11549 --with-zlib-lib options to Configure can be used to supply the location
11550 of the headers and library. Gracefully handle case where zlib library
11555 * Several fixes and enhancements to the OID generation code. The old code
11556 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11557 handle numbers larger than ULONG_MAX, truncated printing and had a
11558 non standard OBJ_obj2txt() behaviour.
11562 * Add support for building of engines under engine/ as shared libraries
11563 under VC++ build system.
11567 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11568 Hopefully, we will not see any false combination of paths any more.
11572 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11574 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11575 (part of SSL_OP_ALL). This option used to disable the
11576 countermeasure against man-in-the-middle protocol-version
11577 rollback in the SSL 2.0 server implementation, which is a bad
11578 idea. ([CVE-2005-2969])
11580 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11581 for Information Security, National Institute of Advanced Industrial
11582 Science and Technology [AIST], Japan)*
11584 * Add two function to clear and return the verify parameter flags.
11588 * Keep cipherlists sorted in the source instead of sorting them at
11589 runtime, thus removing the need for a lock.
11593 * Avoid some small subgroup attacks in Diffie-Hellman.
11595 *Nick Mathewson and Ben Laurie*
11597 * Add functions for well-known primes.
11601 * Extended Windows CE support.
11603 *Satoshi Nakamura and Andy Polyakov*
11605 * Initialize SSL_METHOD structures at compile time instead of during
11606 runtime, thus removing the need for a lock.
11610 * Make PKCS7_decrypt() work even if no certificate is supplied by
11611 attempting to decrypt each encrypted key in turn. Add support to
11616 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11618 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11621 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11625 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11626 key into the same file any more.
11630 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11634 * Add -utf8 command line and config file option to 'ca'.
11636 *Stefan <stf@udoma.org*
11638 * Removed the macro des_crypt(), as it seems to conflict with some
11639 libraries. Use DES_crypt().
11643 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11644 involves renaming the source and generated shared-libs for
11645 both. The engines will accept the corrected or legacy ids
11646 ('ncipher' and '4758_cca' respectively) when binding. NB,
11647 this only applies when building 'shared'.
11649 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11651 * Add attribute functions to EVP_PKEY structure. Modify
11652 PKCS12_create() to recognize a CSP name attribute and
11653 use it. Make -CSP option work again in pkcs12 utility.
11657 * Add new functionality to the bn blinding code:
11658 - automatic re-creation of the BN_BLINDING parameters after
11659 a fixed number of uses (currently 32)
11660 - add new function for parameter creation
11661 - introduce flags to control the update behaviour of the
11662 BN_BLINDING parameters
11663 - hide BN_BLINDING structure
11664 Add a second BN_BLINDING slot to the RSA structure to improve
11665 performance when a single RSA object is shared among several
11670 * Add support for DTLS.
11672 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11674 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11675 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11679 * Remove buggy and incomplete DH cert support from
11680 ssl/ssl_rsa.c and ssl/s3_both.c
11684 * Use SHA-1 instead of MD5 as the default digest algorithm for
11685 the `apps/openssl` commands.
11689 * Compile clean with "-Wall -Wmissing-prototypes
11690 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11691 DEBUG_SAFESTACK must also be set.
11695 * Change ./Configure so that certain algorithms can be disabled by default.
11696 The new counterpiece to "no-xxx" is "enable-xxx".
11698 The patented RC5 and MDC2 algorithms will now be disabled unless
11699 "enable-rc5" and "enable-mdc2", respectively, are specified.
11701 (IDEA remains enabled despite being patented. This is because IDEA
11702 is frequently required for interoperability, and there is no license
11703 fee for non-commercial use. As before, "no-idea" can be used to
11704 avoid this algorithm.)
11708 * Add processing of proxy certificates (see RFC 3820). This work was
11709 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11710 EGEE (Enabling Grids for E-science in Europe).
11714 * RC4 performance overhaul on modern architectures/implementations, such
11715 as Intel P4, IA-64 and AMD64.
11719 * New utility extract-section.pl. This can be used specify an alternative
11720 section number in a pod file instead of having to treat each file as
11721 a separate case in Makefile. This can be done by adding two lines to the
11724 =for comment openssl_section:XXX
11726 The blank line is mandatory.
11730 * New arguments -certform, -keyform and -pass for s_client and s_server
11731 to allow alternative format key and certificate files and passphrase
11736 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11737 update associated structures and add various utility functions.
11739 Add new policy related verify parameters, include policy checking in
11740 standard verify code. Enhance 'smime' application with extra parameters
11741 to support policy checking and print out.
11745 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11746 Nehemiah processors. These extensions support AES encryption in hardware
11747 as well as RNG (though RNG support is currently disabled).
11749 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11751 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11755 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11757 *Andy Polyakov and a number of other people*
11759 * Improved PowerPC platform support. Most notably BIGNUM assembler
11760 implementation contributed by IBM.
11762 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11764 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11765 exponent rather than 'unsigned long'. There is a corresponding change to
11766 the new 'rsa_keygen' element of the RSA_METHOD structure.
11768 *Jelte Jansen, Geoff Thorpe*
11770 * Functionality for creating the initial serial number file is now
11771 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11773 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11774 number file to 1, which is bound to cause problems. To avoid
11775 the problems while respecting compatibility between different 0.9.7
11776 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11777 CA.pl for serial number initialization. With the new release 0.9.8,
11778 we can fix the problem directly in the 'ca' utility.)
11782 * Reduced header interdependencies by declaring more opaque objects in
11783 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11784 give fewer recursive includes, which could break lazy source code - so
11785 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11786 developers should define this symbol when building and using openssl to
11787 ensure they track the recommended behaviour, interfaces, [etc], but
11788 backwards-compatible behaviour prevails when this isn't defined.
11792 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11796 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11797 This will generate a random key of the appropriate length based on the
11798 cipher context. The EVP_CIPHER can provide its own random key generation
11799 routine to support keys of a specific form. This is used in the des and
11800 3des routines to generate a key of the correct parity. Update S/MIME
11801 code to use new functions and hence generate correct parity DES keys.
11802 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11803 valid (weak or incorrect parity).
11807 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11808 as looking them up. This is useful when the verified structure may contain
11809 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11810 present unless the new PKCS7_NO_CRL flag is asserted.
11814 * Extend ASN1 oid configuration module. It now additionally accepts the
11817 shortName = some long name, 1.2.3.4
11821 * Reimplemented the BN_CTX implementation. There is now no more static
11822 limitation on the number of variables it can handle nor the depth of the
11823 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11824 information can now expand as required, and rather than having a single
11825 static array of bignums, BN_CTX now uses a linked-list of such arrays
11826 allowing it to expand on demand whilst maintaining the usefulness of
11827 BN_CTX's "bundling".
11831 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11832 to allow all RSA operations to function using a single BN_CTX.
11836 * Preliminary support for certificate policy evaluation and checking. This
11837 is initially intended to pass the tests outlined in "Conformance Testing
11838 of Relying Party Client Certificate Path Processing Logic" v1.07.
11842 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11843 remained unused and not that useful. A variety of other little bignum
11844 tweaks and fixes have also been made continuing on from the audit (see
11849 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11850 associated ASN1, EVP and SSL functions and old ASN1 macros.
11854 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11855 and this should never fail. So the return value from the use of
11856 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11857 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11861 * BN_CTX_get() should return zero-valued bignums, providing the same
11862 initialised value as BN_new().
11864 *Geoff Thorpe, suggested by Ulf Möller*
11866 * Support for inhibitAnyPolicy certificate extension.
11870 * An audit of the BIGNUM code is underway, for which debugging code is
11871 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11872 is considered valid when processing BIGNUMs, and causes execution to
11873 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11874 further steps are taken to deliberately pollute unused data in BIGNUM
11875 structures to try and expose faulty code further on. For now, openssl will
11876 (in its default mode of operation) continue to tolerate the inconsistent
11877 forms that it has tolerated in the past, but authors and packagers should
11878 consider trying openssl and their own applications when compiled with
11879 these debugging symbols defined. It will help highlight potential bugs in
11880 their own code, and will improve the test coverage for OpenSSL itself. At
11881 some point, these tighter rules will become openssl's default to improve
11882 maintainability, though the assert()s and other overheads will remain only
11883 in debugging configurations. See bn.h for more details.
11885 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11887 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11888 that can only be obtained through BN_CTX_new() (which implicitly
11889 initialises it). The presence of this function only made it possible
11890 to overwrite an existing structure (and cause memory leaks).
11894 * Because of the callback-based approach for implementing LHASH as a
11895 template type, lh_insert() adds opaque objects to hash-tables and
11896 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11897 to clean up those corresponding objects before destroying the hash table
11898 (and losing the object pointers). So some over-zealous constifications in
11899 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11900 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11901 prototyped to have "const" restrictions on the object pointers they are
11902 given (and so aren't required to cast them away any more).
11906 * The tmdiff.h API was so ugly and minimal that our own timing utility
11907 (speed) prefers to use its own implementation. The two implementations
11908 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11909 its object type properly exposed (MS_TM) instead of casting to/from
11910 `char *`. This may still change yet if someone realises MS_TM and
11912 aren't necessarily the greatest nomenclatures - but this is what was used
11913 internally to the implementation so I've used that for now.
11917 * Ensure that deprecated functions do not get compiled when
11918 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11919 the self-tests were still using deprecated key-generation functions so
11920 these have been updated also.
11924 * Reorganise PKCS#7 code to separate the digest location functionality
11925 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11926 New function PKCS7_set_digest() to set the digest type for PKCS#7
11927 digestedData type. Add additional code to correctly generate the
11928 digestedData type and add support for this type in PKCS7 initialization
11933 * New function PKCS7_set0_type_other() this initializes a PKCS7
11934 structure of type "other".
11938 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11939 sure the loop does correctly stop and breaking ("division by zero")
11940 modulus operations are not performed. The (pre-generated) prime
11941 table crypto/bn/bn_prime.h was already correct, but it could not be
11942 re-generated on some platforms because of the "division by zero"
11943 situation in the script.
11945 *Ralf S. Engelschall*
11947 * Update support for ECC-based TLS ciphersuites according to
11948 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11949 SHA-1 now is only used for "small" curves (where the
11950 representation of a field element takes up to 24 bytes); for
11951 larger curves, the field element resulting from ECDH is directly
11952 used as premaster secret.
11954 *Douglas Stebila (Sun Microsystems Laboratories)*
11956 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11957 curve secp160r1 to the tests.
11959 *Douglas Stebila (Sun Microsystems Laboratories)*
11961 * Add the possibility to load symbols globally with DSO.
11963 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11965 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11966 control of the error stack.
11970 * Add support for STORE in ENGINE.
11974 * Add the STORE type. The intention is to provide a common interface
11975 to certificate and key stores, be they simple file-based stores, or
11976 HSM-type store, or LDAP stores, or...
11977 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11981 * Add a generic structure called OPENSSL_ITEM. This can be used to
11982 pass a list of arguments to any function as well as provide a way
11983 for a function to pass data back to the caller.
11987 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11988 works like BUF_strdup() but can be used to duplicate a portion of
11989 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11994 * Add the function sk_find_ex() which works like sk_find(), but will
11995 return an index to an element even if an exact match couldn't be
11996 found. The index is guaranteed to point at the element where the
11997 searched-for key would be inserted to preserve sorting order.
12001 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12002 takes an extra flags argument for optional functionality. Currently,
12003 the following flags are defined:
12005 OBJ_BSEARCH_VALUE_ON_NOMATCH
12006 This one gets OBJ_bsearch_ex() to return a pointer to the first
12007 element where the comparing function returns a negative or zero
12010 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12011 This one gets OBJ_bsearch_ex() to return a pointer to the first
12012 element where the comparing function returns zero. This is useful
12013 if there are more than one element where the comparing function
12018 * Make it possible to create self-signed certificates with 'openssl ca'
12019 in such a way that the self-signed certificate becomes part of the
12020 CA database and uses the same mechanisms for serial number generation
12021 as all other certificate signing. The new flag '-selfsign' enables
12022 this functionality. Adapt CA.sh and CA.pl.in.
12026 * Add functionality to check the public key of a certificate request
12027 against a given private. This is useful to check that a certificate
12028 request can be signed by that key (self-signing).
12032 * Make it possible to have multiple active certificates with the same
12033 subject in the CA index file. This is done only if the keyword
12034 'unique_subject' is set to 'no' in the main CA section (default
12035 if 'CA_default') of the configuration file. The value is saved
12036 with the database itself in a separate index attribute file,
12037 named like the index file with '.attr' appended to the name.
12041 * Generate multi-valued AVAs using '+' notation in config files for
12046 * Support for nameConstraints certificate extension.
12050 * Support for policyConstraints certificate extension.
12054 * Support for policyMappings certificate extension.
12058 * Make sure the default DSA_METHOD implementation only uses its
12059 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12060 and change its own handlers to be NULL so as to remove unnecessary
12061 indirection. This lets alternative implementations fallback to the
12062 default implementation more easily.
12066 * Support for directoryName in GeneralName related extensions
12071 * Make it possible to link applications using Makefile.shared.
12072 Make that possible even when linking against static libraries!
12076 * Support for single pass processing for S/MIME signing. This now
12077 means that S/MIME signing can be done from a pipe, in addition
12078 cleartext signing (multipart/signed type) is effectively streaming
12079 and the signed data does not need to be all held in memory.
12081 This is done with a new flag PKCS7_STREAM. When this flag is set
12082 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12083 is done after the data is output (and digests calculated) in
12084 SMIME_write_PKCS7().
12088 * Add full support for -rpath/-R, both in shared libraries and
12089 applications, at least on the platforms where it's known how
12094 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12095 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12096 will now compute a table of multiples of the generator that
12097 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12098 faster (notably in the case of a single point multiplication,
12099 scalar * generator).
12101 *Nils Larsch, Bodo Moeller*
12103 * IPv6 support for certificate extensions. The various extensions
12104 which use the IP:a.b.c.d can now take IPv6 addresses using the
12105 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12110 * Added an ENGINE that implements RSA by performing private key
12111 exponentiations with the GMP library. The conversions to and from
12112 GMP's mpz_t format aren't optimised nor are any montgomery forms
12113 cached, and on x86 it appears OpenSSL's own performance has caught up.
12114 However there are likely to be other architectures where GMP could
12115 provide a boost. This ENGINE is not built in by default, but it can be
12116 specified at Configure time and should be accompanied by the necessary
12117 linker additions, eg;
12118 ./config -DOPENSSL_USE_GMP -lgmp
12122 * "openssl engine" will not display ENGINE/DSO load failure errors when
12123 testing availability of engines with "-t" - the old behaviour is
12124 produced by increasing the feature's verbosity with "-tt".
12128 * ECDSA routines: under certain error conditions uninitialized BN objects
12129 could be freed. Solution: make sure initialization is performed early
12130 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12135 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12136 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12137 software implementations. For DSA and DH, parameter generation can
12138 also be overridden by providing the appropriate method callbacks.
12142 * Change the "progress" mechanism used in key-generation and
12143 primality testing to functions that take a new BN_GENCB pointer in
12144 place of callback/argument pairs. The new API functions have `_ex`
12145 postfixes and the older functions are reimplemented as wrappers for
12146 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12147 declarations of the old functions to help (graceful) attempts to
12148 migrate to the new functions. Also, the new key-generation API
12149 functions operate on a caller-supplied key-structure and return
12150 success/failure rather than returning a key or NULL - this is to
12151 help make "keygen" another member function of RSA_METHOD etc.
12153 Example for using the new callback interface:
12155 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12156 void *my_arg = ...;
12159 BN_GENCB_set(&my_cb, my_callback, my_arg);
12161 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12162 /* For the meaning of a, b in calls to my_callback(), see the
12163 * documentation of the function that calls the callback.
12164 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12165 * my_callback should return 1 if it wants BN_is_prime_ex()
12166 * to continue, or 0 to stop.
12171 * Change the ZLIB compression method to be stateful, and make it
12172 available to TLS with the number defined in
12173 draft-ietf-tls-compression-04.txt.
12177 * Add the ASN.1 structures and functions for CertificatePair, which
12178 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12180 CertificatePair ::= SEQUENCE {
12181 forward [0] Certificate OPTIONAL,
12182 reverse [1] Certificate OPTIONAL,
12183 -- at least one of the pair shall be present -- }
12185 Also implement the PEM functions to read and write certificate
12186 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12188 This needed to be defined, mostly for the sake of the LDAP
12189 attribute crossCertificatePair, but may prove useful elsewhere as
12194 * Make it possible to inhibit symlinking of shared libraries in
12195 Makefile.shared, for Cygwin's sake.
12199 * Extend the BIGNUM API by creating a function
12200 void BN_set_negative(BIGNUM *a, int neg);
12201 and a macro that behave like
12202 int BN_is_negative(const BIGNUM *a);
12204 to avoid the need to access 'a->neg' directly in applications.
12208 * Implement fast modular reduction for pseudo-Mersenne primes
12209 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12210 EC_GROUP_new_curve_GFp() will now automatically use this
12213 *Nils Larsch <nla@trustcenter.de>*
12215 * Add new lock type (CRYPTO_LOCK_BN).
12219 * Change the ENGINE framework to automatically load engines
12220 dynamically from specific directories unless they could be
12221 found to already be built in or loaded. Move all the
12222 current engines except for the cryptodev one to a new
12223 directory engines/.
12224 The engines in engines/ are built as shared libraries if
12225 the "shared" options was given to ./Configure or ./config.
12226 Otherwise, they are inserted in libcrypto.a.
12227 /usr/local/ssl/engines is the default directory for dynamic
12228 engines, but that can be overridden at configure time through
12229 the usual use of --prefix and/or --openssldir, and at run
12230 time with the environment variable OPENSSL_ENGINES.
12232 *Geoff Thorpe and Richard Levitte*
12234 * Add Makefile.shared, a helper makefile to build shared
12235 libraries. Adapt Makefile.org.
12239 * Add version info to Win32 DLLs.
12241 *Peter 'Luna' Runestig" <peter@runestig.com>*
12243 * Add new 'medium level' PKCS#12 API. Certificates and keys
12244 can be added using this API to created arbitrary PKCS#12
12245 files while avoiding the low-level API.
12247 New options to PKCS12_create(), key or cert can be NULL and
12248 will then be omitted from the output file. The encryption
12249 algorithm NIDs can be set to -1 for no encryption, the mac
12250 iteration count can be set to 0 to omit the mac.
12252 Enhance pkcs12 utility by making the -nokeys and -nocerts
12253 options work when creating a PKCS#12 file. New option -nomac
12254 to omit the mac, NONE can be set for an encryption algorithm.
12255 New code is modified to use the enhanced PKCS12_create()
12256 instead of the low-level API.
12260 * Extend ASN1 encoder to support indefinite length constructed
12261 encoding. This can output sequences tags and octet strings in
12262 this form. Modify pk7_asn1.c to support indefinite length
12263 encoding. This is experimental and needs additional code to
12264 be useful, such as an ASN1 bio and some enhanced streaming
12267 Extend template encode functionality so that tagging is passed
12268 down to the template encoder.
12272 * Let 'openssl req' fail if an argument to '-newkey' is not
12273 recognized instead of using RSA as a default.
12277 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12278 As these are not official, they are not included in "ALL";
12279 the "ECCdraft" ciphersuite group alias can be used to select them.
12281 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12283 * Add ECDH engine support.
12285 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12287 * Add ECDH in new directory crypto/ecdh/.
12289 *Douglas Stebila (Sun Microsystems Laboratories)*
12291 * Let BN_rand_range() abort with an error after 100 iterations
12292 without success (which indicates a broken PRNG).
12296 * Change BN_mod_sqrt() so that it verifies that the input value
12297 is really the square of the return value. (Previously,
12298 BN_mod_sqrt would show GIGO behaviour.)
12302 * Add named elliptic curves over binary fields from X9.62, SECG,
12303 and WAP/WTLS; add OIDs that were still missing.
12305 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12307 * Extend the EC library for elliptic curves over binary fields
12308 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12311 EC_GF2m_simple_method
12315 EC_GROUP_new_curve_GF2m
12316 EC_GROUP_set_curve_GF2m
12317 EC_GROUP_get_curve_GF2m
12318 EC_POINT_set_affine_coordinates_GF2m
12319 EC_POINT_get_affine_coordinates_GF2m
12320 EC_POINT_set_compressed_coordinates_GF2m
12322 Point compression for binary fields is disabled by default for
12323 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12326 As binary polynomials are represented as BIGNUMs, various members
12327 of the EC_GROUP and EC_POINT data structures can be shared
12328 between the implementations for prime fields and binary fields;
12329 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12330 are essentially identical to their `..._GFp` counterparts.
12331 (For simplicity, the `..._GFp` prefix has been dropped from
12332 various internal method names.)
12334 An internal 'field_div' method (similar to 'field_mul' and
12335 'field_sqr') has been added; this is used only for binary fields.
12337 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12339 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12340 through methods ('mul', 'precompute_mult').
12342 The generic implementations (now internally called 'ec_wNAF_mul'
12343 and 'ec_wNAF_precomputed_mult') remain the default if these
12344 methods are undefined.
12346 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12348 * New function EC_GROUP_get_degree, which is defined through
12349 EC_METHOD. For curves over prime fields, this returns the bit
12350 length of the modulus.
12352 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12354 * New functions EC_GROUP_dup, EC_POINT_dup.
12355 (These simply call ..._new and ..._copy).
12357 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12359 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12360 Polynomials are represented as BIGNUMs (where the sign bit is not
12361 used) in the following functions [macros]:
12364 BN_GF2m_sub [= BN_GF2m_add]
12365 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12366 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12367 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12369 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12370 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12371 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12372 BN_GF2m_cmp [= BN_ucmp]
12374 (Note that only the 'mod' functions are actually for fields GF(2^m).
12375 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12377 For some functions, an the irreducible polynomial defining a
12378 field can be given as an 'unsigned int[]' with strictly
12379 decreasing elements giving the indices of those bits that are set;
12380 i.e., p[] represents the polynomial
12381 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12383 p[0] > p[1] > ... > p[k] = 0.
12384 This applies to the following functions:
12387 BN_GF2m_mod_mul_arr
12388 BN_GF2m_mod_sqr_arr
12389 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12390 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12391 BN_GF2m_mod_exp_arr
12392 BN_GF2m_mod_sqrt_arr
12393 BN_GF2m_mod_solve_quad_arr
12397 Conversion can be performed by the following functions:
12402 bntest.c has additional tests for binary polynomial arithmetic.
12404 Two implementations for BN_GF2m_mod_div() are available.
12405 The default algorithm simply uses BN_GF2m_mod_inv() and
12406 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12407 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12408 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12410 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12412 * Add new error code 'ERR_R_DISABLED' that can be used when some
12413 functionality is disabled at compile-time.
12415 *Douglas Stebila <douglas.stebila@sun.com>*
12417 * Change default behaviour of 'openssl asn1parse' so that more
12418 information is visible when viewing, e.g., a certificate:
12420 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12421 mode the content of non-printable OCTET STRINGs is output in a
12422 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12423 avoid the appearance of a printable string.
12425 *Nils Larsch <nla@trustcenter.de>*
12427 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12429 EC_GROUP_set_asn1_flag()
12430 EC_GROUP_get_asn1_flag()
12431 EC_GROUP_set_point_conversion_form()
12432 EC_GROUP_get_point_conversion_form()
12433 These control ASN1 encoding details:
12434 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12435 has been set to OPENSSL_EC_NAMED_CURVE.
12436 - Points are encoded in uncompressed form by default; options for
12437 asn1_for are as for point2oct, namely
12438 POINT_CONVERSION_COMPRESSED
12439 POINT_CONVERSION_UNCOMPRESSED
12440 POINT_CONVERSION_HYBRID
12442 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12444 EC_GROUP_set_seed()
12445 EC_GROUP_get0_seed()
12446 EC_GROUP_get_seed_len()
12447 This is used only for ASN1 purposes (so far).
12449 *Nils Larsch <nla@trustcenter.de>*
12451 * Add 'field_type' member to EC_METHOD, which holds the NID
12452 of the appropriate field type OID. The new function
12453 EC_METHOD_get_field_type() returns this value.
12455 *Nils Larsch <nla@trustcenter.de>*
12458 EC_POINT_point2bn()
12459 EC_POINT_bn2point()
12460 EC_POINT_point2hex()
12461 EC_POINT_hex2point()
12462 providing useful interfaces to EC_POINT_point2oct() and
12463 EC_POINT_oct2point().
12465 *Nils Larsch <nla@trustcenter.de>*
12467 * Change internals of the EC library so that the functions
12468 EC_GROUP_set_generator()
12469 EC_GROUP_get_generator()
12470 EC_GROUP_get_order()
12471 EC_GROUP_get_cofactor()
12472 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12473 to methods, which would lead to unnecessary code duplication when
12474 adding different types of curves.
12476 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12478 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12479 arithmetic, and such that modified wNAFs are generated
12480 (which avoid length expansion in many cases).
12484 * Add a function EC_GROUP_check_discriminant() (defined via
12485 EC_METHOD) that verifies that the curve discriminant is non-zero.
12487 Add a function EC_GROUP_check() that makes some sanity tests
12488 on a EC_GROUP, its generator and order. This includes
12489 EC_GROUP_check_discriminant().
12491 *Nils Larsch <nla@trustcenter.de>*
12493 * Add ECDSA in new directory crypto/ecdsa/.
12495 Add applications 'openssl ecparam' and 'openssl ecdsa'
12496 (these are based on 'openssl dsaparam' and 'openssl dsa').
12498 ECDSA support is also included in various other files across the
12499 library. Most notably,
12500 - 'openssl req' now has a '-newkey ecdsa:file' option;
12501 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12502 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12503 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12504 them suitable for ECDSA where domain parameters must be
12505 extracted before the specific public key;
12506 - ECDSA engine support has been added.
12508 *Nils Larsch <nla@trustcenter.de>*
12510 * Include some named elliptic curves, and add OIDs from X9.62,
12511 SECG, and WAP/WTLS. Each curve can be obtained from the new
12513 EC_GROUP_new_by_curve_name(),
12514 and the list of available named curves can be obtained with
12515 EC_get_builtin_curves().
12516 Also add a 'curve_name' member to EC_GROUP objects, which can be
12518 EC_GROUP_set_curve_name()
12519 EC_GROUP_get_curve_name()
12521 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12523 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12524 was actually never needed) and in BN_mul(). The removal in BN_mul()
12525 required a small change in bn_mul_part_recursive() and the addition
12526 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12527 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12528 bn_sub_words() and bn_add_words() except they take arrays with
12533 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12535 * Cleanse PEM buffers before freeing them since they may contain
12538 *Benjamin Bennett <ben@psc.edu>*
12540 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12541 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12542 authentication-only ciphersuites.
12546 * Since AES128 and AES256 share a single mask bit in the logic of
12547 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12548 kludge to work properly if AES128 is available and AES256 isn't.
12552 * Expand security boundary to match 1.1.1 module.
12556 * Remove redundant features: hash file source, editing of test vectors
12557 modify fipsld to use external fips_premain.c signature.
12561 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12562 run algorithm test programs.
12566 * Make algorithm test programs more tolerant of whitespace.
12570 * Have SSL/TLS server implementation tolerate "mismatched" record
12571 protocol version while receiving ClientHello even if the
12572 ClientHello is fragmented. (The server can't insist on the
12573 particular protocol version it has chosen before the ServerHello
12574 message has informed the client about his choice.)
12578 * Load error codes if they are not already present instead of using a
12579 static variable. This allows them to be cleanly unloaded and reloaded.
12583 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12585 * Introduce limits to prevent malicious keys being able to
12586 cause a denial of service. ([CVE-2006-2940])
12588 *Steve Henson, Bodo Moeller*
12590 * Fix ASN.1 parsing of certain invalid structures that can result
12591 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12593 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12594 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12596 * Fix SSL client code which could crash if connecting to a
12597 malicious SSLv2 server. ([CVE-2006-4343])
12599 *Tavis Ormandy and Will Drewry, Google Security Team*
12601 * Change ciphersuite string processing so that an explicit
12602 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12603 will no longer include "AES128-SHA"), and any other similar
12604 ciphersuite (same bitmap) from *other* protocol versions (so that
12605 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12606 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12607 changes from 0.9.8b and 0.9.8d.
12611 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12613 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12614 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12616 * Change the Unix randomness entropy gathering to use poll() when
12617 possible instead of select(), since the latter has some
12618 undesirable limitations.
12620 *Darryl Miles via Richard Levitte and Bodo Moeller*
12622 * Disable rogue ciphersuites:
12624 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12625 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12626 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12628 The latter two were purportedly from
12629 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12632 Also deactivate the remaining ciphersuites from
12633 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12634 unofficial, and the ID has long expired.
12638 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12639 dual-core machines) and other potential thread-safety issues.
12643 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12645 * Adapt fipsld and the build system to link against the validated FIPS
12646 module in FIPS mode.
12650 * Fixes for VC++ 2005 build under Windows.
12654 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12655 from a Windows bash shell such as MSYS. It is autodetected from the
12656 "config" script when run from a VC++ environment. Modify standard VC++
12657 build to use fipscanister.o from the GNU make build.
12661 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12663 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12664 The value now differs depending on if you build for FIPS or not.
12665 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12666 safely run with a non-FIPSed libcrypto, as it may crash because of
12667 the difference induced by this change.
12671 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12673 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12674 (part of SSL_OP_ALL). This option used to disable the
12675 countermeasure against man-in-the-middle protocol-version
12676 rollback in the SSL 2.0 server implementation, which is a bad
12677 idea. ([CVE-2005-2969])
12679 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12680 for Information Security, National Institute of Advanced Industrial
12681 Science and Technology [AIST, Japan)]*
12683 * Minimal support for X9.31 signatures and PSS padding modes. This is
12684 mainly for FIPS compliance and not fully integrated at this stage.
12688 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12689 the exponentiation using a fixed-length exponent. (Otherwise,
12690 the information leaked through timing could expose the secret key
12691 after many signatures; cf. Bleichenbacher's attack on DSA with
12696 * Make a new fixed-window mod_exp implementation the default for
12697 RSA, DSA, and DH private-key operations so that the sequence of
12698 squares and multiplies and the memory access pattern are
12699 independent of the particular secret key. This will mitigate
12700 cache-timing and potential related attacks.
12702 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12703 and this is automatically used by BN_mod_exp_mont() if the new flag
12704 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12705 will use this BN flag for private exponents unless the flag
12706 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12707 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12709 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12711 * Change the client implementation for SSLv23_method() and
12712 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12713 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12714 (Previously, the SSL 2.0 backwards compatible Client Hello
12715 message format would be used even with SSL_OP_NO_SSLv2.)
12719 * Add support for smime-type MIME parameter in S/MIME messages which some
12724 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12725 a threadsafe manner. Modify rsa code to use new function and add calls
12726 to dsa and dh code (which had race conditions before).
12730 * Include the fixed error library code in the C error file definitions
12731 instead of fixing them up at runtime. This keeps the error code
12732 structures constant.
12736 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12738 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12741 * Fixes for newer kerberos headers. NB: the casts are needed because
12742 the 'length' field is signed on one version and unsigned on another
12743 with no (?) obvious way to tell the difference, without these VC++
12744 complains. Also the "definition" of FAR (blank) is no longer included
12745 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12746 some needed definitions.
12750 * Undo Cygwin change.
12754 * Added support for proxy certificates according to RFC 3820.
12755 Because they may be a security thread to unaware applications,
12756 they must be explicitly allowed in run-time. See
12757 docs/HOWTO/proxy_certificates.txt for further information.
12761 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12763 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12764 server and client random values. Previously
12765 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12766 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12768 This change has negligible security impact because:
12770 1. Server and client random values still have 24 bytes of pseudo random
12773 2. Server and client random values are sent in the clear in the initial
12776 3. The master secret is derived using the premaster secret (48 bytes in
12777 size for static RSA ciphersuites) as well as client server and random
12780 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12783 *Stephen Henson, reported by UK NISCC*
12785 * Use Windows randomness collection on Cygwin.
12789 * Fix hang in EGD/PRNGD query when communication socket is closed
12790 prematurely by EGD/PRNGD.
12792 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12794 * Prompt for pass phrases when appropriate for PKCS12 input format.
12798 * Back-port of selected performance improvements from development
12799 branch, as well as improved support for PowerPC platforms.
12803 * Add lots of checks for memory allocation failure, error codes to indicate
12804 failure and freeing up memory if a failure occurs.
12806 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12808 * Add new -passin argument to dgst.
12812 * Perform some character comparisons of different types in X509_NAME_cmp:
12813 this is needed for some certificates that re-encode DNs into UTF8Strings
12814 (in violation of RFC3280) and can't or won't issue name rollover
12819 * Make an explicit check during certificate validation to see that
12820 the CA setting in each certificate on the chain is correct. As a
12821 side effect always do the following basic checks on extensions,
12822 not just when there's an associated purpose to the check:
12824 - if there is an unhandled critical extension (unless the user
12825 has chosen to ignore this fault)
12826 - if the path length has been exceeded (if one is set at all)
12827 - that certain extensions fit the associated purpose (if one has
12832 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12834 * Avoid a race condition when CRLs are checked in a multi threaded
12835 environment. This would happen due to the reordering of the revoked
12836 entries during signature checking and serial number lookup. Now the
12837 encoding is cached and the serial number sort performed under a lock.
12838 Add new STACK function sk_is_sorted().
12842 * Add Delta CRL to the extension code.
12846 * Various fixes to s3_pkt.c so alerts are sent properly.
12848 *David Holmes <d.holmes@f5.com>*
12850 * Reduce the chances of duplicate issuer name and serial numbers (in
12851 violation of RFC3280) using the OpenSSL certificate creation utilities.
12852 This is done by creating a random 64 bit value for the initial serial
12853 number when a serial number file is created or when a self signed
12854 certificate is created using 'openssl req -x509'. The initial serial
12855 number file is created using 'openssl x509 -next_serial' in CA.pl
12856 rather than being initialized to 1.
12860 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12862 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12863 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12865 *Joe Orton, Steve Henson*
12867 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12870 *Joe Orton, Steve Henson*
12872 * Make it possible to have multiple active certificates with the same
12873 subject in the CA index file. This is done only if the keyword
12874 'unique_subject' is set to 'no' in the main CA section (default
12875 if 'CA_default') of the configuration file. The value is saved
12876 with the database itself in a separate index attribute file,
12877 named like the index file with '.attr' appended to the name.
12881 * X509 verify fixes. Disable broken certificate workarounds when
12882 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12883 keyUsage extension present. Don't accept CRLs with unhandled critical
12884 extensions: since verify currently doesn't process CRL extensions this
12885 rejects a CRL with *any* critical extensions. Add new verify error codes
12890 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12891 A clarification of RFC2560 will require the use of OCTET STRINGs and
12892 some implementations cannot handle the current raw format. Since OpenSSL
12893 copies and compares OCSP nonces as opaque blobs without any attempt at
12894 parsing them this should not create any compatibility issues.
12898 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12899 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12900 this HMAC (and other) operations are several times slower than OpenSSL
12905 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12907 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12909 * Use the correct content when signing type "other".
12913 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12915 * Fix various bugs revealed by running the NISCC test suite:
12917 Stop out of bounds reads in the ASN1 code when presented with
12918 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12920 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12922 If verify callback ignores invalid public key errors don't try to check
12923 certificate signature with the NULL public key.
12927 * New -ignore_err option in ocsp application to stop the server
12928 exiting on the first error in a request.
12932 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12933 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12938 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12939 extra data after the compression methods not only for TLS 1.0
12940 but also for SSL 3.0 (as required by the specification).
12942 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12944 * Change X509_certificate_type() to mark the key as exported/exportable
12945 when it's 512 *bits* long, not 512 bytes.
12949 * Change AES_cbc_encrypt() so it outputs exact multiple of
12950 blocks during encryption.
12954 * Various fixes to base64 BIO and non blocking I/O. On write
12955 flushes were not handled properly if the BIO retried. On read
12956 data was not being buffered properly and had various logic bugs.
12957 This also affects blocking I/O when the data being decoded is a
12962 * Various S/MIME bugfixes and compatibility changes:
12963 output correct application/pkcs7 MIME type if
12964 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12965 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12966 of files as .eml work). Correctly handle very long lines in MIME
12971 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12973 * Countermeasure against the Klima-Pokorny-Rosa extension of
12974 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12975 a protocol version number mismatch like a decryption error
12976 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12980 * Turn on RSA blinding by default in the default implementation
12981 to avoid a timing attack. Applications that don't want it can call
12982 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12983 They would be ill-advised to do so in most cases.
12985 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12987 * Change RSA blinding code so that it works when the PRNG is not
12988 seeded (in this case, the secret RSA exponent is abused as
12989 an unpredictable seed -- if it is not unpredictable, there
12990 is no point in blinding anyway). Make RSA blinding thread-safe
12991 by remembering the creator's thread ID in rsa->blinding and
12992 having all other threads use local one-time blinding factors
12993 (this requires more computation than sharing rsa->blinding, but
12994 avoids excessive locking; and if an RSA object is not shared
12995 between threads, blinding will still be very fast).
12999 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13000 ENGINE as defaults for all supported algorithms irrespective of
13001 the 'flags' parameter. 'flags' is now honoured, so applications
13002 should make sure they are passing it correctly.
13006 * Target "mingw" now allows native Windows code to be generated in
13007 the Cygwin environment as well as with the MinGW compiler.
13011 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13013 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13014 via timing by performing a MAC computation even if incorrect
13015 block cipher padding has been found. This is a countermeasure
13016 against active attacks where the attacker has to distinguish
13017 between bad padding and a MAC verification error. ([CVE-2003-0078])
13019 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13020 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13021 Martin Vuagnoux (EPFL, Ilion)*
13023 * Make the no-err option work as intended. The intention with no-err
13024 is not to have the whole error stack handling routines removed from
13025 libcrypto, it's only intended to remove all the function name and
13026 reason texts, thereby removing some of the footprint that may not
13027 be interesting if those errors aren't displayed anyway.
13029 NOTE: it's still possible for any application or module to have its
13030 own set of error texts inserted. The routines are there, just not
13031 used by default when no-err is given.
13035 * Add support for FreeBSD on IA64.
13037 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13039 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13040 Kerberos function mit_des_cbc_cksum(). Before this change,
13041 the value returned by DES_cbc_cksum() was like the one from
13042 mit_des_cbc_cksum(), except the bytes were swapped.
13044 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13046 * Allow an application to disable the automatic SSL chain building.
13047 Before this a rather primitive chain build was always performed in
13048 ssl3_output_cert_chain(): an application had no way to send the
13049 correct chain if the automatic operation produced an incorrect result.
13051 Now the chain builder is disabled if either:
13053 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13055 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13057 The reasoning behind this is that an application would not want the
13058 auto chain building to take place if extra chain certificates are
13059 present and it might also want a means of sending no additional
13060 certificates (for example the chain has two certificates and the
13065 * Add the possibility to build without the ENGINE framework.
13067 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13069 * Under Win32 gmtime() can return NULL: check return value in
13070 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13074 * DSA routines: under certain error conditions uninitialized BN objects
13075 could be freed. Solution: make sure initialization is performed early
13076 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13077 Nils Larsch <nla@trustcenter.de> via PR#459)
13081 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13082 checked on reconnect on the client side, therefore session resumption
13083 could still fail with a "ssl session id is different" error. This
13084 behaviour is masked when SSL_OP_ALL is used due to
13085 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13086 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13087 followup to PR #377.
13091 * IA-32 assembler support enhancements: unified ELF targets, support
13092 for SCO/Caldera platforms, fix for Cygwin shared build.
13096 * Add support for FreeBSD on sparc64. As a consequence, support for
13097 FreeBSD on non-x86 processors is separate from x86 processors on
13098 the config script, much like the NetBSD support.
13100 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13102 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13104 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13107 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13108 code (06) was taken as the first octet of the session ID and the last
13109 octet was ignored consequently. As a result SSLv2 client side session
13110 caching could not have worked due to the session ID mismatch between
13112 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13117 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13118 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13123 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13124 seems that in spite of existing for more than a year, many application
13125 author have done nothing to provide the necessary callbacks, which
13126 means that this particular engine will not work properly anywhere.
13127 This is a very unfortunate situation which forces us, in the name
13128 of usability, to give the hw_ncipher.c a static lock, which is part
13130 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13131 appear in 0.9.8 or later. We EXPECT application authors to have
13132 dealt properly with this when 0.9.8 is released (unless we actually
13133 make such changes in the libcrypto locking code that changes will
13134 have to be made anyway).
13138 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13139 octets have been read, EOF or an error occurs. Without this change
13140 some truncated ASN1 structures will not produce an error.
13144 * Disable Heimdal support, since it hasn't been fully implemented.
13145 Still give the possibility to force the use of Heimdal, but with
13146 warnings and a request that patches get sent to openssl-dev.
13150 * Add the VC-CE target, introduce the WINCE sysname, and add
13151 INSTALL.WCE and appropriate conditionals to make it build.
13153 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13155 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13156 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13157 edit numbers of the version.
13159 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13161 * Introduce safe string copy and catenation functions
13162 (BUF_strlcpy() and BUF_strlcat()).
13164 *Ben Laurie (CHATS) and Richard Levitte*
13166 * Avoid using fixed-size buffers for one-line DNs.
13168 *Ben Laurie (CHATS)*
13170 * Add BUF_MEM_grow_clean() to avoid information leakage when
13171 resizing buffers containing secrets, and use where appropriate.
13173 *Ben Laurie (CHATS)*
13175 * Avoid using fixed size buffers for configuration file location.
13177 *Ben Laurie (CHATS)*
13179 * Avoid filename truncation for various CA files.
13181 *Ben Laurie (CHATS)*
13183 * Use sizeof in preference to magic numbers.
13185 *Ben Laurie (CHATS)*
13187 * Avoid filename truncation in cert requests.
13189 *Ben Laurie (CHATS)*
13191 * Add assertions to check for (supposedly impossible) buffer
13194 *Ben Laurie (CHATS)*
13196 * Don't cache truncated DNS entries in the local cache (this could
13197 potentially lead to a spoofing attack).
13199 *Ben Laurie (CHATS)*
13201 * Fix various buffers to be large enough for hex/decimal
13202 representations in a platform independent manner.
13204 *Ben Laurie (CHATS)*
13206 * Add CRYPTO_realloc_clean() to avoid information leakage when
13207 resizing buffers containing secrets, and use where appropriate.
13209 *Ben Laurie (CHATS)*
13211 * Add BIO_indent() to avoid much slightly worrying code to do
13214 *Ben Laurie (CHATS)*
13216 * Convert sprintf()/BIO_puts() to BIO_printf().
13218 *Ben Laurie (CHATS)*
13220 * buffer_gets() could terminate with the buffer only half
13223 *Ben Laurie (CHATS)*
13225 * Add assertions to prevent user-supplied crypto functions from
13226 overflowing internal buffers by having large block sizes, etc.
13228 *Ben Laurie (CHATS)*
13230 * New OPENSSL_assert() macro (similar to assert(), but enabled
13233 *Ben Laurie (CHATS)*
13235 * Eliminate unused copy of key in RC4.
13237 *Ben Laurie (CHATS)*
13239 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13241 *Ben Laurie (CHATS)*
13243 * Fix off-by-one error in EGD path.
13245 *Ben Laurie (CHATS)*
13247 * If RANDFILE path is too long, ignore instead of truncating.
13249 *Ben Laurie (CHATS)*
13251 * Eliminate unused and incorrectly sized X.509 structure
13254 *Ben Laurie (CHATS)*
13256 * Eliminate unused and dangerous function knumber().
13258 *Ben Laurie (CHATS)*
13260 * Eliminate unused and dangerous structure, KSSL_ERR.
13262 *Ben Laurie (CHATS)*
13264 * Protect against overlong session ID context length in an encoded
13265 session object. Since these are local, this does not appear to be
13268 *Ben Laurie (CHATS)*
13270 * Change from security patch (see 0.9.6e below) that did not affect
13271 the 0.9.6 release series:
13273 Remote buffer overflow in SSL3 protocol - an attacker could
13274 supply an oversized master key in Kerberos-enabled versions.
13277 *Ben Laurie (CHATS)*
13279 * Change the SSL kerb5 codes to match RFC 2712.
13283 * Make -nameopt work fully for req and add -reqopt switch.
13285 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13287 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13289 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13291 * Make sure tests can be performed even if the corresponding algorithms
13292 have been removed entirely. This was also the last step to make
13293 OpenSSL compilable with DJGPP under all reasonable conditions.
13295 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13297 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13298 to allow version independent disabling of normally unselected ciphers,
13299 which may be activated as a side-effect of selecting a single cipher.
13301 (E.g., cipher list string "RSA" enables ciphersuites that are left
13302 out of "ALL" because they do not provide symmetric encryption.
13303 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13305 *Lutz Jaenicke, Bodo Moeller*
13307 * Add appropriate support for separate platform-dependent build
13308 directories. The recommended way to make a platform-dependent
13309 build directory is the following (tested on Linux), maybe with
13312 # Place yourself outside of the OpenSSL source tree. In
13313 # this example, the environment variable OPENSSL_SOURCE
13314 # is assumed to contain the absolute OpenSSL source directory.
13315 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13316 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13317 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13318 mkdir -p `dirname $F`
13319 ln -s $OPENSSL_SOURCE/$F $F
13322 To be absolutely sure not to disturb the source tree, a "make clean"
13323 is a good thing. If it isn't successful, don't worry about it,
13324 it probably means the source directory is very clean.
13328 * Make sure any ENGINE control commands make local copies of string
13329 pointers passed to them whenever necessary. Otherwise it is possible
13330 the caller may have overwritten (or deallocated) the original string
13331 data when a later ENGINE operation tries to use the stored values.
13333 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13335 * Improve diagnostics in file reading and command-line digests.
13337 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13339 * Add AES modes CFB and OFB to the object database. Correct an
13340 error in AES-CFB decryption.
13344 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13345 allows existing EVP_CIPHER_CTX structures to be reused after
13346 calling `EVP_*Final()`. This behaviour is used by encryption
13347 BIOs and some applications. This has the side effect that
13348 applications must explicitly clean up cipher contexts with
13349 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13353 * Check the values of dna and dnb in bn_mul_recursive before calling
13354 bn_mul_comba (a non zero value means the a or b arrays do not contain
13355 n2 elements) and fallback to bn_mul_normal if either is not zero.
13359 * Fix escaping of non-ASCII characters when using the -subj option
13360 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13364 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13365 form for "surname", serialNumber has no short form.
13366 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13367 therefore remove "mail" short name for "internet 7".
13368 The OID for unique identifiers in X509 certificates is
13369 x500UniqueIdentifier, not uniqueIdentifier.
13370 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13374 * Add an "init" command to the ENGINE config module and auto initialize
13375 ENGINEs. Without any "init" command the ENGINE will be initialized
13376 after all ctrl commands have been executed on it. If init=1 the
13377 ENGINE is initialized at that point (ctrls before that point are run
13378 on the uninitialized ENGINE and after on the initialized one). If
13379 init=0 then the ENGINE will not be initialized at all.
13383 * Fix the 'app_verify_callback' interface so that the user-defined
13384 argument is actually passed to the callback: In the
13385 SSL_CTX_set_cert_verify_callback() prototype, the callback
13386 declaration has been changed from
13389 int (*cb)(X509_STORE_CTX *,void *);
13390 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13391 i=s->ctx->app_verify_callback(&ctx)
13392 has been changed into
13393 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13395 To update applications using SSL_CTX_set_cert_verify_callback(),
13396 a dummy argument can be added to their callback functions.
13398 *D. K. Smetters <smetters@parc.xerox.com>*
13400 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13402 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13404 * Add and OPENSSL_LOAD_CONF define which will cause
13405 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13406 This allows older applications to transparently support certain
13407 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13408 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13409 load the config file and OPENSSL_add_all_algorithms_conf() which will
13410 always load it have also been added.
13414 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13415 Adjust NIDs and EVP layer.
13417 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13419 * Config modules support in openssl utility.
13421 Most commands now load modules from the config file,
13422 though in a few (such as version) this isn't done
13423 because it couldn't be used for anything.
13425 In the case of ca and req the config file used is
13426 the same as the utility itself: that is the -config
13427 command line option can be used to specify an
13432 * Move default behaviour from OPENSSL_config(). If appname is NULL
13433 use "openssl_conf" if filename is NULL use default openssl config file.
13437 * Add an argument to OPENSSL_config() to allow the use of an alternative
13438 config section name. Add a new flag to tolerate a missing config file
13439 and move code to CONF_modules_load_file().
13443 * Support for crypto accelerator cards from Accelerated Encryption
13444 Processing, www.aep.ie. (Use engine 'aep')
13445 The support was copied from 0.9.6c [engine] and adapted/corrected
13446 to work with the new engine framework.
13448 *AEP Inc. and Richard Levitte*
13450 * Support for SureWare crypto accelerator cards from Baltimore
13451 Technologies. (Use engine 'sureware')
13452 The support was copied from 0.9.6c [engine] and adapted
13453 to work with the new engine framework.
13457 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13458 make the newer ENGINE framework commands for the CHIL engine work.
13460 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13462 * Make it possible to produce shared libraries on ReliantUNIX.
13464 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13466 * Add the configuration target debug-linux-ppro.
13467 Make 'openssl rsa' use the general key loading routines
13468 implemented in `apps.c`, and make those routines able to
13469 handle the key format FORMAT_NETSCAPE and the variant
13472 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13474 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13476 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13478 * Add -keyform to rsautl, and document -engine.
13480 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13482 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13483 BIO_R_NO_SUCH_FILE error code rather than the generic
13484 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13488 * Add new functions
13489 ERR_peek_last_error
13490 ERR_peek_last_error_line
13491 ERR_peek_last_error_line_data.
13492 These are similar to
13494 ERR_peek_error_line
13495 ERR_peek_error_line_data,
13496 but report on the latest error recorded rather than the first one
13497 still in the error queue.
13499 *Ben Laurie, Bodo Moeller*
13501 * default_algorithms option in ENGINE config module. This allows things
13503 default_algorithms = ALL
13504 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13508 * Preliminary ENGINE config module.
13512 * New experimental application configuration code.
13516 * Change the AES code to follow the same name structure as all other
13517 symmetric ciphers, and behave the same way. Move everything to
13518 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13520 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13522 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13524 *Ben Laurie and Theo de Raadt*
13526 * Add option to output public keys in req command.
13528 *Massimiliano Pala madwolf@openca.org*
13530 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13531 (up to about 10% better than before for P-192 and P-224).
13535 * New functions/macros
13537 SSL_CTX_set_msg_callback(ctx, cb)
13538 SSL_CTX_set_msg_callback_arg(ctx, arg)
13539 SSL_set_msg_callback(ssl, cb)
13540 SSL_set_msg_callback_arg(ssl, arg)
13542 to request calling a callback function
13544 void cb(int write_p, int version, int content_type,
13545 const void *buf, size_t len, SSL *ssl, void *arg)
13547 whenever a protocol message has been completely received
13548 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13549 protocol version according to which the SSL library interprets
13550 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13551 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13552 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13553 specification (change_cipher_spec(20), alert(21), handshake(22)).
13554 'buf' and 'len' point to the actual message, 'ssl' to the
13555 SSL object, and 'arg' is the application-defined value set by
13556 SSL[_CTX]_set_msg_callback_arg().
13558 'openssl s_client' and 'openssl s_server' have new '-msg' options
13559 to enable a callback that displays all protocol messages.
13563 * Change the shared library support so shared libraries are built as
13564 soon as the corresponding static library is finished, and thereby get
13565 openssl and the test programs linked against the shared library.
13566 This still only happens when the keyword "shard" has been given to
13567 the configuration scripts.
13569 NOTE: shared library support is still an experimental thing, and
13570 backward binary compatibility is still not guaranteed.
13572 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13574 * Add support for Subject Information Access extension.
13576 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13578 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13579 additional bytes when new memory had to be allocated, not just
13580 when reusing an existing buffer.
13584 * New command line and configuration option 'utf8' for the req command.
13585 This allows field values to be specified as UTF8 strings.
13589 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13590 runs for the former and machine-readable output for the latter.
13594 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13595 of the e-mail address in the DN (i.e., it will go into a certificate
13596 extension only). The new configuration file option 'email_in_dn = no'
13597 has the same effect.
13599 *Massimiliano Pala madwolf@openca.org*
13601 * Change all functions with names starting with `des_` to be starting
13602 with `DES_` instead. Add wrappers that are compatible with libdes,
13603 but are named `_ossl_old_des_*`. Finally, add macros that map the
13604 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13605 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13606 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13609 Since we provide two compatibility mappings, the user needs to
13610 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13611 compatibility is desired. The default (i.e., when that macro
13612 isn't defined) is OpenSSL 0.9.6c compatibility.
13614 There are also macros that enable and disable the support of old
13615 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13616 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13617 are defined, the default will apply: to support the old des routines.
13619 In either case, one must include openssl/des.h to get the correct
13620 definitions. Do not try to just include openssl/des_old.h, that
13623 NOTE: This is a major break of an old API into a new one. Software
13624 authors are encouraged to switch to the `DES_` style functions. Some
13625 time in the future, des_old.h and the libdes compatibility functions
13626 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13627 default), and then completely removed.
13631 * Test for certificates which contain unsupported critical extensions.
13632 If such a certificate is found during a verify operation it is
13633 rejected by default: this behaviour can be overridden by either
13634 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13635 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13636 X509_supported_extension() has also been added which returns 1 if a
13637 particular extension is supported.
13641 * Modify the behaviour of EVP cipher functions in similar way to digests
13642 to retain compatibility with existing code.
13646 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13647 compatibility with existing code. In particular the 'ctx' parameter does
13648 not have to be to be initialized before the call to EVP_DigestInit() and
13649 it is tidied up after a call to EVP_DigestFinal(). New function
13650 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13651 EVP_MD_CTX_copy() changed to not require the destination to be
13652 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13653 requires the destination to be valid.
13655 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13656 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13660 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13661 so that complete 'Handshake' protocol structures are kept in memory
13662 instead of overwriting 'msg_type' and 'length' with 'body' data.
13666 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13668 *Massimo Santin via Richard Levitte*
13670 * Major restructuring to the underlying ENGINE code. This includes
13671 reduction of linker bloat, separation of pure "ENGINE" manipulation
13672 (initialisation, etc) from functionality dealing with implementations
13673 of specific crypto interfaces. This change also introduces integrated
13674 support for symmetric ciphers and digest implementations - so ENGINEs
13675 can now accelerate these by providing EVP_CIPHER and EVP_MD
13676 implementations of their own. This is detailed in
13677 [crypto/engine/README.md](crypto/engine/README.md)
13678 as it couldn't be adequately described here. However, there are a few
13679 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13680 were changed in the original introduction of ENGINE code have now
13681 reverted back - the hooking from this code to ENGINE is now a good
13682 deal more passive and at run-time, operations deal directly with
13683 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13684 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13685 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13686 they were not being used by the framework as there is no concept of a
13687 BIGNUM_METHOD and they could not be generalised to the new
13688 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13689 ENGINE_cpy() has been removed as it cannot be consistently defined in
13694 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13698 * Change mkdef.pl to sort symbols that get the same entry number,
13699 and make sure the automatically generated functions `ERR_load_*`
13700 become part of libeay.num as well.
13704 * New function SSL_renegotiate_pending(). This returns true once
13705 renegotiation has been requested (either SSL_renegotiate() call
13706 or HelloRequest/ClientHello received from the peer) and becomes
13707 false once a handshake has been completed.
13708 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13709 sends a HelloRequest, but does not ensure that a handshake takes
13710 place. SSL_renegotiate_pending() is useful for checking if the
13711 client has followed the request.)
13715 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13716 By default, clients may request session resumption even during
13717 renegotiation (if session ID contexts permit); with this option,
13718 session resumption is possible only in the first handshake.
13720 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13721 more bits available for options that should not be part of
13722 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13726 * Add some demos for certificate and certificate request creation.
13730 * Make maximum certificate chain size accepted from the peer application
13731 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13732 "Douglas E. Engert" <deengert@anl.gov>.
13736 * Add support for shared libraries for Unixware-7
13737 (Boyd Lynn Gerber <gerberb@zenez.com>).
13741 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13742 be done prior to destruction. Use this to unload error strings from
13743 ENGINEs that load their own error strings. NB: This adds two new API
13744 functions to "get" and "set" this destroy handler in an ENGINE.
13748 * Alter all existing ENGINE implementations (except "openssl" and
13749 "openbsd") to dynamically instantiate their own error strings. This
13750 makes them more flexible to be built both as statically-linked ENGINEs
13751 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13752 Also, add stub code to each that makes building them as self-contained
13753 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13757 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13758 implementations into applications that are completely implemented in
13759 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13760 commands that can be used to configure what shared-library to load and
13761 to control aspects of the way it is handled. Also, made an update to
13762 the [README-Engine.md](README-Engine.md) file
13763 that brings its information up-to-date and
13764 provides some information and instructions on the "dynamic" ENGINE
13765 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13769 * Make it possible to unload ranges of ERR strings with a new
13770 "ERR_unload_strings" function.
13774 * Add a copy() function to EVP_MD.
13778 * Make EVP_MD routines take a context pointer instead of just the
13779 md_data void pointer.
13783 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13784 that the digest can only process a single chunk of data
13785 (typically because it is provided by a piece of
13786 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13787 is only going to provide a single chunk of data, and hence the
13788 framework needn't accumulate the data for oneshot drivers.
13792 * As with "ERR", make it possible to replace the underlying "ex_data"
13793 functions. This change also alters the storage and management of global
13794 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13795 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13796 index counters. The API functions that use this state have been changed
13797 to take a "class_index" rather than pointers to the class's local STACK
13798 and counter, and there is now an API function to dynamically create new
13799 classes. This centralisation allows us to (a) plug a lot of the
13800 thread-safety problems that existed, and (b) makes it possible to clean
13801 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13802 such data would previously have always leaked in application code and
13803 workarounds were in place to make the memory debugging turn a blind eye
13804 to it. Application code that doesn't use this new function will still
13805 leak as before, but their memory debugging output will announce it now
13806 rather than letting it slide.
13808 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13809 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13810 has a return value to indicate success or failure.
13814 * Make it possible to replace the underlying "ERR" functions such that the
13815 global state (2 LHASH tables and 2 locks) is only used by the "default"
13816 implementation. This change also adds two functions to "get" and "set"
13817 the implementation prior to it being automatically set the first time
13818 any other ERR function takes place. Ie. an application can call "get",
13819 pass the return value to a module it has just loaded, and that module
13820 can call its own "set" function using that value. This means the
13821 module's "ERR" operations will use (and modify) the error state in the
13822 application and not in its own statically linked copy of OpenSSL code.
13826 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13827 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13828 the operation, and provides a more encapsulated way for external code
13829 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13830 to use these functions rather than manually incrementing the counts.
13832 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13836 * Add EVP test program.
13840 * Add symmetric cipher support to ENGINE. Expect the API to change!
13844 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13845 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13846 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13847 These allow a CRL to be built without having to access X509_CRL fields
13848 directly. Modify 'ca' application to use new functions.
13852 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13853 bug workarounds. Rollback attack detection is a security feature.
13854 The problem will only arise on OpenSSL servers when TLSv1 is not
13855 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13856 Software authors not wanting to support TLSv1 will have special reasons
13857 for their choice and can explicitly enable this option.
13859 *Bodo Moeller, Lutz Jaenicke*
13861 * Rationalise EVP so it can be extended: don't include a union of
13862 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13863 (similar to those existing for EVP_CIPHER_CTX).
13868 EVP_MD_CTX_init(&md); /* new function call */
13869 EVP_DigestInit(&md, EVP_sha1());
13870 EVP_DigestUpdate(&md, in, len);
13871 EVP_DigestFinal(&md, out, NULL);
13872 EVP_MD_CTX_cleanup(&md); /* new function call */
13876 * Make DES key schedule conform to the usual scheme, as well as
13877 correcting its structure. This means that calls to DES functions
13878 now have to pass a pointer to a des_key_schedule instead of a
13879 plain des_key_schedule (which was actually always a pointer
13882 des_key_schedule ks;
13884 des_set_key_checked(..., &ks);
13885 des_ncbc_encrypt(..., &ks, ...);
13887 (Note that a later change renames 'des_...' into 'DES_...'.)
13891 * Initial reduction of linker bloat: the use of some functions, such as
13892 PEM causes large amounts of unused functions to be linked in due to
13893 poor organisation. For example pem_all.c contains every PEM function
13894 which has a knock on effect of linking in large amounts of (unused)
13895 ASN1 code. Grouping together similar functions and splitting unrelated
13896 functions prevents this.
13900 * Cleanup of EVP macros.
13904 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13905 correct `_ecb suffix`.
13909 * Add initial OCSP responder support to ocsp application. The
13910 revocation information is handled using the text based index
13911 use by the ca application. The responder can either handle
13912 requests generated internally, supplied in files (for example
13913 via a CGI script) or using an internal minimal server.
13917 * Add configuration choices to get zlib compression for TLS.
13921 * Changes to Kerberos SSL for RFC 2712 compliance:
13922 1. Implemented real KerberosWrapper, instead of just using
13923 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13924 2. Implemented optional authenticator field of KerberosWrapper.
13926 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13927 and authenticator structs; see crypto/krb5/.
13929 Generalized Kerberos calls to support multiple Kerberos libraries.
13930 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13931 via Richard Levitte*
13933 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13934 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13935 values for each of the key sizes rather than having just
13936 parameters (and 'speed' generating keys each time).
13940 * Speed up EVP routines.
13943 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13944 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13945 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13946 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13948 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13949 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13950 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13953 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13955 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13959 * Added the OS2-EMX target.
13961 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13963 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13964 New functions to support `NCONF` routines in extension code.
13965 New function `CONF_set_nconf()`
13966 to allow functions which take an `NCONF` to also handle the old `LHASH`
13967 structure: this means that the old `CONF` compatible routines can be
13968 retained (in particular w.rt. extensions) without having to duplicate the
13969 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13973 * Enhance the general user interface with mechanisms for inner control
13974 and with possibilities to have yes/no kind of prompts.
13978 * Change all calls to low-level digest routines in the library and
13979 applications to use EVP. Add missing calls to HMAC_cleanup() and
13980 don't assume HMAC_CTX can be copied using memcpy().
13982 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13984 * Add the possibility to control engines through control names but with
13985 arbitrary arguments instead of just a string.
13986 Change the key loaders to take a UI_METHOD instead of a callback
13987 function pointer. NOTE: this breaks binary compatibility with earlier
13988 versions of OpenSSL [engine].
13989 Adapt the nCipher code for these new conditions and add a card insertion
13994 * Enhance the general user interface with mechanisms to better support
13995 dialog box interfaces, application-defined prompts, the possibility
13996 to use defaults (for example default passwords from somewhere else)
13997 and interrupts/cancellations.
14001 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14002 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14006 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14007 tidy up some unnecessarily weird code in 'sk_new()').
14009 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14011 * Change the key loading routines for ENGINEs to use the same kind
14012 callback (pem_password_cb) as all other routines that need this
14017 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14018 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14019 than this minimum value is recommended.
14023 * New random seeder for OpenVMS, using the system process statistics
14024 that are easily reachable.
14028 * Windows apparently can't transparently handle global
14029 variables defined in DLLs. Initialisations such as:
14031 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14033 won't compile. This is used by the any applications that need to
14034 declare their own ASN1 modules. This was fixed by adding the option
14035 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14036 needed for static libraries under Win32.
14040 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14041 setting of purpose and trust fields. New X509_STORE trust and
14042 purpose functions and tidy up setting in other SSL functions.
14046 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14047 structure. These are inherited by X509_STORE_CTX when it is
14048 initialised. This allows various defaults to be set in the
14049 X509_STORE structure (such as flags for CRL checking and custom
14050 purpose or trust settings) for functions which only use X509_STORE_CTX
14051 internally such as S/MIME.
14053 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14054 trust settings if they are not set in X509_STORE. This allows X509_STORE
14055 purposes and trust (in S/MIME for example) to override any set by default.
14057 Add command line options for CRL checking to smime, s_client and s_server
14062 * Initial CRL based revocation checking. If the CRL checking flag(s)
14063 are set then the CRL is looked up in the X509_STORE structure and
14064 its validity and signature checked, then if the certificate is found
14065 in the CRL the verify fails with a revoked error.
14067 Various new CRL related callbacks added to X509_STORE_CTX structure.
14069 Command line options added to 'verify' application to support this.
14071 This needs some additional work, such as being able to handle multiple
14072 CRLs with different times, extension based lookup (rather than just
14073 by subject name) and ultimately more complete V2 CRL extension
14078 * Add a general user interface API (crypto/ui/). This is designed
14079 to replace things like des_read_password and friends (backward
14080 compatibility functions using this new API are provided).
14081 The purpose is to remove prompting functions from the DES code
14082 section as well as provide for prompting through dialog boxes in
14083 a window system and the like.
14087 * Add "ex_data" support to ENGINE so implementations can add state at a
14088 per-structure level rather than having to store it globally.
14092 * Make it possible for ENGINE structures to be copied when retrieved by
14093 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14094 This causes the "original" ENGINE structure to act like a template,
14095 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14096 operational state can be localised to each ENGINE structure, despite the
14097 fact they all share the same "methods". New ENGINE structures returned in
14098 this case have no functional references and the return value is the single
14099 structural reference. This matches the single structural reference returned
14100 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14105 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14106 needs to match any other type at all we need to manually clear the
14111 * Changes to the "openssl engine" utility to include;
14112 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14113 about an ENGINE's available control commands.
14114 - executing control commands from command line arguments using the
14115 '-pre' and '-post' switches. '-post' is only used if '-t' is
14116 specified and the ENGINE is successfully initialised. The syntax for
14117 the individual commands are colon-separated, for example;
14118 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14122 * New dynamic control command support for ENGINEs. ENGINEs can now
14123 declare their own commands (numbers), names (strings), descriptions,
14124 and input types for run-time discovery by calling applications. A
14125 subset of these commands are implicitly classed as "executable"
14126 depending on their input type, and only these can be invoked through
14127 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14128 can be based on user input, config files, etc). The distinction is
14129 that "executable" commands cannot return anything other than a boolean
14130 result and can only support numeric or string input, whereas some
14131 discoverable commands may only be for direct use through
14132 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14133 pointers, or other custom uses. The "executable" commands are to
14134 support parameterisations of ENGINE behaviour that can be
14135 unambiguously defined by ENGINEs and used consistently across any
14136 OpenSSL-based application. Commands have been added to all the
14137 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14138 control over shared-library paths without source code alterations.
14142 * Changed all ENGINE implementations to dynamically allocate their
14143 ENGINEs rather than declaring them statically. Apart from this being
14144 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14145 this also allows the implementations to compile without using the
14146 internal engine_int.h header.
14150 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14151 'const' value. Any code that should be able to modify a RAND_METHOD
14152 should already have non-const pointers to it (ie. they should only
14153 modify their own ones).
14157 * Made a variety of little tweaks to the ENGINE code.
14158 - "atalla" and "ubsec" string definitions were moved from header files
14159 to C code. "nuron" string definitions were placed in variables
14160 rather than hard-coded - allowing parameterisation of these values
14161 later on via ctrl() commands.
14162 - Removed unused "#if 0"'d code.
14163 - Fixed engine list iteration code so it uses ENGINE_free() to release
14164 structural references.
14165 - Constified the RAND_METHOD element of ENGINE structures.
14166 - Constified various get/set functions as appropriate and added
14167 missing functions (including a catch-all ENGINE_cpy that duplicates
14168 all ENGINE values onto a new ENGINE except reference counts/state).
14169 - Removed NULL parameter checks in get/set functions. Setting a method
14170 or function to NULL is a way of cancelling out a previously set
14171 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14172 and doesn't justify the extra error symbols and code.
14173 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14174 flags from engine_int.h to engine.h.
14175 - Changed prototypes for ENGINE handler functions (init(), finish(),
14176 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14180 * Implement binary inversion algorithm for BN_mod_inverse in addition
14181 to the algorithm using long division. The binary algorithm can be
14182 used only if the modulus is odd. On 32-bit systems, it is faster
14183 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14184 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14185 up to 450 bits. In 64-bit environments, the binary algorithm
14186 appears to be advantageous for much longer moduli; here we use it
14187 for moduli up to 2048 bits.
14191 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14192 could not support the combine flag in choice fields.
14196 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14197 extensions from a certificate request to the certificate.
14201 * Allow multiple 'certopt' and 'nameopt' options to be separated
14202 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14203 file: this allows the display of the certificate about to be
14204 signed to be customised, to allow certain fields to be included
14205 or excluded and extension details. The old system didn't display
14206 multicharacter strings properly, omitted fields not in the policy
14207 and couldn't display additional details such as extensions.
14211 * Function EC_POINTs_mul for multiple scalar multiplication
14212 of an arbitrary number of elliptic curve points
14213 \sum scalars[i]*points[i],
14214 optionally including the generator defined for the EC_GROUP:
14215 scalar*generator + \sum scalars[i]*points[i].
14217 EC_POINT_mul is a simple wrapper function for the typical case
14218 that the point list has just one item (besides the optional
14223 * First EC_METHODs for curves over GF(p):
14225 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14226 operations and provides various method functions that can also
14227 operate with faster implementations of modular arithmetic.
14229 EC_GFp_mont_method() reuses most functions that are part of
14230 EC_GFp_simple_method, but uses Montgomery arithmetic.
14232 *Bodo Moeller; point addition and point doubling
14233 implementation directly derived from source code provided by
14234 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14236 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14237 crypto/ec/ec_lib.c):
14239 Curves are EC_GROUP objects (with an optional group generator)
14240 based on EC_METHODs that are built into the library.
14242 Points are EC_POINT objects based on EC_GROUP objects.
14244 Most of the framework would be able to handle curves over arbitrary
14245 finite fields, but as there are no obvious types for fields other
14246 than GF(p), some functions are limited to that for now.
14250 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14251 that the file contains a complete HTTP response.
14255 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14256 change the def and num file printf format specifier from "%-40sXXX"
14257 to "%-39s XXX". The latter will always guarantee a space after the
14258 field while the former will cause them to run together if the field
14259 is 40 of more characters long.
14263 * Constify the cipher and digest 'method' functions and structures
14264 and modify related functions to take constant EVP_MD and EVP_CIPHER
14269 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14270 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14274 * Modify `EVP_Digest*()` routines so they now return values. Although the
14275 internal software routines can never fail additional hardware versions
14280 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14282 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14283 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14286 ERR_R_NESTED_ASN1_ERROR
14288 ERR_R_MISSING_ASN1_EOS
14289 were 4 .. 9, conflicting with
14290 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14292 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14293 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14295 Add new error code 'ERR_R_INTERNAL_ERROR'.
14299 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14304 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14305 sets the subject name for a new request or supersedes the
14306 subject name in a given request. Formats that can be parsed are
14307 'CN=Some Name, OU=myOU, C=IT'
14309 'CN=Some Name/OU=myOU/C=IT'.
14311 Add options '-batch' and '-verbose' to 'openssl req'.
14313 *Massimiliano Pala <madwolf@hackmasters.net>*
14315 * Introduce the possibility to access global variables through
14316 functions on platform were that's the best way to handle exporting
14317 global variables in shared libraries. To enable this functionality,
14318 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14319 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14320 is normally done by Configure or something similar).
14322 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14323 in the source file (foo.c) like this:
14325 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14326 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14328 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14329 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14331 OPENSSL_DECLARE_GLOBAL(int,foo);
14332 #define foo OPENSSL_GLOBAL_REF(foo)
14333 OPENSSL_DECLARE_GLOBAL(double,bar);
14334 #define bar OPENSSL_GLOBAL_REF(bar)
14336 The #defines are very important, and therefore so is including the
14337 header file everywhere where the defined globals are used.
14339 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14340 of ASN.1 items, but that structure is a bit different.
14342 The largest change is in util/mkdef.pl which has been enhanced with
14343 better and easier to understand logic to choose which symbols should
14344 go into the Windows .def files as well as a number of fixes and code
14345 cleanup (among others, algorithm keywords are now sorted
14346 lexicographically to avoid constant rewrites).
14350 * In BN_div() keep a copy of the sign of 'num' before writing the
14351 result to 'rm' because if rm==num the value will be overwritten
14352 and produce the wrong result if 'num' is negative: this caused
14353 problems with BN_mod() and BN_nnmod().
14357 * Function OCSP_request_verify(). This checks the signature on an
14358 OCSP request and verifies the signer certificate. The signer
14359 certificate is just checked for a generic purpose and OCSP request
14364 * Add OCSP_check_validity() function to check the validity of OCSP
14365 responses. OCSP responses are prepared in real time and may only
14366 be a few seconds old. Simply checking that the current time lies
14367 between thisUpdate and nextUpdate max reject otherwise valid responses
14368 caused by either OCSP responder or client clock inaccuracy. Instead
14369 we allow thisUpdate and nextUpdate to fall within a certain period of
14370 the current time. The age of the response can also optionally be
14371 checked. Two new options -validity_period and -status_age added to
14376 * If signature or public key algorithm is unrecognized print out its
14377 OID rather that just UNKNOWN.
14381 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14382 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14383 ID to be generated from the issuer certificate alone which can then be
14384 passed to OCSP_id_issuer_cmp().
14388 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14389 ASN1 modules to export functions returning ASN1_ITEM pointers
14390 instead of the ASN1_ITEM structures themselves. This adds several
14391 new macros which allow the underlying ASN1 function/structure to
14392 be accessed transparently. As a result code should not use ASN1_ITEM
14393 references directly (such as &X509_it) but instead use the relevant
14394 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14395 use of the new ASN1 code on platforms where exporting structures
14396 is problematical (for example in shared libraries) but exporting
14397 functions returning pointers to structures is not.
14401 * Add support for overriding the generation of SSL/TLS session IDs.
14402 These callbacks can be registered either in an SSL_CTX or per SSL.
14403 The purpose of this is to allow applications to control, if they wish,
14404 the arbitrary values chosen for use as session IDs, particularly as it
14405 can be useful for session caching in multiple-server environments. A
14406 command-line switch for testing this (and any client code that wishes
14407 to use such a feature) has been added to "s_server".
14409 *Geoff Thorpe, Lutz Jaenicke*
14411 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14412 of the form `#if defined(...) || defined(...) || ...` and
14413 `#if !defined(...) && !defined(...) && ...`. This also avoids
14414 the growing number of special cases it was previously handling.
14418 * Make all configuration macros available for application by making
14419 sure they are available in opensslconf.h, by giving them names starting
14420 with `OPENSSL_` to avoid conflicts with other packages and by making
14421 sure e_os2.h will cover all platform-specific cases together with
14423 Additionally, it is now possible to define configuration/platform-
14424 specific names (called "system identities"). In the C code, these
14425 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14426 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14427 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14432 * New option -set_serial to 'req' and 'x509' this allows the serial
14433 number to use to be specified on the command line. Previously self
14434 signed certificates were hard coded with serial number 0 and the
14435 CA options of 'x509' had to use a serial number in a file which was
14440 * New options to 'ca' utility to support V2 CRL entry extensions.
14441 Currently CRL reason, invalidity date and hold instruction are
14442 supported. Add new CRL extensions to V3 code and some new objects.
14446 * New function EVP_CIPHER_CTX_set_padding() this is used to
14447 disable standard block padding (aka PKCS#5 padding) in the EVP
14448 API, which was previously mandatory. This means that the data is
14449 not padded in any way and so the total length much be a multiple
14450 of the block size, otherwise an error occurs.
14454 * Initial (incomplete) OCSP SSL support.
14458 * New function OCSP_parse_url(). This splits up a URL into its host,
14459 port and path components: primarily to parse OCSP URLs. New -url
14460 option to ocsp utility.
14464 * New nonce behavior. The return value of OCSP_check_nonce() now
14465 reflects the various checks performed. Applications can decide
14466 whether to tolerate certain situations such as an absent nonce
14467 in a response when one was present in a request: the ocsp application
14468 just prints out a warning. New function OCSP_add1_basic_nonce()
14469 this is to allow responders to include a nonce in a response even if
14470 the request is nonce-less.
14474 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14475 skipped when using openssl x509 multiple times on a single input file,
14476 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14480 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14481 set string type: to handle setting ASN1_TIME structures. Fix ca
14482 utility to correctly initialize revocation date of CRLs.
14486 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14487 the clients preferred ciphersuites and rather use its own preferences.
14488 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14489 Internet Explorer by ensuring unchanged hash method during stepup.
14490 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14494 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14495 to aes and add a new 'exist' option to print out symbols that don't
14500 * Additional options to ocsp utility to allow flags to be set and
14501 additional certificates supplied.
14505 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14506 OCSP client a number of certificate to only verify the response
14511 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14512 handle the new API. Currently only ECB, CBC modes supported. Add new
14515 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14516 Encryption Standard (AES) Ciphersuites for Transport Layer
14517 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14518 not enabled by default and were not part of the "ALL" ciphersuite
14519 alias because they were not yet official; they could be
14520 explicitly requested by specifying the "AESdraft" ciphersuite
14521 group alias. In the final release of OpenSSL 0.9.7, the group
14522 alias is called "AES" and is part of "ALL".)
14524 *Ben Laurie, Steve Henson, Bodo Moeller*
14526 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14527 request to response.
14531 * Functions for OCSP responders. OCSP_request_onereq_count(),
14532 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14533 extract information from a certificate request. OCSP_response_create()
14534 creates a response and optionally adds a basic response structure.
14535 OCSP_basic_add1_status() adds a complete single response to a basic
14536 response and returns the OCSP_SINGLERESP structure just added (to allow
14537 extensions to be included for example). OCSP_basic_add1_cert() adds a
14538 certificate to a basic response and OCSP_basic_sign() signs a basic
14539 response with various flags. New helper functions ASN1_TIME_check()
14540 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14541 (converts ASN1_TIME to GeneralizedTime).
14545 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14546 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14547 structure from a certificate. X509_pubkey_digest() digests the public_key
14548 contents: this is used in various key identifiers.
14552 * Make sk_sort() tolerate a NULL argument.
14554 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14556 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14557 passed by the function are trusted implicitly. If any of them signed the
14558 response then it is assumed to be valid and is not verified.
14562 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14563 to data. This was previously part of the PKCS7 ASN1 code. This
14564 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14565 *Steve Henson, reported by Kenneth R. Robinette
14566 <support@securenetterm.com>*
14568 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14569 routines: without these tracing memory leaks is very painful.
14570 Fix leaks in PKCS12 and PKCS7 routines.
14574 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14575 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14576 effectively meant GeneralizedTime would never be used. Now it
14577 is initialised to -1 but X509_time_adj() now has to check the value
14578 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14579 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14580 *Steve Henson, reported by Kenneth R. Robinette
14581 <support@securenetterm.com>*
14583 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14584 result in a zero length in the ASN1_INTEGER structure which was
14585 not consistent with the structure when d2i_ASN1_INTEGER() was used
14586 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14587 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14588 where it did not print out a minus for negative ASN1_INTEGER.
14592 * Add summary printout to ocsp utility. The various functions which
14593 convert status values to strings have been renamed to:
14594 OCSP_response_status_str(), OCSP_cert_status_str() and
14595 OCSP_crl_reason_str() and are no longer static. New options
14596 to verify nonce values and to disable verification. OCSP response
14597 printout format cleaned up.
14601 * Add additional OCSP certificate checks. These are those specified
14602 in RFC2560. This consists of two separate checks: the CA of the
14603 certificate being checked must either be the OCSP signer certificate
14604 or the issuer of the OCSP signer certificate. In the latter case the
14605 OCSP signer certificate must contain the OCSP signing extended key
14606 usage. This check is performed by attempting to match the OCSP
14607 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14608 in the OCSP_CERTID structures of the response.
14612 * Initial OCSP certificate verification added to OCSP_basic_verify()
14613 and related routines. This uses the standard OpenSSL certificate
14614 verify routines to perform initial checks (just CA validity) and
14615 to obtain the certificate chain. Then additional checks will be
14616 performed on the chain. Currently the root CA is checked to see
14617 if it is explicitly trusted for OCSP signing. This is used to set
14618 a root CA as a global signing root: that is any certificate that
14619 chains to that CA is an acceptable OCSP signing certificate.
14623 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14624 extensions from a separate configuration file.
14625 As when reading extensions from the main configuration file,
14626 the '-extensions ...' option may be used for specifying the
14629 *Massimiliano Pala <madwolf@comune.modena.it>*
14631 * New OCSP utility. Allows OCSP requests to be generated or
14632 read. The request can be sent to a responder and the output
14633 parsed, outputted or printed in text form. Not complete yet:
14634 still needs to check the OCSP response validity.
14638 * New subcommands for 'openssl ca':
14639 `openssl ca -status <serial>` prints the status of the cert with
14640 the given serial number (according to the index file).
14641 `openssl ca -updatedb` updates the expiry status of certificates
14644 *Massimiliano Pala <madwolf@comune.modena.it>*
14646 * New '-newreq-nodes' command option to CA.pl. This is like
14647 '-newreq', but calls 'openssl req' with the '-nodes' option
14648 so that the resulting key is not encrypted.
14650 *Damien Miller <djm@mindrot.org>*
14652 * New configuration for the GNU Hurd.
14654 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14656 * Initial code to implement OCSP basic response verify. This
14657 is currently incomplete. Currently just finds the signer's
14658 certificate and verifies the signature on the response.
14662 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14663 value of OPENSSLDIR. This is available via the new '-d' option
14664 to 'openssl version', and is also included in 'openssl version -a'.
14668 * Allowing defining memory allocation callbacks that will be given
14669 file name and line number information in additional arguments
14670 (a `const char*` and an int). The basic functionality remains, as
14671 well as the original possibility to just replace malloc(),
14672 realloc() and free() by functions that do not know about these
14673 additional arguments. To register and find out the current
14674 settings for extended allocation functions, the following
14675 functions are provided:
14677 CRYPTO_set_mem_ex_functions
14678 CRYPTO_set_locked_mem_ex_functions
14679 CRYPTO_get_mem_ex_functions
14680 CRYPTO_get_locked_mem_ex_functions
14682 These work the same way as CRYPTO_set_mem_functions and friends.
14683 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14684 extended allocation function is enabled.
14685 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14686 a conventional allocation function is enabled.
14688 *Richard Levitte, Bodo Moeller*
14690 * Finish off removing the remaining LHASH function pointer casts.
14691 There should no longer be any prototype-casting required when using
14692 the LHASH abstraction, and any casts that remain are "bugs". See
14693 the callback types and macros at the head of lhash.h for details
14694 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14698 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14699 If /dev/[u]random devices are not available or do not return enough
14700 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14702 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14703 /etc/entropy will be queried once each in this sequence, querying stops
14704 when enough entropy was collected without querying more sockets.
14708 * Change the Unix RAND_poll() variant to be able to poll several
14709 random devices, as specified by DEVRANDOM, until a sufficient amount
14710 of data has been collected. We spend at most 10 ms on each file
14711 (select timeout) and read in non-blocking mode. DEVRANDOM now
14712 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14713 (previously it was just the string "/dev/urandom"), so on typical
14714 platforms the 10 ms delay will never occur.
14715 Also separate out the Unix variant to its own file, rand_unix.c.
14716 For VMS, there's a currently-empty rand_vms.c.
14720 * Move OCSP client related routines to ocsp_cl.c. These
14721 provide utility functions which an application needing
14722 to issue a request to an OCSP responder and analyse the
14723 response will typically need: as opposed to those which an
14724 OCSP responder itself would need which will be added later.
14726 OCSP_request_sign() signs an OCSP request with an API similar
14727 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14728 response. OCSP_response_get1_basic() extracts basic response
14729 from response. OCSP_resp_find_status(): finds and extracts status
14730 information from an OCSP_CERTID structure (which will be created
14731 when the request structure is built). These are built from lower
14732 level functions which work on OCSP_SINGLERESP structures but
14733 won't normally be used unless the application wishes to examine
14734 extensions in the OCSP response for example.
14736 Replace nonce routines with a pair of functions.
14737 OCSP_request_add1_nonce() adds a nonce value and optionally
14738 generates a random value. OCSP_check_nonce() checks the
14739 validity of the nonce in an OCSP response.
14743 * Change function OCSP_request_add() to OCSP_request_add0_id().
14744 This doesn't copy the supplied OCSP_CERTID and avoids the
14745 need to free up the newly created id. Change return type
14746 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14747 This can then be used to add extensions to the request.
14748 Deleted OCSP_request_new(), since most of its functionality
14749 is now in OCSP_REQUEST_new() (and the case insensitive name
14750 clash) apart from the ability to set the request name which
14751 will be added elsewhere.
14755 * Update OCSP API. Remove obsolete extensions argument from
14756 various functions. Extensions are now handled using the new
14757 OCSP extension code. New simple OCSP HTTP function which
14758 can be used to send requests and parse the response.
14762 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14763 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14764 uses the special reorder version of SET OF to sort the attributes
14765 and reorder them to match the encoded order. This resolves a long
14766 standing problem: a verify on a PKCS7 structure just after signing
14767 it used to fail because the attribute order did not match the
14768 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14769 it uses the received order. This is necessary to tolerate some broken
14770 software that does not order SET OF. This is handled by encoding
14771 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14772 to produce the required SET OF.
14776 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14777 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14778 files to get correct declarations of the ASN.1 item variables.
14782 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14783 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14784 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14785 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14786 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14787 ASN1_ITEM and no wrapper functions.
14791 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14792 replace the old function pointer based I/O routines. Change most of
14793 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14797 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14798 lines, recognize more "algorithms" that can be deselected, and make
14799 it complain about algorithm deselection that isn't recognised.
14803 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14804 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14805 to use new functions. Add NO_ASN1_OLD which can be set to remove
14806 some old style ASN1 functions: this can be used to determine if old
14807 code will still work when these eventually go away.
14811 * New extension functions for OCSP structures, these follow the
14812 same conventions as certificates and CRLs.
14816 * New function X509V3_add1_i2d(). This automatically encodes and
14817 adds an extension. Its behaviour can be customised with various
14818 flags to append, replace or delete. Various wrappers added for
14819 certificates and CRLs.
14823 * Fix to avoid calling the underlying ASN1 print routine when
14824 an extension cannot be parsed. Correct a typo in the
14825 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14829 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14830 entries for variables.
14834 * Add functionality to `apps/openssl.c` for detecting locking
14835 problems: As the program is single-threaded, all we have
14836 to do is register a locking callback using an array for
14837 storing which locks are currently held by the program.
14841 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14842 SSL_get_ex_data_X509_STORE_idx(), which is used in
14843 ssl_verify_cert_chain() and thus can be called at any time
14844 during TLS/SSL handshakes so that thread-safety is essential.
14845 Unfortunately, the ex_data design is not at all suited
14846 for multi-threaded use, so it probably should be abolished.
14850 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14852 *Broadcom, tweaked and integrated by Geoff Thorpe*
14854 * Move common extension printing code to new function
14855 X509V3_print_extensions(). Reorganise OCSP print routines and
14856 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14860 * New function X509_signature_print() to remove duplication in some
14865 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14866 set (this was treated exactly the same as SET OF previously). This
14867 is used to reorder the STACK representing the structure to match the
14868 encoding. This will be used to get round a problem where a PKCS7
14869 structure which was signed could not be verified because the STACK
14870 order did not reflect the encoded order.
14874 * Reimplement the OCSP ASN1 module using the new code.
14878 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14879 for its ASN1 operations. The old style function pointers still exist
14880 for now but they will eventually go away.
14884 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14885 completely replaces the old ASN1 functionality with a table driven
14886 encoder and decoder which interprets an ASN1_ITEM structure describing
14887 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14888 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14889 has also been converted to the new form.
14893 * Change BN_mod_exp_recp so that negative moduli are tolerated
14894 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14895 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14896 for negative moduli.
14900 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14901 of not touching the result's sign bit.
14905 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14910 * Changed the LHASH code to use prototypes for callbacks, and created
14911 macros to declare and implement thin (optionally static) functions
14912 that provide type-safety and avoid function pointer casting for the
14913 type-specific callbacks.
14917 * Added Kerberos Cipher Suites to be used with TLS, as written in
14919 *Veers Staats <staatsvr@asc.hpc.mil>,
14920 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14922 * Reformat the FAQ so the different questions and answers can be divided
14923 in sections depending on the subject.
14927 * Have the zlib compression code load ZLIB.DLL dynamically under
14932 * New function BN_mod_sqrt for computing square roots modulo a prime
14933 (using the probabilistic Tonelli-Shanks algorithm unless
14934 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14935 be handled deterministically).
14937 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14939 * Make BN_mod_inverse faster by explicitly handling small quotients
14940 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14941 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14945 * New function BN_kronecker.
14949 * Fix BN_gcd so that it works on negative inputs; the result is
14950 positive unless both parameters are zero.
14951 Previously something reasonably close to an infinite loop was
14952 possible because numbers could be growing instead of shrinking
14953 in the implementation of Euclid's algorithm.
14957 * Fix BN_is_word() and BN_is_one() macros to take into account the
14958 sign of the number in question.
14960 Fix BN_is_word(a,w) to work correctly for w == 0.
14962 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14963 because its test if the absolute value of 'a' equals 'w'.
14964 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14965 it exists mostly for use in the implementations of BN_is_zero(),
14966 BN_is_one(), and BN_is_word().
14970 * New function BN_swap.
14974 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14975 the exponentiation functions are more likely to produce reasonable
14976 results on negative inputs.
14980 * Change BN_mod_mul so that the result is always non-negative.
14981 Previously, it could be negative if one of the factors was negative;
14982 I don't think anyone really wanted that behaviour.
14986 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14987 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14988 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14989 and add new functions:
14998 BN_mod_lshift1_quick
15000 BN_mod_lshift_quick
15002 These functions always generate non-negative results.
15004 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15005 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15007 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15008 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15009 be reduced modulo `m`.
15011 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15014 The following entry accidentally appeared in the CHANGES file
15015 distributed with OpenSSL 0.9.7. The modifications described in
15016 it do *not* apply to OpenSSL 0.9.7.
15018 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15019 was actually never needed) and in BN_mul(). The removal in BN_mul()
15020 required a small change in bn_mul_part_recursive() and the addition
15021 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15022 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15023 bn_sub_words() and bn_add_words() except they take arrays with
15029 * In 'openssl passwd', verify passwords read from the terminal
15030 unless the '-salt' option is used (which usually means that
15031 verification would just waste user's time since the resulting
15032 hash is going to be compared with some given password hash)
15033 or the new '-noverify' option is used.
15035 This is an incompatible change, but it does not affect
15036 non-interactive use of 'openssl passwd' (passwords on the command
15037 line, '-stdin' option, '-in ...' option) and thus should not
15038 cause any problems.
15042 * Remove all references to RSAref, since there's no more need for it.
15046 * Make DSO load along a path given through an environment variable
15047 (SHLIB_PATH) with shl_load().
15051 * Constify the ENGINE code as a result of BIGNUM constification.
15052 Also constify the RSA code and most things related to it. In a
15053 few places, most notable in the depth of the ASN.1 code, ugly
15054 casts back to non-const were required (to be solved at a later
15059 * Make it so the openssl application has all engines loaded by default.
15063 * Constify the BIGNUM routines a little more.
15067 * Add the following functions:
15069 ENGINE_load_cswift()
15071 ENGINE_load_atalla()
15072 ENGINE_load_nuron()
15073 ENGINE_load_builtin_engines()
15075 That way, an application can itself choose if external engines that
15076 are built-in in OpenSSL shall ever be used or not. The benefit is
15077 that applications won't have to be linked with libdl or other dso
15078 libraries unless it's really needed.
15080 Changed 'openssl engine' to load all engines on demand.
15081 Changed the engine header files to avoid the duplication of some
15082 declarations (they differed!).
15086 * 'openssl engine' can now list capabilities.
15090 * Better error reporting in 'openssl engine'.
15094 * Never call load_dh_param(NULL) in s_server.
15098 * Add engine application. It can currently list engines by name and
15099 identity, and test if they are actually available.
15103 * Improve RPM specification file by forcing symbolic linking and making
15104 sure the installed documentation is also owned by root.root.
15106 *Damien Miller <djm@mindrot.org>*
15108 * Give the OpenSSL applications more possibilities to make use of
15109 keys (public as well as private) handled by engines.
15113 * Add OCSP code that comes from CertCo.
15117 * Add VMS support for the Rijndael code.
15121 * Added untested support for Nuron crypto accelerator.
15125 * Add support for external cryptographic devices. This code was
15126 previously distributed separately as the "engine" branch.
15128 *Geoff Thorpe, Richard Levitte*
15130 * Rework the filename-translation in the DSO code. It is now possible to
15131 have far greater control over how a "name" is turned into a filename
15132 depending on the operating environment and any oddities about the
15133 different shared library filenames on each system.
15137 * Support threads on FreeBSD-elf in Configure.
15141 * Fix for SHA1 assembly problem with MASM: it produces
15142 warnings about corrupt line number information when assembling
15143 with debugging information. This is caused by the overlapping
15146 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15149 NCONF_get_number() has no error checking at all. As a replacement,
15150 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15151 promoted strongly. The old NCONF_get_number is kept around for
15152 binary backward compatibility.
15153 Make it possible for methods to load from something other than a BIO,
15154 by providing a function pointer that is given a name instead of a BIO.
15155 For example, this could be used to load configuration data from an
15160 * Fix for non blocking accept BIOs. Added new I/O special reason
15161 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15162 with non blocking I/O was not possible because no retry code was
15163 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15168 * Added the beginnings of Rijndael support.
15172 * Fix for bug in DirectoryString mask setting. Add support for
15173 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15174 to allow certificate printing to more controllable, additional
15175 'certopt' option to 'x509' to allow new printing options to be
15180 * Clean old EAY MD5 hack from e_os.h.
15184 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15186 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15187 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15189 *Joe Orton, Steve Henson*
15191 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15193 * Fix additional bug revealed by the NISCC test suite:
15195 Stop bug triggering large recursion when presented with
15196 certain ASN.1 tags ([CVE-2003-0851])
15200 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15202 * Fix various bugs revealed by running the NISCC test suite:
15204 Stop out of bounds reads in the ASN1 code when presented with
15205 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15207 If verify callback ignores invalid public key errors don't try to check
15208 certificate signature with the NULL public key.
15212 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15213 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15218 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15219 extra data after the compression methods not only for TLS 1.0
15220 but also for SSL 3.0 (as required by the specification).
15222 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15224 * Change X509_certificate_type() to mark the key as exported/exportable
15225 when it's 512 *bits* long, not 512 bytes.
15229 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15231 * Countermeasure against the Klima-Pokorny-Rosa extension of
15232 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15233 a protocol version number mismatch like a decryption error
15234 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15238 * Turn on RSA blinding by default in the default implementation
15239 to avoid a timing attack. Applications that don't want it can call
15240 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15241 They would be ill-advised to do so in most cases.
15243 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15245 * Change RSA blinding code so that it works when the PRNG is not
15246 seeded (in this case, the secret RSA exponent is abused as
15247 an unpredictable seed -- if it is not unpredictable, there
15248 is no point in blinding anyway). Make RSA blinding thread-safe
15249 by remembering the creator's thread ID in rsa->blinding and
15250 having all other threads use local one-time blinding factors
15251 (this requires more computation than sharing rsa->blinding, but
15252 avoids excessive locking; and if an RSA object is not shared
15253 between threads, blinding will still be very fast).
15257 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15259 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15260 via timing by performing a MAC computation even if incorrect
15261 block cipher padding has been found. This is a countermeasure
15262 against active attacks where the attacker has to distinguish
15263 between bad padding and a MAC verification error. ([CVE-2003-0078])
15265 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15266 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15267 Martin Vuagnoux (EPFL, Ilion)*
15269 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15271 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15272 memory from its contents. This is done with a counter that will
15273 place alternating values in each byte. This can be used to solve
15274 two issues: 1) the removal of calls to memset() by highly optimizing
15275 compilers, and 2) cleansing with other values than 0, since those can
15276 be read through on certain media, for example a swap space on disk.
15280 * Bugfix: client side session caching did not work with external caching,
15281 because the session->cipher setting was not restored when reloading
15282 from the external cache. This problem was masked, when
15283 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15284 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15288 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15289 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15291 *Zeev Lieber <zeev-l@yahoo.com>*
15293 * Undo an undocumented change introduced in 0.9.6e which caused
15294 repeated calls to OpenSSL_add_all_ciphers() and
15295 OpenSSL_add_all_digests() to be ignored, even after calling
15300 * Change the default configuration reader to deal with last line not
15301 being properly terminated.
15305 * Change X509_NAME_cmp() so it applies the special rules on handling
15306 DN values that are of type PrintableString, as well as RDNs of type
15307 emailAddress where the value has the type ia5String.
15309 *stefank@valicert.com via Richard Levitte*
15311 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15312 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15313 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15314 the bitwise-OR of the two for use by the majority of applications
15315 wanting this behaviour, and update the docs. The documented
15316 behaviour and actual behaviour were inconsistent and had been
15317 changing anyway, so this is more a bug-fix than a behavioural
15320 *Geoff Thorpe, diagnosed by Nadav Har'El*
15322 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15323 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15327 * Fix initialization code race conditions in
15328 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15329 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15330 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15331 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15332 ssl2_get_cipher_by_char(),
15333 ssl3_get_cipher_by_char().
15335 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15337 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15338 the cached sessions are flushed, as the remove_cb() might use ex_data
15339 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15340 (see [openssl.org #212]).
15342 *Geoff Thorpe, Lutz Jaenicke*
15344 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15345 length, instead of the encoding length to d2i_ASN1_OBJECT.
15349 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15351 * [In 0.9.6g-engine release:]
15352 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15354 *Lynn Gazis <lgazis@rainbow.com>*
15356 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15358 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15359 and get fix the header length calculation.
15360 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15361 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15363 * Use proper error handling instead of 'assertions' in buffer
15364 overflow checks added in 0.9.6e. This prevents DoS (the
15365 assertions could call abort()).
15367 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15369 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15371 * Add various sanity checks to asn1_get_length() to reject
15372 the ASN1 length bytes if they exceed sizeof(long), will appear
15373 negative or the content length exceeds the length of the
15376 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15378 * Fix cipher selection routines: ciphers without encryption had no flags
15379 for the cipher strength set and where therefore not handled correctly
15380 by the selection routines (PR #130).
15384 * Fix EVP_dsa_sha macro.
15389 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15390 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15391 that was added in OpenSSL 0.9.6d.
15393 As the countermeasure turned out to be incompatible with some
15394 broken SSL implementations, the new option is part of SSL_OP_ALL.
15395 SSL_OP_ALL is usually employed when compatibility with weird SSL
15396 implementations is desired (e.g. '-bugs' option to 's_client' and
15397 's_server'), so the new option is automatically set in many
15402 * Changes in security patch:
15404 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15405 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15406 Air Force Materiel Command, USAF, under agreement number
15409 * Add various sanity checks to asn1_get_length() to reject
15410 the ASN1 length bytes if they exceed sizeof(long), will appear
15411 negative or the content length exceeds the length of the
15412 supplied buffer. ([CVE-2002-0659])
15414 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15416 * Assertions for various potential buffer overflows, not known to
15417 happen in practice.
15419 *Ben Laurie (CHATS)*
15421 * Various temporary buffers to hold ASCII versions of integers were
15422 too small for 64 bit platforms. ([CVE-2002-0655])
15423 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15425 * Remote buffer overflow in SSL3 protocol - an attacker could
15426 supply an oversized session ID to a client. ([CVE-2002-0656])
15428 *Ben Laurie (CHATS)*
15430 * Remote buffer overflow in SSL2 protocol - an attacker could
15431 supply an oversized client master key. ([CVE-2002-0656])
15433 *Ben Laurie (CHATS)*
15435 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15437 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15438 encoded as NULL) with id-dsa-with-sha1.
15440 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15442 * Check various `X509_...()` return values in `apps/req.c`.
15444 *Nils Larsch <nla@trustcenter.de>*
15446 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15447 an end-of-file condition would erroneously be flagged, when the CRLF
15448 was just at the end of a processed block. The bug was discovered when
15449 processing data through a buffering memory BIO handing the data to a
15450 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15451 <ptsekov@syntrex.com> and Nedelcho Stanev.
15455 * Implement a countermeasure against a vulnerability recently found
15456 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15457 before application data chunks to avoid the use of known IVs
15458 with data potentially chosen by the attacker.
15462 * Fix length checks in ssl3_get_client_hello().
15466 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15467 to prevent ssl3_read_internal() from incorrectly assuming that
15468 ssl3_read_bytes() found application data while handshake
15469 processing was enabled when in fact s->s3->in_read_app_data was
15470 merely automatically cleared during the initial handshake.
15472 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15474 * Fix object definitions for Private and Enterprise: they were not
15475 recognized in their shortname (=lowercase) representation. Extend
15476 obj_dat.pl to issue an error when using undefined keywords instead
15477 of silently ignoring the problem (Svenning Sorensen
15478 <sss@sss.dnsalias.net>).
15482 * Fix DH_generate_parameters() so that it works for 'non-standard'
15483 generators, i.e. generators other than 2 and 5. (Previously, the
15484 code did not properly initialise the 'add' and 'rem' values to
15485 BN_generate_prime().)
15487 In the new general case, we do not insist that 'generator' is
15488 actually a primitive root: This requirement is rather pointless;
15489 a generator of the order-q subgroup is just as good, if not
15494 * Map new X509 verification errors to alerts. Discovered and submitted by
15495 Tom Wu <tom@arcot.com>.
15499 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15500 returning non-zero before the data has been completely received
15501 when using non-blocking I/O.
15503 *Bodo Moeller; problem pointed out by John Hughes*
15505 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15507 *Ben Laurie, Lutz Jaenicke*
15509 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15510 Yoram Zahavi <YoramZ@gilian.com>).
15514 * Add information about CygWin 1.3 and on, and preserve proper
15515 configuration for the versions before that.
15517 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15519 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15520 check whether we deal with a copy of a session and do not delete from
15521 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15522 <izhar@checkpoint.com>.
15526 * Do not store session data into the internal session cache, if it
15527 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15528 flag is set). Proposed by Aslam <aslam@funk.com>.
15532 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15537 * [In 0.9.6d-engine release:]
15538 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15540 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15542 * Add the configuration target linux-s390x.
15544 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15546 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15547 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15548 variable as an indication that a ClientHello message has been
15549 received. As the flag value will be lost between multiple
15550 invocations of ssl3_accept when using non-blocking I/O, the
15551 function may not be aware that a handshake has actually taken
15552 place, thus preventing a new session from being added to the
15555 To avoid this problem, we now set s->new_session to 2 instead of
15556 using a local variable.
15558 *Lutz Jaenicke, Bodo Moeller*
15560 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15561 if the SSL_R_LENGTH_MISMATCH error is detected.
15563 *Geoff Thorpe, Bodo Moeller*
15565 * New 'shared_ldflag' column in Configure platform table.
15569 * Fix EVP_CIPHER_mode macro.
15571 *"Dan S. Camper" <dan@bti.net>*
15573 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15574 type, we must throw them away by setting rr->length to 0.
15576 *D P Chang <dpc@qualys.com>*
15578 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15580 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15581 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15582 worked incorrectly for those cases where range = `10..._2` and
15583 `3*range` is two bits longer than range.)
15587 * Only add signing time to PKCS7 structures if it is not already
15592 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15593 OBJ_ld_ce should be OBJ_id_ce.
15594 Also some ip-pda OIDs in crypto/objects/objects.txt were
15595 incorrect (cf. RFC 3039).
15597 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15599 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15600 returns early because it has nothing to do.
15602 *Andy Schneider <andy.schneider@bjss.co.uk>*
15604 * [In 0.9.6c-engine release:]
15605 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15607 *Andy Schneider <andy.schneider@bjss.co.uk>*
15609 * [In 0.9.6c-engine release:]
15610 Add support for Cryptographic Appliance's keyserver technology.
15611 (Use engine 'keyclient')
15613 *Cryptographic Appliances and Geoff Thorpe*
15615 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15616 is called via tools/c89.sh because arguments have to be
15617 rearranged (all '-L' options must appear before the first object
15620 *Richard Shapiro <rshapiro@abinitio.com>*
15622 * [In 0.9.6c-engine release:]
15623 Add support for Broadcom crypto accelerator cards, backported
15626 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15628 * [In 0.9.6c-engine release:]
15629 Add support for SureWare crypto accelerator cards from
15630 Baltimore Technologies. (Use engine 'sureware')
15632 *Baltimore Technologies and Mark Cox*
15634 * [In 0.9.6c-engine release:]
15635 Add support for crypto accelerator cards from Accelerated
15636 Encryption Processing, www.aep.ie. (Use engine 'aep')
15638 *AEP Inc. and Mark Cox*
15640 * Add a configuration entry for gcc on UnixWare.
15642 *Gary Benson <gbenson@redhat.com>*
15644 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15645 messages are stored in a single piece (fixed-length part and
15646 variable-length part combined) and fix various bugs found on the way.
15650 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15651 instead. BIO_gethostbyname() does not know what timeouts are
15652 appropriate, so entries would stay in cache even when they have
15654 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15656 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15657 faced with a pathologically small ClientHello fragment that does
15658 not contain client_version: Instead of aborting with an error,
15659 simply choose the highest available protocol version (i.e.,
15660 TLS 1.0 unless it is disabled). In practice, ClientHello
15661 messages are never sent like this, but this change gives us
15662 strictly correct behaviour at least for TLS.
15666 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15667 never resets s->method to s->ctx->method when called from within
15668 one of the SSL handshake functions.
15670 *Bodo Moeller; problem pointed out by Niko Baric*
15672 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15673 (sent using the client's version number) if client_version is
15674 smaller than the protocol version in use. Also change
15675 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15676 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15677 the client will at least see that alert.
15681 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15686 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15687 client receives HelloRequest while in a handshake.
15689 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15691 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15692 should end in 'break', not 'goto end' which circumvents various
15693 cleanups done in state SSL_ST_OK. But session related stuff
15694 must be disabled for SSL_ST_OK in the case that we just sent a
15697 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15698 before just sending a HelloRequest.
15700 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15702 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15703 reveal whether illegal block cipher padding was found or a MAC
15704 verification error occurred. (Neither SSLerr() codes nor alerts
15705 are directly visible to potential attackers, but the information
15706 may leak via logfiles.)
15708 Similar changes are not required for the SSL 2.0 implementation
15709 because the number of padding bytes is sent in clear for SSL 2.0,
15710 and the extra bytes are just ignored. However ssl/s2_pkt.c
15711 failed to verify that the purported number of padding bytes is in
15716 * Add OpenUNIX-8 support including shared libraries
15717 (Boyd Lynn Gerber <gerberb@zenez.com>).
15721 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15722 'wristwatch attack' using huge encoding parameters (cf.
15723 James H. Manger's CRYPTO 2001 paper). Note that the
15724 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15725 encoding parameters and hence was not vulnerable.
15729 * BN_sqr() bug fix.
15731 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15733 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15734 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15735 followed by modular reduction.
15737 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15739 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15740 equivalent based on BN_pseudo_rand() instead of BN_rand().
15744 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15745 This function was broken, as the check for a new client hello message
15746 to handle SGC did not allow these large messages.
15747 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15751 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15755 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15756 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15760 * Rework the configuration and shared library support for Tru64 Unix.
15761 The configuration part makes use of modern compiler features and
15762 still retains old compiler behavior for those that run older versions
15763 of the OS. The shared library support part includes a variant that
15764 uses the RPATH feature, and is available through the special
15765 configuration target "alpha-cc-rpath", which will never be selected
15768 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15770 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15771 with the same message size as in ssl3_get_certificate_request().
15772 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15773 messages might inadvertently be reject as too long.
15775 *Petr Lampa <lampa@fee.vutbr.cz>*
15777 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15781 * Modified SSL library such that the verify_callback that has been set
15782 specifically for an SSL object with SSL_set_verify() is actually being
15783 used. Before the change, a verify_callback set with this function was
15784 ignored and the verify_callback() set in the SSL_CTX at the time of
15785 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15786 to allow the necessary settings.
15790 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15791 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15792 done automatically (in contradiction to the requirements of the C
15793 standard). This made problems when used from OpenSSH.
15797 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15798 dh->length and always used
15800 BN_rand_range(priv_key, dh->p).
15802 BN_rand_range() is not necessary for Diffie-Hellman, and this
15803 specific range makes Diffie-Hellman unnecessarily inefficient if
15804 dh->length (recommended exponent length) is much smaller than the
15805 length of dh->p. We could use BN_rand_range() if the order of
15806 the subgroup was stored in the DH structure, but we only have
15811 BN_rand(priv_key, l, ...)
15813 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15820 RSA_eay_public_encrypt
15821 RSA_eay_private_decrypt
15822 RSA_eay_private_encrypt (signing)
15823 RSA_eay_public_decrypt (signature verification)
15825 (default implementations for RSA_public_encrypt,
15826 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15827 always reject numbers >= n.
15831 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15832 to synchronize access to 'locking_thread'. This is necessary on
15833 systems where access to 'locking_thread' (an 'unsigned long'
15834 variable) is not atomic.
15838 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15839 *before* setting the 'crypto_lock_rand' flag. The previous code had
15840 a race condition if 0 is a valid thread ID.
15842 *Travis Vitek <vitek@roguewave.com>*
15844 * Add support for shared libraries under Irix.
15846 *Albert Chin-A-Young <china@thewrittenword.com>*
15848 * Add configuration option to build on Linux on both big-endian and
15849 little-endian MIPS.
15851 *Ralf Baechle <ralf@uni-koblenz.de>*
15853 * Add the possibility to create shared libraries on HP-UX.
15857 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15859 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15860 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15861 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15862 PRNG state recovery was possible based on the output of
15863 one PRNG request appropriately sized to gain knowledge on
15864 'md' followed by enough consecutive 1-byte PRNG requests
15865 to traverse all of 'state'.
15867 1. When updating 'md_local' (the current thread's copy of 'md')
15868 during PRNG output generation, hash all of the previous
15869 'md_local' value, not just the half used for PRNG output.
15871 2. Make the number of bytes from 'state' included into the hash
15872 independent from the number of PRNG bytes requested.
15874 The first measure alone would be sufficient to avoid
15875 Markku-Juhani's attack. (Actually it had never occurred
15876 to me that the half of 'md_local' used for chaining was the
15877 half from which PRNG output bytes were taken -- I had always
15878 assumed that the secret half would be used.) The second
15879 measure makes sure that additional data from 'state' is never
15880 mixed into 'md_local' in small portions; this heuristically
15881 further strengthens the PRNG.
15885 * Fix crypto/bn/asm/mips3.s.
15889 * When only the key is given to "enc", the IV is undefined. Print out
15890 an error message in this case.
15894 * Handle special case when X509_NAME is empty in X509 printing routines.
15898 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15899 positive and less than q.
15903 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15904 used: it isn't thread safe and the add_lock_callback should handle
15907 *Paul Rose <Paul.Rose@bridge.com>*
15909 * Verify that incoming data obeys the block size in
15910 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15916 *Ulf Möller, Bodo Möller*
15918 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15919 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15920 when fixing the server behaviour for backwards-compatible 'client
15921 hello' messages. (Note that the attack is impractical against
15922 SSL 3.0 and TLS 1.0 anyway because length and version checking
15923 means that the probability of guessing a valid ciphertext is
15924 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15927 Before 0.9.5, the countermeasure (hide the error by generating a
15928 random 'decryption result') did not work properly because
15929 ERR_clear_error() was missing, meaning that SSL_get_error() would
15930 detect the supposedly ignored error.
15932 Both problems are now fixed.
15936 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15937 (previously it was 1024).
15941 * Fix for compatibility mode trust settings: ignore trust settings
15942 unless some valid trust or reject settings are present.
15946 * Fix for blowfish EVP: its a variable length cipher.
15950 * Fix various bugs related to DSA S/MIME verification. Handle missing
15951 parameters in DSA public key structures and return an error in the
15952 DSA routines if parameters are absent.
15956 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15957 in the current directory if neither $RANDFILE nor $HOME was set.
15958 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15959 caused some confusion to Windows users who haven't defined $HOME.
15960 Thus RAND_file_name() is changed again: e_os.h can define a
15961 DEFAULT_HOME, which will be used if $HOME is not set.
15962 For Windows, we use "C:"; on other platforms, we still require
15963 environment variables.
15965 * Move 'if (!initialized) RAND_poll()' into regions protected by
15966 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15967 having multiple threads call RAND_poll() concurrently.
15971 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15972 combination of a flag and a thread ID variable.
15973 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15974 flag), *other* threads can enter ssleay_add_bytes without obeying
15975 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15976 that they do not hold after the first thread unsets add_do_not_lock).
15980 * Change bctest again: '-x' expressions are not available in all
15981 versions of 'test'.
15985 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15987 * Fix a couple of memory leaks in PKCS7_dataDecode()
15989 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15991 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15992 the default extension for executables, if any. Also, make the perl
15993 scripts that use symlink() to test if it really exists and use "cp"
15994 if it doesn't. All this made OpenSSL compilable and installable in
15999 * Fix for asn1_GetSequence() for indefinite length constructed data.
16000 If SEQUENCE is length is indefinite just set c->slen to the total
16001 amount of data available.
16003 *Steve Henson, reported by shige@FreeBSD.org*
16005 *This change does not apply to 0.9.7.*
16007 * Change bctest to avoid here-documents inside command substitution
16008 (workaround for FreeBSD /bin/sh bug).
16009 For compatibility with Ultrix, avoid shell functions (introduced
16010 in the bctest version that searches along $PATH).
16014 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16015 with des_encrypt() defined on some operating systems, like Solaris
16020 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16021 On the Importance of Eliminating Errors in Cryptographic
16022 Computations, J. Cryptology 14 (2001) 2, 101-119,
16023 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16027 * MIPS assembler BIGNUM division bug fix.
16031 * Disabled incorrect Alpha assembler code.
16035 * Fix PKCS#7 decode routines so they correctly update the length
16036 after reading an EOC for the EXPLICIT tag.
16040 *This change does not apply to 0.9.7.*
16042 * Fix bug in PKCS#12 key generation routines. This was triggered
16043 if a 3DES key was generated with a 0 initial byte. Include
16044 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16045 (but broken) behaviour.
16049 * Enhance bctest to search for a working bc along $PATH and print
16052 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16054 * Fix memory leaks in err.c: free err_data string if necessary;
16055 don't write to the wrong index in ERR_set_error_data.
16059 * Implement ssl23_peek (analogous to ssl23_read), which previously
16064 * Replace rdtsc with `_emit` statements for VC++ version 5.
16066 *Jeremy Cooper <jeremy@baymoo.org>*
16068 * Make it possible to reuse SSLv2 sessions.
16072 * In copy_email() check for >= 0 as a return value for
16073 X509_NAME_get_index_by_NID() since 0 is a valid index.
16075 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16077 * Avoid coredump with unsupported or invalid public keys by checking if
16078 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16079 PKCS7_verify() fails with non detached data.
16083 * Don't use getenv in library functions when run as setuid/setgid.
16084 New function OPENSSL_issetugid().
16088 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16089 due to incorrect handling of multi-threading:
16091 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16093 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16095 3. Count how many times MemCheck_off() has been called so that
16096 nested use can be treated correctly. This also avoids
16097 inband-signalling in the previous code (which relied on the
16098 assumption that thread ID 0 is impossible).
16102 * Add "-rand" option also to s_client and s_server.
16106 * Fix CPU detection on Irix 6.x.
16107 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16108 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16110 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16115 *This change does not apply to 0.9.7.*
16117 * Use the cached encoding of an X509_NAME structure rather than
16118 copying it. This is apparently the reason for the libsafe "errors"
16119 but the code is actually correct.
16123 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16124 Bleichenbacher's DSA attack.
16125 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16126 to be set and top=0 forces the highest bit to be set; top=-1 is new
16127 and leaves the highest bit random.
16129 *Ulf Moeller, Bodo Moeller*
16131 * In the `NCONF_...`-based implementations for `CONF_...` queries
16132 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16133 a temporary CONF structure with the data component set to NULL
16134 (which gives segmentation faults in lh_retrieve).
16135 Instead, use NULL for the CONF pointer in CONF_get_string and
16136 CONF_get_number (which may use environment variables) and directly
16137 return NULL from CONF_get_section.
16141 * Fix potential buffer overrun for EBCDIC.
16145 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16146 keyUsage if basicConstraints absent for a CA.
16150 * Make SMIME_write_PKCS7() write mail header values with a format that
16151 is more generally accepted (no spaces before the semicolon), since
16152 some programs can't parse those values properly otherwise. Also make
16153 sure BIO's that break lines after each write do not create invalid
16158 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16159 macros previously used would not encode an empty SEQUENCE OF
16160 and break the signature.
16164 *This change does not apply to 0.9.7.*
16166 * Zero the premaster secret after deriving the master secret in
16171 * Add some EVP_add_digest_alias registrations (as found in
16172 OpenSSL_add_all_digests()) to SSL_library_init()
16173 aka OpenSSL_add_ssl_algorithms(). This provides improved
16174 compatibility with peers using X.509 certificates
16175 with unconventional AlgorithmIdentifier OIDs.
16179 * Fix for Irix with NO_ASM.
16181 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16183 * ./config script fixes.
16185 *Ulf Moeller, Richard Levitte*
16187 * Fix 'openssl passwd -1'.
16191 * Change PKCS12_key_gen_asc() so it can cope with non null
16192 terminated strings whose length is passed in the passlen
16193 parameter, for example from PEM callbacks. This was done
16194 by adding an extra length parameter to asc2uni().
16196 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16198 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16199 call failed, free the DSA structure.
16203 * Fix to uni2asc() to cope with zero length Unicode strings.
16204 These are present in some PKCS#12 files.
16208 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16209 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16210 when writing a 32767 byte record.
16212 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16214 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16215 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16217 (RSA objects have a reference count access to which is protected
16218 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16219 so they are meant to be shared between threads.)
16220 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16221 "Reddie, Steven" <Steven.Reddie@ca.com>*
16223 * Fix a deadlock in CRYPTO_mem_leaks().
16227 * Use better test patterns in bntest.
16231 * rand_win.c fix for Borland C.
16235 * BN_rshift bugfix for n == 0.
16239 * Add a 'bctest' script that checks for some known 'bc' bugs
16240 so that 'make test' does not abort just because 'bc' is broken.
16244 * Store verify_result within SSL_SESSION also for client side to
16245 avoid potential security hole. (Re-used sessions on the client side
16246 always resulted in verify_result==X509_V_OK, not using the original
16247 result of the server certificate verification.)
16251 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16252 SSL3_RT_APPLICATION_DATA, return 0.
16253 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16258 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16259 releases, have been re-implemented by renaming the previous
16260 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16261 and ssl3_read_internal, respectively, and adding 'peek' parameters
16262 to them. The new ssl[23]_{read,peek} functions are calls to
16263 ssl[23]_read_internal with the 'peek' flag set appropriately.
16264 A 'peek' parameter has also been added to ssl3_read_bytes, which
16265 does the actual work for ssl3_read_internal.
16269 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16270 the method-specific "init()" handler. Also clean up ex_data after
16271 calling the method-specific "finish()" handler. Previously, this was
16272 happening the other way round.
16276 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16277 The previous value, 12, was not always sufficient for BN_mod_exp().
16281 * Make sure that shared libraries get the internal name engine with
16282 the full version number and not just 0. This should mark the
16283 shared libraries as not backward compatible. Of course, this should
16284 be changed again when we can guarantee backward binary compatibility.
16288 * Fix typo in get_cert_by_subject() in by_dir.c
16290 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16292 * Rework the system to generate shared libraries:
16294 - Make note of the expected extension for the shared libraries and
16295 if there is a need for symbolic links from for example libcrypto.so.0
16296 to libcrypto.so.0.9.7. There is extended info in Configure for
16299 - Make as few rebuilds of the shared libraries as possible.
16301 - Still avoid linking the OpenSSL programs with the shared libraries.
16303 - When installing, install the shared libraries separately from the
16308 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16310 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16311 and not in SSL_clear because the latter is also used by the
16312 accept/connect functions; previously, the settings made by
16313 SSL_set_read_ahead would be lost during the handshake.
16315 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16317 * Correct util/mkdef.pl to be selective about disabled algorithms.
16318 Previously, it would create entries for disabled algorithms no
16323 * Added several new manual pages for SSL_* function.
16327 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16329 * In ssl23_get_client_hello, generate an error message when faced
16330 with an initial SSL 3.0/TLS record that is too small to contain the
16331 first two bytes of the ClientHello message, i.e. client_version.
16332 (Note that this is a pathologic case that probably has never happened
16333 in real life.) The previous approach was to use the version number
16334 from the record header as a substitute; but our protocol choice
16335 should not depend on that one because it is not authenticated
16336 by the Finished messages.
16340 * More robust randomness gathering functions for Windows.
16342 *Jeffrey Altman <jaltman@columbia.edu>*
16344 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16345 not set then we don't setup the error code for issuer check errors
16346 to avoid possibly overwriting other errors which the callback does
16347 handle. If an application does set the flag then we assume it knows
16348 what it is doing and can handle the new informational codes
16353 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16354 a general "ANY" type, as such it should be able to decode anything
16355 including tagged types. However it didn't check the class so it would
16356 wrongly interpret tagged types in the same way as their universal
16357 counterpart and unknown types were just rejected. Changed so that the
16358 tagged and unknown types are handled in the same way as a SEQUENCE:
16359 that is the encoding is stored intact. There is also a new type
16360 "V_ASN1_OTHER" which is used when the class is not universal, in this
16361 case we have no idea what the actual type is so we just lump them all
16366 * On VMS, stdout may very well lead to a file that is written to
16367 in a record-oriented fashion. That means that every write() will
16368 write a separate record, which will be read separately by the
16369 programs trying to read from it. This can be very confusing.
16371 The solution is to put a BIO filter in the way that will buffer
16372 text until a linefeed is reached, and then write everything a
16373 line at a time, so every record written will be an actual line,
16374 not chunks of lines and not (usually doesn't happen, but I've
16375 seen it once) several lines in one record. BIO_f_linebuffer() is
16378 Currently, it's a VMS-only method, because that's where it has
16379 been tested well enough.
16383 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16384 it can return incorrect results.
16385 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16386 but it was in 0.9.6-beta[12].)
16390 * Disable the check for content being present when verifying detached
16391 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16392 include zero length content when signing messages.
16396 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16397 BIO_ctrl (for BIO pairs).
16401 * Add DSO method for VMS.
16405 * Bug fix: Montgomery multiplication could produce results with the
16410 * Add RPM specification openssl.spec and modify it to build three
16411 packages. The default package contains applications, application
16412 documentation and run-time libraries. The devel package contains
16413 include files, static libraries and function documentation. The
16414 doc package contains the contents of the doc directory. The original
16415 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16419 * Add a large number of documentation files for many SSL routines.
16421 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16423 * Add a configuration entry for Sony News 4.
16425 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16427 * Don't set the two most significant bits to one when generating a
16428 random number < q in the DSA library.
16432 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16433 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16434 the underlying transport is blocking) if a handshake took place.
16435 (The default behaviour is needed by applications such as s_client
16436 and s_server that use select() to determine when to use SSL_read;
16437 but for applications that know in advance when to expect data, it
16438 just makes things more complicated.)
16442 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16447 * Add a few more EBCDIC conditionals that make `req` and `x509`
16448 work better on such systems.
16450 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16452 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16453 Update PKCS12_parse() so it copies the friendlyName and the
16454 keyid to the certificates aux info.
16458 * Fix bug in PKCS7_verify() which caused an infinite loop
16459 if there was more than one signature.
16461 *Sven Uszpelkat <su@celocom.de>*
16463 * Major change in util/mkdef.pl to include extra information
16464 about each symbol, as well as presenting variables as well
16465 as functions. This change means that there's n more need
16466 to rebuild the .num files when some algorithms are excluded.
16470 * Allow the verify time to be set by an application,
16471 rather than always using the current time.
16475 * Phase 2 verify code reorganisation. The certificate
16476 verify code now looks up an issuer certificate by a
16477 number of criteria: subject name, authority key id
16478 and key usage. It also verifies self signed certificates
16479 by the same criteria. The main comparison function is
16480 X509_check_issued() which performs these checks.
16482 Lot of changes were necessary in order to support this
16483 without completely rewriting the lookup code.
16485 Authority and subject key identifier are now cached.
16487 The LHASH 'certs' is X509_STORE has now been replaced
16488 by a STACK_OF(X509_OBJECT). This is mainly because an
16489 LHASH can't store or retrieve multiple objects with
16490 the same hash value.
16492 As a result various functions (which were all internal
16493 use only) have changed to handle the new X509_STORE
16494 structure. This will break anything that messed round
16495 with X509_STORE internally.
16497 The functions X509_STORE_add_cert() now checks for an
16498 exact match, rather than just subject name.
16500 The X509_STORE API doesn't directly support the retrieval
16501 of multiple certificates matching a given criteria, however
16502 this can be worked round by performing a lookup first
16503 (which will fill the cache with candidate certificates)
16504 and then examining the cache for matches. This is probably
16505 the best we can do without throwing out X509_LOOKUP
16506 entirely (maybe later...).
16508 The X509_VERIFY_CTX structure has been enhanced considerably.
16510 All certificate lookup operations now go via a get_issuer()
16511 callback. Although this currently uses an X509_STORE it
16512 can be replaced by custom lookups. This is a simple way
16513 to bypass the X509_STORE hackery necessary to make this
16514 work and makes it possible to use more efficient techniques
16515 in future. A very simple version which uses a simple
16516 STACK for its trusted certificate store is also provided
16517 using X509_STORE_CTX_trusted_stack().
16519 The verify_cb() and verify() callbacks now have equivalents
16520 in the X509_STORE_CTX structure.
16522 X509_STORE_CTX also has a 'flags' field which can be used
16523 to customise the verify behaviour.
16527 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16528 excludes S/MIME capabilities.
16532 * When a certificate request is read in keep a copy of the
16533 original encoding of the signed data and use it when outputting
16534 again. Signatures then use the original encoding rather than
16535 a decoded, encoded version which may cause problems if the
16536 request is improperly encoded.
16540 * For consistency with other BIO_puts implementations, call
16541 buffer_write(b, ...) directly in buffer_puts instead of calling
16544 In BIO_puts, increment b->num_write as in BIO_write.
16546 *Peter.Sylvester@EdelWeb.fr*
16548 * Fix BN_mul_word for the case where the word is 0. (We have to use
16549 BN_zero, we may not return a BIGNUM with an array consisting of
16550 words set to zero.)
16554 * Avoid calling abort() from within the library when problems are
16555 detected, except if preprocessor symbols have been defined
16556 (such as REF_CHECK, BN_DEBUG etc.).
16560 * New openssl application 'rsautl'. This utility can be
16561 used for low-level RSA operations. DER public key
16562 BIO/fp routines also added.
16566 * New Configure entry and patches for compiling on QNX 4.
16568 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16570 * A demo state-machine implementation was sponsored by
16571 Nuron (<http://www.nuron.com/>) and is now available in
16572 demos/state_machine.
16576 * New options added to the 'dgst' utility for signature
16577 generation and verification.
16581 * Unrecognized PKCS#7 content types are now handled via a
16582 catch all ASN1_TYPE structure. This allows unsupported
16583 types to be stored as a "blob" and an application can
16584 encode and decode it manually.
16588 * Fix various signed/unsigned issues to make a_strex.c
16589 compile under VC++.
16591 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16593 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16594 length if passed a buffer. ASN1_INTEGER_to_BN failed
16595 if passed a NULL BN and its argument was negative.
16597 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16599 * Modification to PKCS#7 encoding routines to output definite
16600 length encoding. Since currently the whole structures are in
16601 memory there's not real point in using indefinite length
16602 constructed encoding. However if OpenSSL is compiled with
16603 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16607 * Added BIO_vprintf() and BIO_vsnprintf().
16611 * Added more prefixes to parse for in the strings written
16612 through a logging bio, to cover all the levels that are available
16613 through syslog. The prefixes are now:
16615 PANIC, EMERG, EMR => LOG_EMERG
16616 ALERT, ALR => LOG_ALERT
16617 CRIT, CRI => LOG_CRIT
16618 ERROR, ERR => LOG_ERR
16619 WARNING, WARN, WAR => LOG_WARNING
16620 NOTICE, NOTE, NOT => LOG_NOTICE
16621 INFO, INF => LOG_INFO
16622 DEBUG, DBG => LOG_DEBUG
16624 and as before, if none of those prefixes are present at the
16625 beginning of the string, LOG_ERR is chosen.
16627 On Win32, the `LOG_*` levels are mapped according to this:
16629 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16630 LOG_WARNING => EVENTLOG_WARNING_TYPE
16631 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16635 * Made it possible to reconfigure with just the configuration
16636 argument "reconf" or "reconfigure". The command line arguments
16637 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16638 and are retrieved from there when reconfiguring.
16644 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16646 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16650 * The obj_dat.pl script was messing up the sorting of object
16651 names. The reason was that it compared the quoted version
16652 of strings as a result "OCSP" > "OCSP Signing" because
16653 " > SPACE. Changed script to store unquoted versions of
16654 names and add quotes on output. It was also omitting some
16655 names from the lookup table if they were given a default
16656 value (that is if SN is missing it is given the same
16657 value as LN and vice versa), these are now added on the
16658 grounds that if an object has a name we should be able to
16659 look it up. Finally added warning output when duplicate
16660 short or long names are found.
16664 * Changes needed for Tandem NSK.
16666 *Scott Uroff <scott@xypro.com>*
16668 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16669 RSA_padding_check_SSLv23(), special padding was never detected
16670 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16671 version rollback attacks was not effective.
16673 In s23_clnt.c, don't use special rollback-attack detection padding
16674 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16675 client; similarly, in s23_srvr.c, don't do the rollback check if
16676 SSL 2.0 is the only protocol enabled in the server.
16680 * Make it possible to get hexdumps of unprintable data with 'openssl
16681 asn1parse'. By implication, the functions ASN1_parse_dump() and
16682 BIO_dump_indent() are added.
16686 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16687 these print out strings and name structures based on various
16688 flags including RFC2253 support and proper handling of
16689 multibyte characters. Added options to the 'x509' utility
16690 to allow the various flags to be set.
16694 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16695 Also change the functions X509_cmp_current_time() and
16696 X509_gmtime_adj() work with an ASN1_TIME structure,
16697 this will enable certificates using GeneralizedTime in validity
16698 dates to be checked.
16702 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16703 negative public key encodings) on by default,
16704 NO_NEG_PUBKEY_BUG can be set to disable it.
16708 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16709 content octets. An i2c_ASN1_OBJECT is unnecessary because
16710 the encoding can be trivially obtained from the structure.
16714 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16715 not read locks (`CRYPTO_r_[un]lock`).
16719 * A first attempt at creating official support for shared
16720 libraries through configuration. I've kept it so the
16721 default is static libraries only, and the OpenSSL programs
16722 are always statically linked for now, but there are
16723 preparations for dynamic linking in place.
16724 This has been tested on Linux and Tru64.
16728 * Randomness polling function for Win9x, as described in:
16729 Peter Gutmann, Software Generation of Practically Strong
16734 * Fix so PRNG is seeded in req if using an already existing
16739 * New options to smime application. -inform and -outform
16740 allow alternative formats for the S/MIME message including
16741 PEM and DER. The -content option allows the content to be
16742 specified separately. This should allow things like Netscape
16743 form signing output easier to verify.
16747 * Fix the ASN1 encoding of tags using the 'long form'.
16751 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16752 STRING types. These convert content octets to and from the
16753 underlying type. The actual tag and length octets are
16754 already assumed to have been read in and checked. These
16755 are needed because all other string types have virtually
16756 identical handling apart from the tag. By having versions
16757 of the ASN1 functions that just operate on content octets
16758 IMPLICIT tagging can be handled properly. It also allows
16759 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16760 and ASN1_INTEGER are identical apart from the tag.
16764 * Change the handling of OID objects as follows:
16766 - New object identifiers are inserted in objects.txt, following
16767 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16768 - objects.pl is used to process obj_mac.num and create a new
16770 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16773 This is currently kind of a hack, and the perl code in objects.pl
16774 isn't very elegant, but it works as I intended. The simplest way
16775 to check that it worked correctly is to look in obj_dat.h and
16776 check the array nid_objs and make sure the objects haven't moved
16777 around (this is important!). Additions are OK, as well as
16778 consistent name changes.
16782 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16786 * Addition of the command line parameter '-rand file' to 'openssl req'.
16787 The given file adds to whatever has already been seeded into the
16788 random pool through the RANDFILE configuration file option or
16789 environment variable, or the default random state file.
16793 * mkstack.pl now sorts each macro group into lexical order.
16794 Previously the output order depended on the order the files
16795 appeared in the directory, resulting in needless rewriting
16800 * Patches to make OpenSSL compile under Win32 again. Mostly
16801 work arounds for the VC++ problem that it treats func() as
16802 func(void). Also stripped out the parts of mkdef.pl that
16803 added extra typesafe functions: these no longer exist.
16807 * Reorganisation of the stack code. The macros are now all
16808 collected in safestack.h . Each macro is defined in terms of
16809 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16810 DEBUG_SAFESTACK is now handled in terms of function casts,
16811 this has the advantage of retaining type safety without the
16812 use of additional functions. If DEBUG_SAFESTACK is not defined
16813 then the non typesafe macros are used instead. Also modified the
16814 mkstack.pl script to handle the new form. Needs testing to see
16815 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16816 the default if no major problems. Similar behaviour for ASN1_SET_OF
16817 and PKCS12_STACK_OF.
16821 * When some versions of IIS use the 'NET' form of private key the
16822 key derivation algorithm is different. Normally MD5(password) is
16823 used as a 128 bit RC4 key. In the modified case
16824 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16825 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16826 as the old Netscape_RSA functions except they have an additional
16827 'sgckey' parameter which uses the modified algorithm. Also added
16828 an -sgckey command line option to the rsa utility. Thanks to
16829 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16830 algorithm to openssl-dev.
16834 * The evp_local.h macros were using 'c.##kname' which resulted in
16835 invalid expansion on some systems (SCO 5.0.5 for example).
16836 Corrected to 'c.kname'.
16838 *Phillip Porch <root@theporch.com>*
16840 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16841 a STACK of email addresses from a certificate or request, these look
16842 in the subject name and the subject alternative name extensions and
16843 omit any duplicate addresses.
16847 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16848 This makes DSA verification about 2 % faster.
16852 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16853 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16854 plus overhead for 1024 bit moduli).
16855 This makes exponentiations about 0.5 % faster for 1024 bit
16856 exponents (as measured by "openssl speed rsa2048").
16860 * Rename memory handling macros to avoid conflicts with other
16862 Malloc => OPENSSL_malloc
16863 Malloc_locked => OPENSSL_malloc_locked
16864 Realloc => OPENSSL_realloc
16865 Free => OPENSSL_free
16869 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16870 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16874 * CygWin32 support.
16876 *John Jarvie <jjarvie@newsguy.com>*
16878 * The type-safe stack code has been rejigged. It is now only compiled
16879 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16880 by default all type-specific stack functions are "#define"d back to
16881 standard stack functions. This results in more streamlined output
16882 but retains the type-safety checking possibilities of the original
16887 * The STACK code has been cleaned up, and certain type declarations
16888 that didn't make a lot of sense have been brought in line. This has
16889 also involved a cleanup of sorts in safestack.h to more correctly
16890 map type-safe stack functions onto their plain stack counterparts.
16891 This work has also resulted in a variety of "const"ifications of
16892 lots of the code, especially `_cmp` operations which should normally
16893 be prototyped with "const" parameters anyway.
16897 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16898 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16899 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16900 where all of 'md' is used each time the PRNG is used, but 'state'
16901 is used only indexed by a cyclic counter. As entropy may not be
16902 well distributed from the beginning, 'md' is important as a
16903 chaining variable. However, the output function chains only half
16904 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16905 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16906 in all of 'state' being rewritten, with the new values depending
16907 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16911 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16912 the handshake is continued after ssl_verify_cert_chain();
16913 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16914 can lead to 'unexplainable' connection aborts later.
16916 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16918 * Major EVP API cipher revision.
16919 Add hooks for extra EVP features. This allows various cipher
16920 parameters to be set in the EVP interface. Support added for variable
16921 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16922 setting of RC2 and RC5 parameters.
16924 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16927 Remove lots of duplicated code from the EVP library. For example *every*
16928 cipher init() function handles the 'iv' in the same way according to the
16929 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16930 for CFB and OFB modes they zero ctx->num.
16932 New functionality allows removal of S/MIME code RC2 hack.
16934 Most of the routines have the same form and so can be declared in terms
16937 By shifting this to the top level EVP_CipherInit() it can be removed from
16938 all individual ciphers. If the cipher wants to handle IVs or keys
16939 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16942 Change lots of functions like EVP_EncryptUpdate() to now return a
16943 value: although software versions of the algorithms cannot fail
16944 any installed hardware versions can.
16948 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16949 this option is set, tolerate broken clients that send the negotiated
16950 protocol version number instead of the requested protocol version
16955 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16956 i.e. non-zero for export ciphersuites, zero otherwise.
16957 Previous versions had this flag inverted, inconsistent with
16958 rsa_tmp_cb (..._TMP_RSA_CB).
16960 *Bodo Moeller; problem reported by Amit Chopra*
16962 * Add missing DSA library text string. Work around for some IIS
16963 key files with invalid SEQUENCE encoding.
16967 * Add a document (doc/standards.txt) that list all kinds of standards
16968 and so on that are implemented in OpenSSL.
16972 * Enhance c_rehash script. Old version would mishandle certificates
16973 with the same subject name hash and wouldn't handle CRLs at all.
16974 Added -fingerprint option to crl utility, to support new c_rehash
16979 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16983 * Fix for SSL server purpose checking. Server checking was
16984 rejecting certificates which had extended key usage present
16985 but no ssl client purpose.
16987 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16989 * Make PKCS#12 code work with no password. The PKCS#12 spec
16990 is a little unclear about how a blank password is handled.
16991 Since the password in encoded as a BMPString with terminating
16992 double NULL a zero length password would end up as just the
16993 double NULL. However no password at all is different and is
16994 handled differently in the PKCS#12 key generation code. NS
16995 treats a blank password as zero length. MSIE treats it as no
16996 password on export: but it will try both on import. We now do
16997 the same: PKCS12_parse() tries zero length and no password if
16998 the password is set to "" or NULL (NULL is now a valid password:
16999 it wasn't before) as does the pkcs12 application.
17003 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17004 perror when PEM_read_bio_X509_REQ fails, the error message must
17005 be obtained from the error queue.
17009 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17010 it in ERR_remove_state if appropriate, and change ERR_get_state
17011 accordingly to avoid race conditions (this is necessary because
17012 thread_hash is no longer constant once set).
17016 * Bugfix for linux-elf makefile.one.
17020 * RSA_get_default_method() will now cause a default
17021 RSA_METHOD to be chosen if one doesn't exist already.
17022 Previously this was only set during a call to RSA_new()
17023 or RSA_new_method(NULL) meaning it was possible for
17024 RSA_get_default_method() to return NULL.
17028 * Added native name translation to the existing DSO code
17029 that will convert (if the flag to do so is set) filenames
17030 that are sufficiently small and have no path information
17031 into a canonical native form. Eg. "blah" converted to
17032 "libblah.so" or "blah.dll" etc.
17036 * New function ERR_error_string_n(e, buf, len) which is like
17037 ERR_error_string(e, buf), but writes at most 'len' bytes
17038 including the 0 terminator. For ERR_error_string_n, 'buf'
17041 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17043 * CONF library reworked to become more general. A new CONF
17044 configuration file reader "class" is implemented as well as a
17045 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17046 old `CONF_*` functions are still there, but are reimplemented to
17047 work in terms of the new functions. Also, a set of functions
17048 to handle the internal storage of the configuration data is
17049 provided to make it easier to write new configuration file
17050 reader "classes" (I can definitely see something reading a
17051 configuration file in XML format, for example), called `_CONF_*`,
17052 or "the configuration storage API"...
17054 The new configuration file reading functions are:
17056 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17057 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17059 NCONF_default, NCONF_WIN32
17061 NCONF_dump_fp, NCONF_dump_bio
17063 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17064 NCONF_new creates a new CONF object. This works in the same way
17065 as other interfaces in OpenSSL, like the BIO interface.
17066 `NCONF_dump_*` dump the internal storage of the configuration file,
17067 which is useful for debugging. All other functions take the same
17068 arguments as the old `CONF_*` functions with the exception of the
17069 first that must be a `CONF *` instead of a `LHASH *`.
17071 To make it easier to use the new classes with the old `CONF_*` functions,
17072 the function CONF_set_default_method is provided.
17076 * Add '-tls1' option to 'openssl ciphers', which was already
17077 mentioned in the documentation but had not been implemented.
17078 (This option is not yet really useful because even the additional
17079 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17083 * Initial DSO code added into libcrypto for letting OpenSSL (and
17084 OpenSSL-based applications) load shared libraries and bind to
17085 them in a portable way.
17087 *Geoff Thorpe, with contributions from Richard Levitte*
17089 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17091 * Make sure _lrotl and _lrotr are only used with MSVC.
17093 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17094 (the default implementation of RAND_status).
17096 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17097 to '-clrext' (= clear extensions), as intended and documented.
17098 *Bodo Moeller; inconsistency pointed out by Michael Attili
17099 <attili@amaxo.com>*
17101 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17102 was larger than the MD block size.
17104 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17106 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17107 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17108 using the passed key: if the passed key was a private key the result
17109 of X509_print(), for example, would be to print out all the private key
17114 * des_quad_cksum() byte order bug fix.
17115 *Ulf Möller, using the problem description in krb4-0.9.7, where
17116 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17118 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17121 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17123 * For easily testing in shell scripts whether some command
17124 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17125 returns with exit code 0 iff no command of the given name is available.
17126 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17127 the output goes to stdout and nothing is printed to stderr.
17128 Additional arguments are always ignored.
17130 Since for each cipher there is a command of the same name,
17131 the 'no-cipher' compilation switches can be tested this way.
17133 ('openssl no-XXX' is not able to detect pseudo-commands such
17134 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17138 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17142 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17143 is set; it will be thrown away anyway because each handshake creates
17145 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17146 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17147 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17148 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17152 * New s_client option -ign_eof: EOF at stdin is ignored, and
17153 'Q' and 'R' lose their special meanings (quit/renegotiate).
17154 This is part of what -quiet does; unlike -quiet, -ign_eof
17155 does not suppress any output.
17159 * Add compatibility options to the purpose and trust code. The
17160 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17161 accepts a certificate or CA, this was the previous behaviour,
17162 with all the associated security issues.
17164 X509_TRUST_COMPAT is the old trust behaviour: only and
17165 automatically trust self signed roots in certificate store. A
17166 new trust setting X509_TRUST_DEFAULT is used to specify that
17167 a purpose has no associated trust setting and it should instead
17168 use the value in the default purpose.
17172 * Fix the PKCS#8 DSA private key code so it decodes keys again
17173 and fix a memory leak.
17177 * In util/mkerr.pl (which implements 'make errors'), preserve
17178 reason strings from the previous version of the .c file, as
17179 the default to have only downcase letters (and digits) in
17180 automatically generated reasons codes is not always appropriate.
17184 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17185 using strerror. Previously, ERR_reason_error_string() returned
17186 library names as reason strings for SYSerr; but SYSerr is a special
17187 case where small numbers are errno values, not library numbers.
17191 * Add '-dsaparam' option to 'openssl dhparam' application. This
17192 converts DSA parameters into DH parameters. (When creating parameters,
17193 DSA_generate_parameters is used.)
17197 * Include 'length' (recommended exponent length) in C code generated
17198 by 'openssl dhparam -C'.
17202 * The second argument to set_label in perlasm was already being used
17203 so couldn't be used as a "file scope" flag. Moved to third argument
17208 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17209 instead of RAND_bytes for encryption IVs and salts.
17213 * Include RAND_status() into RAND_METHOD instead of implementing
17214 it only for md_rand.c Otherwise replacing the PRNG by calling
17215 RAND_set_rand_method would be impossible.
17219 * Don't let DSA_generate_key() enter an infinite loop if the random
17220 number generation fails.
17224 * New 'rand' application for creating pseudo-random output.
17228 * Added configuration support for Linux/IA64
17230 *Rolf Haberrecker <rolf@suse.de>*
17232 * Assembler module support for Mingw32.
17236 * Shared library support for HPUX (in shlib/).
17238 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17240 * Shared library support for Solaris gcc.
17242 *Lutz Behnke <behnke@trustcenter.de>*
17244 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17246 * PKCS7_encrypt() was adding text MIME headers twice because they
17247 were added manually and by SMIME_crlf_copy().
17251 * In bntest.c don't call BN_rand with zero bits argument.
17253 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17255 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17256 case was implemented. This caused BN_div_recp() to fail occasionally.
17260 * Add an optional second argument to the set_label() in the perl
17261 assembly language builder. If this argument exists and is set
17262 to 1 it signals that the assembler should use a symbol whose
17263 scope is the entire file, not just the current function. This
17264 is needed with MASM which uses the format label:: for this scope.
17266 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17268 * Change the ASN1 types so they are typedefs by default. Before
17269 almost all types were #define'd to ASN1_STRING which was causing
17270 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17275 * Change names of new functions to the new get1/get0 naming
17276 convention: After 'get1', the caller owns a reference count
17277 and has to call `..._free`; 'get0' returns a pointer to some
17278 data structure without incrementing reference counters.
17279 (Some of the existing 'get' functions increment a reference
17280 counter, some don't.)
17281 Similarly, 'set1' and 'add1' functions increase reference
17282 counters or duplicate objects.
17286 * Allow for the possibility of temp RSA key generation failure:
17287 the code used to assume it always worked and crashed on failure.
17291 * Fix potential buffer overrun problem in BIO_printf().
17292 *Ulf Möller, using public domain code by Patrick Powell; problem
17293 pointed out by David Sacerdote <das33@cornell.edu>*
17295 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17296 RAND_egd() and RAND_status(). In the command line application,
17297 the EGD socket can be specified like a seed file using RANDFILE
17302 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17303 Some CAs (e.g. Verisign) distribute certificates in this form.
17307 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17308 list to exclude them. This means that no special compilation option
17309 is needed to use anonymous DH: it just needs to be included in the
17314 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17315 EVP_MD_type. The old functionality is available in a new macro called
17316 EVP_MD_md(). Change code that uses it and update docs.
17320 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17321 where the `void *` argument is replaced by a function pointer argument.
17322 Previously `void *` was abused to point to functions, which works on
17323 many platforms, but is not correct. As these functions are usually
17324 called by macros defined in OpenSSL header files, most source code
17325 should work without changes.
17329 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17330 sections with information on -D... compiler switches used for
17331 compiling the library so that applications can see them. To enable
17332 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17333 must be defined. E.g.,
17334 #define OPENSSL_ALGORITHM_DEFINES
17335 #include <openssl/opensslconf.h>
17336 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17338 *Richard Levitte, Ulf and Bodo Möller*
17340 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17345 * Change the 'other' type in certificate aux info to a STACK_OF
17346 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17347 the required ASN1 format: arbitrary types determined by an OID.
17351 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17352 argument to 'req'. This is not because the function is newer or
17353 better than others it just uses the work 'NEW' in the certificate
17354 request header lines. Some software needs this.
17358 * Reorganise password command line arguments: now passwords can be
17359 obtained from various sources. Delete the PEM_cb function and make
17360 it the default behaviour: i.e. if the callback is NULL and the
17361 usrdata argument is not NULL interpret it as a null terminated pass
17362 phrase. If usrdata and the callback are NULL then the pass phrase
17363 is prompted for as usual.
17367 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17368 the support is automatically enabled. The resulting binaries will
17369 autodetect the card and use it if present.
17371 *Ben Laurie and Compaq Inc.*
17373 * Work around for Netscape hang bug. This sends certificate request
17374 and server done in one record. Since this is perfectly legal in the
17375 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17376 the bugs/SSLv3 entry for more info.
17380 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17384 * Add -rand argument to smime and pkcs12 applications and read/write
17389 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17393 * Add command line password options to the remaining applications.
17397 * Bug fix for BN_div_recp() for numerators with an even number of
17402 * More tests in bntest.c, and changed test_bn output.
17406 * ./config recognizes MacOS X now.
17410 * Bug fix for BN_div() when the first words of num and divisor are
17411 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17415 * Add support for various broken PKCS#8 formats, and command line
17416 options to produce them.
17420 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17421 get temporary BIGNUMs from a BN_CTX.
17425 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17430 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17431 include a #define from the old name to the new. The original intent
17432 was that statically linked binaries could for example just call
17433 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17434 link with digests. This never worked because SSLeay_add_all_digests()
17435 and SSLeay_add_all_ciphers() were in the same source file so calling
17436 one would link with the other. They are now in separate source files.
17440 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17444 * Use a less unusual form of the Miller-Rabin primality test (it used
17445 a binary algorithm for exponentiation integrated into the Miller-Rabin
17446 loop, our standard modexp algorithms are faster).
17450 * Support for the EBCDIC character set completed.
17452 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17454 * Source code cleanups: use const where appropriate, eliminate casts,
17455 use `void *` instead of `char *` in lhash.
17459 * Bugfix: ssl3_send_server_key_exchange was not restartable
17460 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17461 this the server could overwrite ephemeral keys that the client
17466 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17467 using 50 iterations of the Rabin-Miller test.
17469 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17470 iterations of the Rabin-Miller test as required by the appendix
17471 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17472 As BN_is_prime_fasttest includes trial division, DSA parameter
17473 generation becomes much faster.
17475 This implies a change for the callback functions in DSA_is_prime
17476 and DSA_generate_parameters: The callback function is called once
17477 for each positive witness in the Rabin-Miller test, not just
17478 occasionally in the inner loop; and the parameters to the
17479 callback function now provide an iteration count for the outer
17480 loop rather than for the current invocation of the inner loop.
17481 DSA_generate_parameters additionally can call the callback
17482 function with an 'iteration count' of -1, meaning that a
17483 candidate has passed the trial division test (when q is generated
17484 from an application-provided seed, trial division is skipped).
17488 * New function BN_is_prime_fasttest that optionally does trial
17489 division before starting the Rabin-Miller test and has
17490 an additional BN_CTX * argument (whereas BN_is_prime always
17491 has to allocate at least one BN_CTX).
17492 'callback(1, -1, cb_arg)' is called when a number has passed the
17493 trial division stage.
17497 * Fix for bug in CRL encoding. The validity dates weren't being handled
17502 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17506 * New function BN_pseudo_rand().
17510 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17511 bignum version of BN_from_montgomery() with the working code from
17512 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17517 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17518 made it impossible to use the same SSL_SESSION data structure in
17519 SSL2 clients in multiple threads.
17523 * The return value of RAND_load_file() no longer counts bytes obtained
17524 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17525 to seed the PRNG (previously an explicit byte count was required).
17527 *Ulf Möller, Bodo Möller*
17529 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17530 used `char *` instead of `void *` and had casts all over the place.
17534 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17538 * Retain source code compatibility for BN_prime_checks macro:
17539 BN_is_prime(..., BN_prime_checks, ...) now uses
17540 BN_prime_checks_for_size to determine the appropriate number of
17541 Rabin-Miller iterations.
17545 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17546 DH_CHECK_P_NOT_SAFE_PRIME.
17547 (Check if this is true? OpenPGP calls them "strong".)
17551 * Merge the functionality of "dh" and "gendh" programs into a new program
17552 "dhparam". The old programs are retained for now but will handle DH keys
17553 (instead of parameters) in future.
17557 * Make the ciphers, s_server and s_client programs check the return values
17558 when a new cipher list is set.
17562 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17563 ciphers. Before when the 56bit ciphers were enabled the sorting was
17566 The syntax for the cipher sorting has been extended to support sorting by
17567 cipher-strength (using the strength_bits hard coded in the tables).
17568 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17570 Fix a bug in the cipher-command parser: when supplying a cipher command
17571 string with an "undefined" symbol (neither command nor alphanumeric
17572 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17573 an error is flagged.
17575 Due to the strength-sorting extension, the code of the
17576 ssl_create_cipher_list() function was completely rearranged. I hope that
17577 the readability was also increased :-)
17579 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17581 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17582 for the first serial number and places 2 in the serial number file. This
17583 avoids problems when the root CA is created with serial number zero and
17584 the first user certificate has the same issuer name and serial number
17589 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17590 the new code. Add documentation for this stuff.
17594 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17595 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17596 structures and behave in an analogous way to the X509v3 functions:
17597 they shouldn't be called directly but wrapper functions should be used
17600 So we also now have some wrapper functions that call the X509at functions
17601 when passed certificate requests. (TO DO: similar things can be done with
17602 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17603 things. Some of these need some d2i or i2d and print functionality
17604 because they handle more complex structures.)
17608 * Add missing #ifndefs that caused missing symbols when building libssl
17609 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17610 NO_RSA in `ssl/s2*.c`.
17612 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17614 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17615 has a return value which indicates the quality of the random data
17616 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17617 error queue. New function RAND_pseudo_bytes() generates output that is
17618 guaranteed to be unique but not unpredictable. RAND_add is like
17619 RAND_seed, but takes an extra argument for an entropy estimate
17620 (RAND_seed always assumes full entropy).
17624 * Do more iterations of Rabin-Miller probable prime test (specifically,
17625 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17626 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17627 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17628 false-positive rate of at most 2^-80 for random input.
17632 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17636 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17637 in the 0.9.5 release), this returns the chain
17638 from an X509_CTX structure with a dup of the stack and all
17639 the X509 reference counts upped: so the stack will exist
17640 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17643 Also make SSL_SESSION_print() print out the verify return
17648 * Add manpage for the pkcs12 command. Also change the default
17649 behaviour so MAC iteration counts are used unless the new
17650 -nomaciter option is used. This improves file security and
17651 only older versions of MSIE (4.0 for example) need it.
17655 * Honor the no-xxx Configure options when creating .DEF files.
17659 * Add PKCS#10 attributes to field table: challengePassword,
17660 unstructuredName and unstructuredAddress. These are taken from
17661 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17662 international characters are used.
17664 More changes to X509_ATTRIBUTE code: allow the setting of types
17665 based on strings. Remove the 'loc' parameter when adding
17666 attributes because these will be a SET OF encoding which is sorted
17671 * Initial changes to the 'req' utility to allow request generation
17672 automation. This will allow an application to just generate a template
17673 file containing all the field values and have req construct the
17676 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17677 used all over the place including certificate requests and PKCS#7
17678 structures. They are currently handled manually where necessary with
17679 some primitive wrappers for PKCS#7. The new functions behave in a
17680 manner analogous to the X509 extension functions: they allow
17681 attributes to be looked up by NID and added.
17683 Later something similar to the X509V3 code would be desirable to
17684 automatically handle the encoding, decoding and printing of the
17685 more complex types. The string types like challengePassword can
17686 be handled by the string table functions.
17688 Also modified the multi byte string table handling. Now there is
17689 a 'global mask' which masks out certain types. The table itself
17690 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17691 is useful when for example there is only one permissible type
17692 (as in countryName) and using the mask might result in no valid
17697 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17698 SSL_get_peer_finished to allow applications to obtain the latest
17699 Finished messages sent to the peer or expected from the peer,
17700 respectively. (SSL_get_peer_finished is usually the Finished message
17701 actually received from the peer, otherwise the protocol will be aborted.)
17703 As the Finished message are message digests of the complete handshake
17704 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17705 be used for external authentication procedures when the authentication
17706 provided by SSL/TLS is not desired or is not enough.
17710 * Enhanced support for Alpha Linux is added. Now ./config checks if
17711 the host supports BWX extension and if Compaq C is present on the
17712 $PATH. Just exploiting of the BWX extension results in 20-30%
17713 performance kick for some algorithms, e.g. DES and RC4 to mention
17714 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17719 * Add support for MS "fast SGC". This is arguably a violation of the
17720 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17721 weak crypto and after checking the certificate is SGC a second one
17722 with strong crypto. MS SGC stops the first handshake after receiving
17723 the server certificate message and sends a second client hello. Since
17724 a server will typically do all the time consuming operations before
17725 expecting any further messages from the client (server key exchange
17726 is the most expensive) there is little difference between the two.
17728 To get OpenSSL to support MS SGC we have to permit a second client
17729 hello message after we have sent server done. In addition we have to
17730 reset the MAC if we do get this second client hello.
17734 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17735 if a DER encoded private key is RSA or DSA traditional format. Changed
17736 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17737 format DER encoded private key. Newer code should use PKCS#8 format which
17738 has the key type encoded in the ASN1 structure. Added DER private key
17739 support to pkcs8 application.
17743 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17744 ciphersuites has been selected (as required by the SSL 3/TLS 1
17745 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17746 is set, we interpret this as a request to violate the specification
17747 (the worst that can happen is a handshake failure, and 'correct'
17748 behaviour would result in a handshake failure anyway).
17752 * In SSL_CTX_add_session, take into account that there might be multiple
17753 SSL_SESSION structures with the same session ID (e.g. when two threads
17754 concurrently obtain them from an external cache).
17755 The internal cache can handle only one SSL_SESSION with a given ID,
17756 so if there's a conflict, we now throw out the old one to achieve
17761 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17762 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17763 some routines that use cipher OIDs: some ciphers do not have OIDs
17764 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17769 * Simplify the trust setting structure and code. Now we just have
17770 two sequences of OIDs for trusted and rejected settings. These will
17771 typically have values the same as the extended key usage extension
17772 and any application specific purposes.
17774 The trust checking code now has a default behaviour: it will just
17775 check for an object with the same NID as the passed id. Functions can
17776 be provided to override either the default behaviour or the behaviour
17777 for a given id. SSL client, server and email already have functions
17778 in place for compatibility: they check the NID and also return "trusted"
17779 if the certificate is self signed.
17783 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17784 traditional format into an EVP_PKEY structure.
17788 * Add a password callback function PEM_cb() which either prompts for
17789 a password if usr_data is NULL or otherwise assumes it is a null
17790 terminated password. Allow passwords to be passed on command line
17791 environment or config files in a few more utilities.
17795 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17796 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17797 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17798 Update documentation.
17802 * Support for ASN1 "NULL" type. This could be handled before by using
17803 ASN1_TYPE but there wasn't any function that would try to read a NULL
17804 and produce an error if it couldn't. For compatibility we also have
17805 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17806 don't allocate anything because they don't need to.
17810 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17813 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17815 * Rebuild of the memory allocation routines used by OpenSSL code and
17816 possibly others as well. The purpose is to make an interface that
17817 provide hooks so anyone can build a separate set of allocation and
17818 deallocation routines to be used by OpenSSL, for example memory
17819 pool implementations, or something else, which was previously hard
17820 since Malloc(), Realloc() and Free() were defined as macros having
17821 the values malloc, realloc and free, respectively (except for Win32
17822 compilations). The same is provided for memory debugging code.
17823 OpenSSL already comes with functionality to find memory leaks, but
17824 this gives people a chance to debug other memory problems.
17826 With these changes, a new set of functions and macros have appeared:
17828 CRYPTO_set_mem_debug_functions() [F]
17829 CRYPTO_get_mem_debug_functions() [F]
17830 CRYPTO_dbg_set_options() [F]
17831 CRYPTO_dbg_get_options() [F]
17832 CRYPTO_malloc_debug_init() [M]
17834 The memory debug functions are NULL by default, unless the library
17835 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17836 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17837 gives the standard debugging functions that come with OpenSSL) or
17838 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17839 provided by the library user) must be used. When the standard
17840 debugging functions are used, CRYPTO_dbg_set_options can be used to
17841 request additional information:
17842 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17843 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17845 Also, things like CRYPTO_set_mem_functions will always give the
17846 expected result (the new set of functions is used for allocation
17847 and deallocation) at all times, regardless of platform and compiler
17850 To finish it up, some functions that were never use in any other
17851 way than through macros have a new API and new semantic:
17853 CRYPTO_dbg_malloc()
17854 CRYPTO_dbg_realloc()
17857 All macros of value have retained their old syntax.
17859 *Richard Levitte and Bodo Moeller*
17861 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17862 ordering of SMIMECapabilities wasn't in "strength order" and there
17863 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17868 * Some ASN1 types with illegal zero length encoding (INTEGER,
17869 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17871 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17873 * Merge in my S/MIME library for OpenSSL. This provides a simple
17874 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17875 functionality to handle multipart/signed properly) and a utility
17876 called 'smime' to call all this stuff. This is based on code I
17877 originally wrote for Celo who have kindly allowed it to be
17878 included in OpenSSL.
17882 * Add variants des_set_key_checked and des_set_key_unchecked of
17883 des_set_key (aka des_key_sched). Global variable des_check_key
17884 decides which of these is called by des_set_key; this way
17885 des_check_key behaves as it always did, but applications and
17886 the library itself, which was buggy for des_check_key == 1,
17887 have a cleaner way to pick the version they need.
17891 * New function PKCS12_newpass() which changes the password of a
17896 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17897 dynamic mix. In both cases the ids can be used as an index into the
17898 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17899 functions so they accept a list of the field values and the
17900 application doesn't need to directly manipulate the X509_TRUST
17905 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17910 * Modify the way the V3 extension code looks up extensions. This now
17911 works in a similar way to the object code: we have some "standard"
17912 extensions in a static table which is searched with OBJ_bsearch()
17913 and the application can add dynamic ones if needed. The file
17914 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17915 updated whenever a new extension is added to the core code and kept
17916 in ext_nid order. There is a simple program 'tabtest.c' which checks
17917 this. New extensions are not added too often so this file can readily
17918 be maintained manually.
17920 There are two big advantages in doing things this way. The extensions
17921 can be looked up immediately and no longer need to be "added" using
17922 X509V3_add_standard_extensions(): this function now does nothing.
17923 Side note: I get *lots* of email saying the extension code doesn't
17924 work because people forget to call this function.
17925 Also no dynamic allocation is done unless new extensions are added:
17926 so if we don't add custom extensions there is no need to call
17927 X509V3_EXT_cleanup().
17931 * Modify enc utility's salting as follows: make salting the default. Add a
17932 magic header, so unsalted files fail gracefully instead of just decrypting
17933 to garbage. This is because not salting is a big security hole, so people
17934 should be discouraged from doing it.
17938 * Fixes and enhancements to the 'x509' utility. It allowed a message
17939 digest to be passed on the command line but it only used this
17940 parameter when signing a certificate. Modified so all relevant
17941 operations are affected by the digest parameter including the
17942 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17943 DSA key was used because it didn't fix the digest.
17947 * Initial certificate chain verify code. Currently tests the untrusted
17948 certificates for consistency with the verify purpose (which is set
17949 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17951 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17952 this is because it will reject chains with invalid extensions whereas
17953 every previous version of OpenSSL and SSLeay made no checks at all.
17955 Trust code: checks the root CA for the relevant trust settings. Trust
17956 settings have an initial value consistent with the verify purpose: e.g.
17957 if the verify purpose is for SSL client use it expects the CA to be
17958 trusted for SSL client use. However the default value can be changed to
17959 permit custom trust settings: one example of this would be to only trust
17960 certificates from a specific "secure" set of CAs.
17962 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17963 which should be used for version portability: especially since the
17964 verify structure is likely to change more often now.
17966 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17967 to set them. If not set then assume SSL clients will verify SSL servers
17970 Two new options to the verify program: -untrusted allows a set of
17971 untrusted certificates to be passed in and -purpose which sets the
17972 intended purpose of the certificate. If a purpose is set then the
17973 new chain verify code is used to check extension consistency.
17977 * Support for the authority information access extension.
17981 * Modify RSA and DSA PEM read routines to transparently handle
17982 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17983 public keys in a format compatible with certificate
17984 SubjectPublicKeyInfo structures. Unfortunately there were already
17985 functions called *_PublicKey_* which used various odd formats so
17986 these are retained for compatibility: however the DSA variants were
17987 never in a public release so they have been deleted. Changed dsa/rsa
17988 utilities to handle the new format: note no releases ever handled public
17989 keys so we should be OK.
17991 The primary motivation for this change is to avoid the same fiasco
17992 that dogs private keys: there are several incompatible private key
17993 formats some of which are standard and some OpenSSL specific and
17994 require various evil hacks to allow partial transparent handling and
17995 even then it doesn't work with DER formats. Given the option anything
17996 other than PKCS#8 should be dumped: but the other formats have to
17997 stay in the name of compatibility.
17999 With public keys and the benefit of hindsight one standard format
18000 is used which works with EVP_PKEY, RSA or DSA structures: though
18001 it clearly returns an error if you try to read the wrong kind of key.
18003 Added a -pubkey option to the 'x509' utility to output the public key.
18004 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18005 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18006 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18007 that do the same as the `EVP_PKEY_assign_*()` except they up the
18008 reference count of the added key (they don't "swallow" the
18013 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18014 CRLs would fail if the file contained no certificates or no CRLs:
18015 added a new function to read in both types and return the number
18016 read: this means that if none are read it will be an error. The
18017 DER versions of the certificate and CRL reader would always fail
18018 because it isn't possible to mix certificates and CRLs in DER format
18019 without choking one or the other routine. Changed this to just read
18020 a certificate: this is the best we can do. Also modified the code
18021 in `apps/verify.c` to take notice of return codes: it was previously
18022 attempting to read in certificates from NULL pointers and ignoring
18023 any errors: this is one reason why the cert and CRL reader seemed
18024 to work. It doesn't check return codes from the default certificate
18025 routines: these may well fail if the certificates aren't installed.
18029 * Code to support otherName option in GeneralName.
18033 * First update to verify code. Change the verify utility
18034 so it warns if it is passed a self signed certificate:
18035 for consistency with the normal behaviour. X509_verify
18036 has been modified to it will now verify a self signed
18037 certificate if *exactly* the same certificate appears
18038 in the store: it was previously impossible to trust a
18039 single self signed certificate. This means that:
18040 openssl verify ss.pem
18041 now gives a warning about a self signed certificate but
18042 openssl verify -CAfile ss.pem ss.pem
18047 * For servers, store verify_result in SSL_SESSION data structure
18048 (and add it to external session representation).
18049 This is needed when client certificate verifications fails,
18050 but an application-provided verification callback (set by
18051 SSL_CTX_set_cert_verify_callback) allows accepting the session
18052 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18053 but returns 1): When the session is reused, we have to set
18054 ssl->verify_result to the appropriate error code to avoid
18057 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18059 * Fix a bug in the new PKCS#7 code: it didn't consider the
18060 case in PKCS7_dataInit() where the signed PKCS7 structure
18061 didn't contain any existing data because it was being created.
18063 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18065 * Add a salt to the key derivation routines in enc.c. This
18066 forms the first 8 bytes of the encrypted file. Also add a
18067 -S option to allow a salt to be input on the command line.
18071 * New function X509_cmp(). Oddly enough there wasn't a function
18072 to compare two certificates. We do this by working out the SHA1
18073 hash and comparing that. X509_cmp() will be needed by the trust
18078 * SSL_get1_session() is like SSL_get_session(), but increments
18079 the reference count in the SSL_SESSION returned.
18081 *Geoff Thorpe <geoff@eu.c2.net>*
18083 * Fix for 'req': it was adding a null to request attributes.
18084 Also change the X509_LOOKUP and X509_INFO code to handle
18085 certificate auxiliary information.
18089 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18094 * Add the possibility to add extra information to the memory leak
18095 detecting output, to form tracebacks, showing from where each
18096 allocation was originated: CRYPTO_push_info("constant string") adds
18097 the string plus current file name and line number to a per-thread
18098 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18099 is like calling CYRPTO_pop_info() until the stack is empty.
18100 Also updated memory leak detection code to be multi-thread-safe.
18104 * Add options -text and -noout to pkcs7 utility and delete the
18105 encryption options which never did anything. Update docs.
18109 * Add options to some of the utilities to allow the pass phrase
18110 to be included on either the command line (not recommended on
18111 OSes like Unix) or read from the environment. Update the
18112 manpages and fix a few bugs.
18116 * Add a few manpages for some of the openssl commands.
18120 * Fix the -revoke option in ca. It was freeing up memory twice,
18121 leaking and not finding already revoked certificates.
18125 * Extensive changes to support certificate auxiliary information.
18126 This involves the use of X509_CERT_AUX structure and X509_AUX
18127 functions. An X509_AUX function such as PEM_read_X509_AUX()
18128 can still read in a certificate file in the usual way but it
18129 will also read in any additional "auxiliary information". By
18130 doing things this way a fair degree of compatibility can be
18131 retained: existing certificates can have this information added
18132 using the new 'x509' options.
18134 Current auxiliary information includes an "alias" and some trust
18135 settings. The trust settings will ultimately be used in enhanced
18136 certificate chain verification routines: currently a certificate
18137 can only be trusted if it is self signed and then it is trusted
18142 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18143 The problem was that one of the replacement routines had not been working
18144 since SSLeay releases. For now the offending routine has been replaced
18145 with non-optimised assembler. Even so, this now gives around 95%
18146 performance improvement for 1024 bit RSA signs.
18150 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18151 handling. Most clients have the effective key size in bits equal to
18152 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18153 A few however don't do this and instead use the size of the decrypted key
18154 to determine the RC2 key length and the AlgorithmIdentifier to determine
18155 the effective key length. In this case the effective key length can still
18156 be 40 bits but the key length can be 168 bits for example. This is fixed
18157 by manually forcing an RC2 key into the EVP_PKEY structure because the
18158 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18159 the key length and effective key length are equal.
18163 * Add a bunch of functions that should simplify the creation of
18164 X509_NAME structures. Now you should be able to do:
18165 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18166 and have it automatically work out the correct field type and fill in
18167 the structures. The more adventurous can try:
18168 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18169 and it will (hopefully) work out the correct multibyte encoding.
18173 * Change the 'req' utility to use the new field handling and multibyte
18174 copy routines. Before the DN field creation was handled in an ad hoc
18175 way in req, ca, and x509 which was rather broken and didn't support
18176 BMPStrings or UTF8Strings. Since some software doesn't implement
18177 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18178 using the dirstring_type option. See the new comment in the default
18179 openssl.cnf for more info.
18183 * Make crypto/rand/md_rand.c more robust:
18184 - Assure unique random numbers after fork().
18185 - Make sure that concurrent threads access the global counter and
18186 md serializably so that we never lose entropy in them
18187 or use exactly the same state in multiple threads.
18188 Access to the large state is not always serializable because
18189 the additional locking could be a performance killer, and
18190 md should be large enough anyway.
18194 * New file `apps/app_rand.c` with commonly needed functionality
18195 for handling the random seed file.
18197 Use the random seed file in some applications that previously did not:
18199 dsaparam -genkey (which also ignored its '-rand' option),
18202 x509 (when signing).
18203 Except on systems with /dev/urandom, it is crucial to have a random
18204 seed file at least for key creation, DSA signing, and for DH exchanges;
18205 for RSA signatures we could do without one.
18207 gendh and gendsa (unlike genrsa) used to read only the first byte
18208 of each file listed in the '-rand' option. The function as previously
18209 found in genrsa is now in app_rand.c and is used by all programs
18210 that support '-rand'.
18214 * In RAND_write_file, use mode 0600 for creating files;
18215 don't just chmod when it may be too late.
18219 * Report an error from X509_STORE_load_locations
18220 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18224 * New function ASN1_mbstring_copy() this copies a string in either
18225 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18226 into an ASN1_STRING type. A mask of permissible types is passed
18227 and it chooses the "minimal" type to use or an error if not type
18232 * Add function equivalents to the various macros in asn1.h. The old
18233 macros are retained with an `M_` prefix. Code inside the library can
18234 use the `M_` macros. External code (including the openssl utility)
18235 should *NOT* in order to be "shared library friendly".
18239 * Add various functions that can check a certificate's extensions
18240 to see if it usable for various purposes such as SSL client,
18241 server or S/MIME and CAs of these types. This is currently
18242 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18243 verification. Also added a -purpose flag to x509 utility to
18244 print out all the purposes.
18248 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18253 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18254 for, obtain and decode and extension and obtain its critical flag.
18255 This allows all the necessary extension code to be handled in a
18256 single function call.
18260 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18261 platforms. See crypto/rc4/rc4_enc.c for further details.
18265 * New -noout option to asn1parse. This causes no output to be produced
18266 its main use is when combined with -strparse and -out to extract data
18267 from a file (which may not be in ASN.1 format).
18271 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18272 when producing the local key id.
18274 *Richard Levitte <levitte@stacken.kth.se>*
18276 * New option -dhparam in s_server. This allows a DH parameter file to be
18277 stated explicitly. If it is not stated then it tries the first server
18278 certificate file. The previous behaviour hard coded the filename
18283 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18284 a public key to be input or output. For example:
18285 openssl rsa -in key.pem -pubout -out pubkey.pem
18286 Also added necessary DSA public key functions to handle this.
18290 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18291 in the message. This was handled by allowing
18292 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18294 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18296 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18297 to the end of the strings whereas this didn't. This would cause problems
18298 if strings read with d2i_ASN1_bytes() were later modified.
18300 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18302 * Fix for base64 decode bug. When a base64 bio reads only one line of
18303 data and it contains EOF it will end up returning an error. This is
18304 caused by input 46 bytes long. The cause is due to the way base64
18305 BIOs find the start of base64 encoded data. They do this by trying a
18306 trial decode on each line until they find one that works. When they
18307 do a flag is set and it starts again knowing it can pass all the
18308 data directly through the decoder. Unfortunately it doesn't reset
18309 the context it uses. This means that if EOF is reached an attempt
18310 is made to pass two EOFs through the context and this causes the
18311 resulting error. This can also cause other problems as well. As is
18312 usual with these problems it takes *ages* to find and the fix is
18313 trivial: move one line.
18315 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18317 * Ugly workaround to get s_client and s_server working under Windows. The
18318 old code wouldn't work because it needed to select() on sockets and the
18319 tty (for keypresses and to see if data could be written). Win32 only
18320 supports select() on sockets so we select() with a 1s timeout on the
18321 sockets and then see if any characters are waiting to be read, if none
18322 are present then we retry, we also assume we can always write data to
18323 the tty. This isn't nice because the code then blocks until we've
18324 received a complete line of data and it is effectively polling the
18325 keyboard at 1s intervals: however it's quite a bit better than not
18326 working at all :-) A dedicated Windows application might handle this
18327 with an event loop for example.
18331 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18332 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18333 will be called when RSA_sign() and RSA_verify() are used. This is useful
18334 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18335 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18336 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18337 This necessitated the support of an extra signature type NID_md5_sha1
18338 for SSL signatures and modifications to the SSL library to use it instead
18339 of calling RSA_public_decrypt() and RSA_private_encrypt().
18343 * Add new -verify -CAfile and -CApath options to the crl program, these
18344 will lookup a CRL issuers certificate and verify the signature in a
18345 similar way to the verify program. Tidy up the crl program so it
18346 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18347 less strict. It will now permit CRL extensions even if it is not
18348 a V2 CRL: this will allow it to tolerate some broken CRLs.
18352 * Initialize all non-automatic variables each time one of the openssl
18353 sub-programs is started (this is necessary as they may be started
18354 multiple times from the "OpenSSL>" prompt).
18356 *Lennart Bang, Bodo Moeller*
18358 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18359 removing all other RSA functionality (this is what NO_RSA does). This
18360 is so (for example) those in the US can disable those operations covered
18361 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18366 * Non-copying interface to BIO pairs.
18367 (still largely untested)
18371 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18372 ASCII string. This was handled independently in various places before.
18376 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18377 UTF8 strings a character at a time.
18381 * Use client_version from client hello to select the protocol
18382 (s23_srvr.c) and for RSA client key exchange verification
18383 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18387 * Add various utility functions to handle SPKACs, these were previously
18388 handled by poking round in the structure internals. Added new function
18389 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18390 print, verify and generate SPKACs. Based on an original idea from
18391 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18395 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18399 * Allow the config file extension section to be overwritten on the
18400 command line. Based on an original idea from Massimiliano Pala
18401 <madwolf@comune.modena.it>. The new option is called -extensions
18402 and can be applied to ca, req and x509. Also -reqexts to override
18403 the request extensions in req and -crlexts to override the crl extensions
18408 * Add new feature to the SPKAC handling in ca. Now you can include
18409 the same field multiple times by preceding it by "XXXX." for example:
18412 this is the same syntax as used in the req config file.
18416 * Allow certificate extensions to be added to certificate requests. These
18417 are specified in a 'req_extensions' option of the req section of the
18418 config file. They can be printed out with the -text option to req but
18419 are otherwise ignored at present.
18423 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18424 data read consists of only the final block it would not decrypted because
18425 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18426 A misplaced 'break' also meant the decrypted final block might not be
18427 copied until the next read.
18431 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18432 a few extra parameters to the DH structure: these will be useful if
18433 for example we want the value of 'q' or implement X9.42 DH.
18437 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18438 provides hooks that allow the default DSA functions or functions on a
18439 "per key" basis to be replaced. This allows hardware acceleration and
18440 hardware key storage to be handled without major modification to the
18441 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18442 associated functions.
18446 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18447 as "read only": it can't be written to and the buffer it points to will
18448 not be freed. Reading from a read only BIO is much more efficient than
18449 a normal memory BIO. This was added because there are several times when
18450 an area of memory needs to be read from a BIO. The previous method was
18451 to create a memory BIO and write the data to it, this results in two
18452 copies of the data and an O(n^2) reading algorithm. There is a new
18453 function BIO_new_mem_buf() which creates a read only memory BIO from
18454 an area of memory. Also modified the PKCS#7 routines to use read only
18459 * Bugfix: ssl23_get_client_hello did not work properly when called in
18460 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18461 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18462 but a retry condition occurred while trying to read the rest.
18466 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18467 NID_pkcs7_encrypted by default: this was wrong since this should almost
18468 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18469 the encrypted data type: this is a more sensible place to put it and it
18470 allows the PKCS#12 code to be tidied up that duplicated this
18475 * Changed obj_dat.pl script so it takes its input and output files on
18476 the command line. This should avoid shell escape redirection problems
18481 * Initial support for certificate extension requests, these are included
18482 in things like Xenroll certificate requests. Included functions to allow
18483 extensions to be obtained and added.
18487 * -crlf option to s_client and s_server for sending newlines as
18488 CRLF (as required by many protocols).
18492 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18494 * Install libRSAglue.a when OpenSSL is built with RSAref.
18496 *Ralf S. Engelschall*
18498 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18500 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18502 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18507 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18508 DH parameters/keys (q is lost during that conversion, but the resulting
18509 DH parameters contain its length).
18511 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18512 much faster than DH_generate_parameters (which creates parameters
18513 where `p = 2*q + 1`), and also the smaller q makes DH computations
18514 much more efficient (160-bit exponentiation instead of 1024-bit
18515 exponentiation); so this provides a convenient way to support DHE
18516 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18517 utter importance to use
18518 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18520 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18521 when such DH parameters are used, because otherwise small subgroup
18522 attacks may become possible!
18526 * Avoid memory leak in i2d_DHparams.
18530 * Allow the -k option to be used more than once in the enc program:
18531 this allows the same encrypted message to be read by multiple recipients.
18535 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18536 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18537 it will always use the numerical form of the OID, even if it has a short
18542 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18543 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18544 otherwise bn_mod_exp was called. In the case of hardware keys for example
18545 no private key components need be present and it might store extra data
18546 in the RSA structure, which cannot be accessed from bn_mod_exp.
18547 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18548 private key operations.
18552 * Added support for SPARC Linux.
18556 * pem_password_cb function type incompatibly changed from
18557 typedef int pem_password_cb(char *buf, int size, int rwflag);
18559 ....(char *buf, int size, int rwflag, void *userdata);
18560 so that applications can pass data to their callbacks:
18561 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18562 additional void * argument, which is just handed through whenever
18563 the password callback is called.
18565 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18567 New function SSL_CTX_set_default_passwd_cb_userdata.
18569 Compatibility note: As many C implementations push function arguments
18570 onto the stack in reverse order, the new library version is likely to
18571 interoperate with programs that have been compiled with the old
18572 pem_password_cb definition (PEM_whatever takes some data that
18573 happens to be on the stack as its last argument, and the callback
18574 just ignores this garbage); but there is no guarantee whatsoever that
18577 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18578 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18579 problems not only on Windows, but also on some Unix platforms.
18580 To avoid problematic command lines, these definitions are now in an
18581 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18582 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18586 * MIPS III/IV assembler module is reimplemented.
18590 * More DES library cleanups: remove references to srand/rand and
18591 delete an unused file.
18595 * Add support for the free Netwide assembler (NASM) under Win32,
18596 since not many people have MASM (ml) and it can be hard to obtain.
18597 This is currently experimental but it seems to work OK and pass all
18598 the tests. Check out INSTALL.W32 for info.
18602 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18603 without temporary keys kept an extra copy of the server key,
18604 and connections with temporary keys did not free everything in case
18609 * New function RSA_check_key and new openssl rsa option -check
18610 for verifying the consistency of RSA keys.
18612 *Ulf Moeller, Bodo Moeller*
18614 * Various changes to make Win32 compile work:
18615 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18616 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18617 comparison" warnings.
18618 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18622 * Add a debugging option to PKCS#5 v2 key generation function: when
18623 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18624 derived keys are printed to stderr.
18628 * Copy the flags in ASN1_STRING_dup().
18630 *Roman E. Pavlov <pre@mo.msk.ru>*
18632 * The x509 application mishandled signing requests containing DSA
18633 keys when the signing key was also DSA and the parameters didn't match.
18635 It was supposed to omit the parameters when they matched the signing key:
18636 the verifying software was then supposed to automatically use the CA's
18637 parameters if they were absent from the end user certificate.
18639 Omitting parameters is no longer recommended. The test was also
18640 the wrong way round! This was probably due to unusual behaviour in
18641 EVP_cmp_parameters() which returns 1 if the parameters match.
18642 This meant that parameters were omitted when they *didn't* match and
18643 the certificate was useless. Certificates signed with 'ca' didn't have
18646 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18648 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18649 The interface is as follows:
18650 Applications can use
18651 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18652 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18653 "off" is now the default.
18654 The library internally uses
18655 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18656 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18657 to disable memory-checking temporarily.
18659 Some inconsistent states that previously were possible (and were
18660 even the default) are now avoided.
18662 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18663 with each memory chunk allocated; this is occasionally more helpful
18664 than just having a counter.
18666 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18668 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18673 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18674 which largely parallels "options", but is for changing API behaviour,
18675 whereas "options" are about protocol behaviour.
18676 Initial "mode" flags are:
18678 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18679 a single record has been written.
18680 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18681 retries use the same buffer location.
18682 (But all of the contents must be
18687 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18690 * Fix problems with no-hmac etc.
18692 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18694 * New functions RSA_get_default_method(), RSA_set_method() and
18695 RSA_get_method(). These allows replacement of RSA_METHODs without having
18696 to mess around with the internals of an RSA structure.
18700 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18701 Also really enable memory leak checks in openssl.c and in some
18704 *Chad C. Mulligan, Bodo Moeller*
18706 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18707 up the length of negative integers. This has now been simplified to just
18708 store the length when it is first determined and use it later, rather
18709 than trying to keep track of where data is copied and updating it to
18711 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18713 * Add a new function PKCS7_signatureVerify. This allows the verification
18714 of a PKCS#7 signature but with the signing certificate passed to the
18715 function itself. This contrasts with PKCS7_dataVerify which assumes the
18716 certificate is present in the PKCS#7 structure. This isn't always the
18717 case: certificates can be omitted from a PKCS#7 structure and be
18718 distributed by "out of band" means (such as a certificate database).
18722 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18723 function prototypes in pem.h, also change util/mkdef.pl to add the
18724 necessary function names.
18728 * mk1mf.pl (used by Windows builds) did not properly read the
18729 options set by Configure in the top level Makefile, and Configure
18730 was not even able to write more than one option correctly.
18731 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18735 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18736 file to be loaded from a BIO or FILE pointer. The BIO version will
18737 for example allow memory BIOs to contain config info.
18741 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18742 Whoever hopes to achieve shared-library compatibility across versions
18743 must use this, not the compile-time macro.
18744 (Exercise 0.9.4: Which is the minimum library version required by
18746 Note: All this applies only to multi-threaded programs, others don't
18751 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18752 through a BIO pair triggered the default case, i.e.
18753 SSLerr(...,SSL_R_UNKNOWN_STATE).
18757 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18758 can use the SSL library even if none of the specific BIOs is
18763 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18764 for the encoded length.
18766 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18768 * Add initial documentation of the X509V3 functions.
18772 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18773 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18774 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18775 secure PKCS#8 private key format with a high iteration count.
18779 * Fix determination of Perl interpreter: A perl or perl5
18780 *directory* in $PATH was also accepted as the interpreter.
18782 *Ralf S. Engelschall*
18784 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18785 wrong with it but it was very old and did things like calling
18786 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18787 unusual formatting.
18791 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18792 to use the new extension code.
18796 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18797 with macros. This should make it easier to change their form, add extra
18798 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18803 * Add to configuration table a new entry that can specify an alternative
18804 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18805 according to Mark Crispin <MRC@Panda.COM>.
18809 * DES CBC did not update the IV. Weird.
18813 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18814 Changing the behaviour of the former might break existing programs --
18815 where IV updating is needed, des_ncbc_encrypt can be used.
18818 * When bntest is run from "make test" it drives bc to check its
18819 calculations, as well as internally checking them. If an internal check
18820 fails, it needs to cause bc to give a non-zero result or make test carries
18821 on without noticing the failure. Fixed.
18825 * DES library cleanups.
18829 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18830 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18831 ciphers. NOTE: although the key derivation function has been verified
18832 against some published test vectors it has not been extensively tested
18833 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18838 * Instead of "mkdir -p", which is not fully portable, use new
18839 Perl script "util/mkdir-p.pl".
18843 * Rewrite the way password based encryption (PBE) is handled. It used to
18844 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18845 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18846 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18847 the 'parameter' field of the AlgorithmIdentifier is passed to the
18848 underlying key generation function so it must do its own ASN1 parsing.
18849 This has also changed the EVP_PBE_CipherInit() function which now has a
18850 'parameter' argument instead of literal salt and iteration count values
18851 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18855 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18856 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18857 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18858 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18859 value was just used as a "magic string" and not used directly its
18860 value doesn't matter.
18864 * Introduce some semblance of const correctness to BN. Shame C doesn't
18869 * "linux-sparc64" configuration (ultrapenguin).
18871 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18872 "linux-sparc" configuration.
18874 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18876 * config now generates no-xxx options for missing ciphers.
18880 * Support the EBCDIC character set (work in progress).
18881 File ebcdic.c not yet included because it has a different license.
18883 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18885 * Support BS2000/OSD-POSIX.
18887 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18889 * Make callbacks for key generation use `void *` instead of `char *`.
18893 * Make S/MIME samples compile (not yet tested).
18897 * Additional typesafe stacks.
18901 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18905 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18907 * New configuration variant "sco5-gcc".
18909 * Updated some demos.
18911 *Sean O Riordain, Wade Scholine*
18913 * Add missing BIO_free at exit of pkcs12 application.
18917 * Fix memory leak in conf.c.
18921 * Updates for Win32 to assembler version of MD5.
18925 * Set #! path to perl in `apps/der_chop` to where we found it
18926 instead of using a fixed path.
18930 * SHA library changes for irix64-mips4-cc.
18934 * Improvements for VMS support.
18938 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18940 * Bignum library bug fix. IRIX 6 passes "make test" now!
18941 This also avoids the problems with SC4.2 and unpatched SC5.
18943 *Andy Polyakov <appro@fy.chalmers.se>*
18945 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18946 These are required because of the typesafe stack would otherwise break
18947 existing code. If old code used a structure member which used to be STACK
18948 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18949 sk_num or sk_value it would produce an error because the num, data members
18950 are not present in STACK_OF. Now it just produces a warning. sk_set
18951 replaces the old method of assigning a value to sk_value
18952 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18953 that does this will no longer work (and should use sk_set instead) but
18954 this could be regarded as a "questionable" behaviour anyway.
18958 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18959 correctly handle encrypted S/MIME data.
18963 * Change type of various DES function arguments from des_cblock
18964 (which means, in function argument declarations, pointer to char)
18965 to des_cblock * (meaning pointer to array with 8 char elements),
18966 which allows the compiler to do more typechecking; it was like
18967 that back in SSLeay, but with lots of ugly casts.
18969 Introduce new type const_des_cblock.
18973 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18974 problems: find RecipientInfo structure that matches recipient certificate
18975 and initialise the ASN1 structures properly based on passed cipher.
18979 * Belatedly make the BN tests actually check the results.
18983 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18984 to and from BNs: it was completely broken. New compilation option
18985 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18986 key elements as negative integers.
18990 * Reorganize and speed up MD5.
18992 *Andy Polyakov <appro@fy.chalmers.se>*
18996 *Richard Levitte <richard@levitte.org>*
18998 * New option -out to asn1parse to allow the parsed structure to be
18999 output to a file. This is most useful when combined with the -strparse
19000 option to examine the output of things like OCTET STRINGS.
19004 * Make SSL library a little more fool-proof by not requiring any longer
19005 that `SSL_set_{accept,connect}_state` be called before
19006 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19007 in many applications because usually everything *appeared* to work as
19008 intended anyway -- now it really works as intended).
19012 * Move openssl.cnf out of lib/.
19016 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19017 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19018 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19020 *Ralf S. Engelschall*
19022 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19023 handle PKCS#7 enveloped data properly.
19025 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19027 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19028 copying pointers. The cert_st handling is changed by this in
19029 various ways (and thus what used to be known as ctx->default_cert
19030 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19031 any longer when s->cert does not give us what we need).
19032 ssl_cert_instantiate becomes obsolete by this change.
19033 As soon as we've got the new code right (possibly it already is?),
19034 we have solved a couple of bugs of the earlier code where s->cert
19035 was used as if it could not have been shared with other SSL structures.
19037 Note that using the SSL API in certain dirty ways now will result
19038 in different behaviour than observed with earlier library versions:
19039 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19040 does not influence s as it used to.
19042 In order to clean up things more thoroughly, inside SSL_SESSION
19043 we don't use CERT any longer, but a new structure SESS_CERT
19044 that holds per-session data (if available); currently, this is
19045 the peer's certificate chain and, for clients, the server's certificate
19046 and temporary key. CERT holds only those values that can have
19047 meaningful defaults in an SSL_CTX.
19051 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19052 from the internal representation. Various PKCS#7 fixes: remove some
19053 evil casts and set the enc_dig_alg field properly based on the signing
19058 * Allow PKCS#12 password to be set from the command line or the
19059 environment. Let 'ca' get its config file name from the environment
19060 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19065 * Allow certificate policies extension to use an IA5STRING for the
19066 organization field. This is contrary to the PKIX definition but
19067 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19072 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19073 without disallowing inline assembler and the like for non-pedantic builds.
19077 * Support Borland C++ builder.
19079 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19085 * SHA-1 cleanups and performance enhancements.
19087 *Andy Polyakov <appro@fy.chalmers.se>*
19089 * Sparc v8plus assembler for the bignum library.
19091 *Andy Polyakov <appro@fy.chalmers.se>*
19093 * Accept any -xxx and +xxx compiler options in Configure.
19097 * Update HPUX configuration.
19101 * Add missing `sk_<type>_unshift()` function to safestack.h
19103 *Ralf S. Engelschall*
19105 * New function SSL_CTX_use_certificate_chain_file that sets the
19106 "extra_cert"s in addition to the certificate. (This makes sense
19107 only for "PEM" format files, as chains as a whole are not
19112 * Support verify_depth from the SSL API.
19113 x509_vfy.c had what can be considered an off-by-one-error:
19114 Its depth (which was not part of the external interface)
19115 was actually counting the number of certificates in a chain;
19116 now it really counts the depth.
19120 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19121 instead of X509err, which often resulted in confusing error
19122 messages since the error codes are not globally unique
19123 (e.g. an alleged error in ssl3_accept when a certificate
19124 didn't match the private key).
19126 * New function SSL_CTX_set_session_id_context that allows to set a default
19127 value (so that you don't need SSL_set_session_id_context for each
19128 connection using the SSL_CTX).
19132 * OAEP decoding bug fix.
19136 * Support INSTALL_PREFIX for package builders, as proposed by
19141 * New Configure options "threads" and "no-threads". For systems
19142 where the proper compiler options are known (currently Solaris
19143 and Linux), "threads" is the default.
19147 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19151 * Install various scripts to $(OPENSSLDIR)/misc, not to
19152 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19153 such as /usr/local/bin.
19157 * "make linux-shared" to build shared libraries.
19159 *Niels Poppe <niels@netbox.org>*
19161 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19165 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19166 extension adding in x509 utility.
19170 * Remove NOPROTO sections and error code comments.
19174 * Partial rewrite of the DEF file generator to now parse the ANSI
19179 * New Configure options --prefix=DIR and --openssldir=DIR.
19183 * Complete rewrite of the error code script(s). It is all now handled
19184 by one script at the top level which handles error code gathering,
19185 header rewriting and C source file generation. It should be much better
19186 than the old method: it now uses a modified version of Ulf's parser to
19187 read the ANSI prototypes in all header files (thus the old K&R definitions
19188 aren't needed for error creation any more) and do a better job of
19189 translating function codes into names. The old 'ASN1 error code embedded
19190 in a comment' is no longer necessary and it doesn't use .err files which
19191 have now been deleted. Also the error code call doesn't have to appear all
19192 on one line (which resulted in some large lines...).
19196 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19200 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19201 0 (which usually indicates a closed connection), but continue reading.
19205 * Fix some race conditions.
19209 * Add support for CRL distribution points extension. Add Certificate
19210 Policies and CRL distribution points documentation.
19214 * Move the autogenerated header file parts to crypto/opensslconf.h.
19218 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19219 8 of keying material. Merlin has also confirmed interop with this fix
19220 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19222 *Merlin Hughes <merlin@baltimore.ie>*
19224 * Fix lots of warnings.
19226 *Richard Levitte <levitte@stacken.kth.se>*
19228 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19229 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19231 *Richard Levitte <levitte@stacken.kth.se>*
19233 * Fix problems with sizeof(long) == 8.
19235 *Andy Polyakov <appro@fy.chalmers.se>*
19237 * Change functions to ANSI C.
19241 * Fix typos in error codes.
19243 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19245 * Remove defunct assembler files from Configure.
19249 * SPARC v8 assembler BIGNUM implementation.
19251 *Andy Polyakov <appro@fy.chalmers.se>*
19253 * Support for Certificate Policies extension: both print and set.
19254 Various additions to support the r2i method this uses.
19258 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19259 return a const string when you are expecting an allocated buffer.
19263 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19264 types DirectoryString and DisplayText.
19268 * Add code to allow r2i extensions to access the configuration database,
19269 add an LHASH database driver and add several ctx helper functions.
19273 * Fix an evil bug in bn_expand2() which caused various BN functions to
19274 fail when they extended the size of a BIGNUM.
19278 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19279 support typesafe stack.
19283 * Fix typo in SSL_[gs]et_options().
19285 *Nils Frostberg <nils@medcom.se>*
19287 * Delete various functions and files that belonged to the (now obsolete)
19288 old X509V3 handling code.
19292 * New Configure option "rsaref".
19296 * Don't auto-generate pem.h.
19300 * Introduce type-safe ASN.1 SETs.
19304 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19306 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19308 * Introduce type-safe STACKs. This will almost certainly break lots of code
19309 that links with OpenSSL (well at least cause lots of warnings), but fear
19310 not: the conversion is trivial, and it eliminates loads of evil casts. A
19311 few STACKed things have been converted already. Feel free to convert more.
19312 In the fullness of time, I'll do away with the STACK type altogether.
19316 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19317 specified in `<certfile>` by updating the entry in the index.txt file.
19318 This way one no longer has to edit the index.txt file manually for
19319 revoking a certificate. The -revoke option does the gory details now.
19321 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19323 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19324 `-text` option at all and this way the `-noout -text` combination was
19325 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19327 *Ralf S. Engelschall*
19329 * Make sure a corresponding plain text error message exists for the
19330 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19331 verify callback function determined that a certificate was revoked.
19333 *Ralf S. Engelschall*
19335 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19336 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19337 all available ciphers including rc5, which was forgotten until now.
19338 In order to let the testing shell script know which algorithms
19339 are available, a new (up to now undocumented) command
19340 `openssl list-cipher-commands` is used.
19344 * Bugfix: s_client occasionally would sleep in select() when
19345 it should have checked SSL_pending() first.
19349 * New functions DSA_do_sign and DSA_do_verify to provide access to
19350 the raw DSA values prior to ASN.1 encoding.
19354 * Tweaks to Configure
19356 *Niels Poppe <niels@netbox.org>*
19358 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19363 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19367 * New config option to avoid instructions that are illegal on the 80386.
19368 The default code is faster, but requires at least a 486.
19372 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19373 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19374 same as SSL2_VERSION anyway.
19378 * New "-showcerts" option for s_client.
19382 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19383 application. Various cleanups and fixes.
19387 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19388 modify error routines to work internally. Add error codes and PBE init
19389 to library startup routines.
19393 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19394 packing functions to asn1 and evp. Changed function names and error
19395 codes along the way.
19399 * PKCS12 integration: and so it begins... First of several patches to
19400 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19401 objects to objects.h
19405 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19406 and display support for Thawte strong extranet extension.
19410 * Add LinuxPPC support.
19412 *Jeff Dubrule <igor@pobox.org>*
19414 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19415 bn_div_words in alpha.s.
19417 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19419 * Make sure the RSA OAEP test is skipped under -DRSAref because
19420 OAEP isn't supported when OpenSSL is built with RSAref.
19422 *Ulf Moeller <ulf@fitug.de>*
19424 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19425 so they no longer are missing under -DNOPROTO.
19427 *Soren S. Jorvang <soren@t.dk>*
19429 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19431 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19432 doesn't work when the session is reused. Coming soon!
19436 * Fix a security hole, that allows sessions to be reused in the wrong
19437 context thus bypassing client cert protection! All software that uses
19438 client certs and session caches in multiple contexts NEEDS PATCHING to
19439 allow session reuse! A fuller solution is in the works.
19441 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19443 * Some more source tree cleanups (removed obsolete files
19444 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19445 permission on "config" script to be executable) and a fix for the INSTALL
19448 *Ulf Moeller <ulf@fitug.de>*
19450 * Remove some legacy and erroneous uses of malloc, free instead of
19453 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19455 * Make rsa_oaep_test return non-zero on error.
19457 *Ulf Moeller <ulf@fitug.de>*
19459 * Add support for native Solaris shared libraries. Configure
19460 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19461 if someone would make that last step automatic.
19463 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19465 * ctx_size was not built with the right compiler during "make links". Fixed.
19469 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19470 except NULL ciphers". This means the default cipher list will no longer
19471 enable NULL ciphers. They need to be specifically enabled e.g. with
19472 the string "DEFAULT:eNULL".
19476 * Fix to RSA private encryption routines: if p < q then it would
19477 occasionally produce an invalid result. This will only happen with
19478 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19482 * Be less restrictive and allow also `perl util/perlpath.pl
19483 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19484 because this way one can also use an interpreter named `perl5` (which is
19485 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19486 installed as `perl`).
19488 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19490 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19492 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19494 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19495 advapi32.lib to Win32 build and change the pem test comparison
19496 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19497 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19498 and crypto/des/ede_cbcm_enc.c.
19502 * DES quad checksum was broken on big-endian architectures. Fixed.
19506 * Comment out two functions in bio.h that aren't implemented. Fix up the
19507 Win32 test batch file so it (might) work again. The Win32 test batch file
19508 is horrible: I feel ill....
19512 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19513 in e_os.h. Audit of header files to check ANSI and non ANSI
19514 sections: 10 functions were absent from non ANSI section and not exported
19515 from Windows DLLs. Fixed up libeay.num for new functions.
19519 * Make `openssl version` output lines consistent.
19521 *Ralf S. Engelschall*
19523 * Fix Win32 symbol export lists for BIO functions: Added
19524 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19525 to ms/libeay{16,32}.def.
19527 *Ralf S. Engelschall*
19529 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19530 fine under Unix and passes some trivial tests I've now added. But the
19531 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19532 added to make sure no one expects that this stuff really works in the
19533 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19534 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19537 *Ralf S. Engelschall*
19539 * Fix the generation of two part addresses in perl.
19541 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19543 * Add config entry for Linux on MIPS.
19545 *John Tobey <jtobey@channel1.com>*
19547 * Make links whenever Configure is run, unless we are on Windoze.
19551 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19552 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19557 * Add a useful kludge to allow package maintainers to specify compiler and
19558 other platforms details on the command line without having to patch the
19559 Configure script every time: One now can use
19560 `perl Configure <id>:<details>`,
19561 i.e. platform ids are allowed to have details appended
19562 to them (separated by colons). This is treated as there would be a static
19563 pre-configured entry in Configure's %table under key `<id>` with value
19564 `<details>` and `perl Configure <id>` is called. So, when you want to
19565 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19566 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19567 now, which overrides the FreeBSD-elf entry on-the-fly.
19569 *Ralf S. Engelschall*
19571 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19575 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19576 on the `perl Configure ...` command line. This way one can compile
19577 OpenSSL libraries with Position Independent Code (PIC) which is needed
19578 for linking it into DSOs.
19580 *Ralf S. Engelschall*
19582 * Remarkably, export ciphers were totally broken and no-one had noticed!
19587 * Cleaned up the LICENSE document: The official contact for any license
19588 questions now is the OpenSSL core team under openssl-core@openssl.org.
19589 And add a paragraph about the dual-license situation to make sure people
19590 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19591 to the OpenSSL toolkit.
19593 *Ralf S. Engelschall*
19595 * General source tree makefile cleanups: Made `making xxx in yyy...`
19596 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19597 Additionally cleaned up the `make links` target: Remove unnecessary
19598 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19599 to speed processing and no longer clutter the display with confusing
19600 stuff. Instead only the actually done links are displayed.
19602 *Ralf S. Engelschall*
19604 * Permit null encryption ciphersuites, used for authentication only. It used
19605 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19606 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19611 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19612 signed attributes when verifying signatures (this would break them),
19613 the detached data encoding was wrong and public keys obtained using
19614 X509_get_pubkey() weren't freed.
19618 * Add text documentation for the BUFFER functions. Also added a work around
19619 to a Win95 console bug. This was triggered by the password read stuff: the
19620 last character typed gets carried over to the next fread(). If you were
19621 generating a new cert request using 'req' for example then the last
19622 character of the passphrase would be CR which would then enter the first
19627 * Added the new 'Includes OpenSSL Cryptography Software' button as
19628 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19629 button and can be used by applications based on OpenSSL to show the
19630 relationship to the OpenSSL project.
19632 *Ralf S. Engelschall*
19634 * Remove confusing variables in function signatures in files
19635 ssl/ssl_lib.c and ssl/ssl.h.
19637 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19639 * Don't install bss_file.c under PREFIX/include/
19641 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19643 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19644 functions that return function pointers and has support for NT specific
19645 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19646 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19647 unsigned to signed types: this was killing the Win32 compile.
19651 * Add new certificate file to stack functions,
19652 SSL_add_dir_cert_subjects_to_stack() and
19653 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19654 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19655 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19656 This means that Apache-SSL and similar packages don't have to mess around
19657 to add as many CAs as they want to the preferred list.
19661 * Experiment with doxygen documentation. Currently only partially applied to
19663 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19664 openssl.doxy as the configuration file.
19668 * Get rid of remaining C++-style comments which strict C compilers hate.
19670 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19672 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19673 compiled in by default: it has problems with large keys.
19677 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19678 DH private keys and/or callback functions which directly correspond to
19679 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19680 is needed for applications which have to configure certificates on a
19681 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19683 For the RSA certificate situation is makes no difference, but
19684 for the DSA certificate situation this fixes the "no shared cipher"
19685 problem where the OpenSSL cipher selection procedure failed because the
19686 temporary keys were not overtaken from the context and the API provided
19687 no way to reconfigure them.
19688 The new functions now let applications reconfigure the stuff and they
19689 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19690 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19691 non-public-API function ssl_cert_instantiate() is used as a helper
19692 function and also to reduce code redundancy inside ssl_rsa.c.
19694 *Ralf S. Engelschall*
19696 * Move s_server -dcert and -dkey options out of the undocumented feature
19697 area because they are useful for the DSA situation and should be
19698 recognized by the users.
19700 *Ralf S. Engelschall*
19702 * Fix the cipher decision scheme for export ciphers: the export bits are
19703 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19704 SSL_EXP_MASK. So, the original variable has to be used instead of the
19705 already masked variable.
19707 *Richard Levitte <levitte@stacken.kth.se>*
19709 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19711 *Richard Levitte <levitte@stacken.kth.se>*
19713 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19714 from `int` to `unsigned int` because it is a length and initialized by
19715 EVP_DigestFinal() which expects an `unsigned int *`.
19717 *Richard Levitte <levitte@stacken.kth.se>*
19719 * Don't hard-code path to Perl interpreter on shebang line of Configure
19720 script. Instead use the usual Shell->Perl transition trick.
19722 *Ralf S. Engelschall*
19724 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19725 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19726 -noout -modulus` as it's already the case for `openssl rsa -noout
19727 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19728 currently the public key is printed (a decision which was already done by
19729 `openssl dsa -modulus` in the past) which serves a similar purpose.
19730 Additionally the NO_RSA no longer completely removes the whole -modulus
19731 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19734 *Ralf S. Engelschall*
19736 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19737 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19739 *Arne Ansper <arne@ats.cyber.ee>*
19741 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19742 to be added. Now both 'req' and 'ca' can use new objects defined in the
19747 * Add cool BIO that does syslog (or event log on NT).
19749 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19751 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19752 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19753 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19754 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19758 * Add preliminary config info for new extension code.
19762 * Make RSA_NO_PADDING really use no padding.
19764 *Ulf Moeller <ulf@fitug.de>*
19766 * Generate errors when private/public key check is done.
19770 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19771 for some CRL extensions and new objects added.
19775 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19776 key usage extension and fuller support for authority key id.
19780 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19781 padding method for RSA, which is recommended for new applications in PKCS
19782 #1 v2.0 (RFC 2437, October 1998).
19783 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19784 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19785 against Bleichbacher's attack on RSA.
19786 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19789 * Updates to the new SSL compression code
19791 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19793 * Fix so that the version number in the master secret, when passed
19794 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19795 (because the server will not accept higher), that the version number
19796 is 0x03,0x01, not 0x03,0x00
19798 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19800 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19801 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19802 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19806 * Support for RAW extensions where an arbitrary extension can be
19807 created by including its DER encoding. See `apps/openssl.cnf` for
19812 * Make sure latest Perl versions don't interpret some generated C array
19813 code as Perl array code in the crypto/err/err_genc.pl script.
19815 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19817 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19818 not many people have the assembler. Various Win32 compilation fixes and
19819 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19820 build instructions.
19824 * Modify configure script 'Configure' to automatically create crypto/date.h
19825 file under Win32 and also build pem.h from pem.org. New script
19826 util/mkfiles.pl to create the MINFO file on environments that can't do a
19827 'make files': perl util/mkfiles.pl >MINFO should work.
19831 * Major rework of DES function declarations, in the pursuit of correctness
19832 and purity. As a result, many evil casts evaporated, and some weirdness,
19833 too. You may find this causes warnings in your code. Zapping your evil
19834 casts will probably fix them. Mostly.
19838 * Fix for a typo in asn1.h. Bug fix to object creation script
19839 obj_dat.pl. It considered a zero in an object definition to mean
19840 "end of object": none of the objects in objects.h have any zeros
19841 so it wasn't spotted.
19843 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19845 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19846 Masking (CBCM). In the absence of test vectors, the best I have been able
19847 to do is check that the decrypt undoes the encrypt, so far. Send me test
19848 vectors if you have them.
19852 * Correct calculation of key length for export ciphers (too much space was
19853 allocated for null ciphers). This has not been tested!
19857 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19858 message is now correct (it understands "crypto" and "ssl" on its
19859 command line). There is also now an "update" option. This will update
19860 the util/ssleay.num and util/libeay.num files with any new functions.
19862 perl util/mkdef.pl crypto ssl update
19863 it will update them.
19867 * Overhauled the Perl interface:
19868 - ported BN stuff to OpenSSL's different BN library
19869 - made the perl/ source tree CVS-aware
19870 - renamed the package from SSLeay to OpenSSL (the files still contain
19871 their history because I've copied them in the repository)
19872 - removed obsolete files (the test scripts will be replaced
19873 by better Test::Harness variants in the future)
19875 *Ralf S. Engelschall*
19877 * First cut for a very conservative source tree cleanup:
19878 1. merge various obsolete readme texts into doc/ssleay.txt
19879 where we collect the old documents and readme texts.
19880 2. remove the first part of files where I'm already sure that we no
19881 longer need them because of three reasons: either they are just temporary
19882 files which were left by Eric or they are preserved original files where
19883 I've verified that the diff is also available in the CVS via "cvs diff
19884 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19885 the crypto/md/ stuff).
19887 *Ralf S. Engelschall*
19889 * More extension code. Incomplete support for subject and issuer alt
19890 name, issuer and authority key id. Change the i2v function parameters
19891 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19892 what that's for :-) Fix to ASN1 macro which messed up
19893 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19897 * Preliminary support for ENUMERATED type. This is largely copied from the
19902 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19904 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19906 * Make sure `make rehash` target really finds the `openssl` program.
19908 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19910 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19911 like to hear about it if this slows down other processors.
19915 * Add CygWin32 platform information to Configure script.
19917 *Alan Batie <batie@aahz.jf.intel.com>*
19919 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19921 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19923 * New program nseq to manipulate netscape certificate sequences
19927 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19932 * Fixes to BN code. Previously the default was to define BN_RECURSION
19933 but the BN code had some problems that would cause failures when
19934 doing certificate verification and some other functions.
19936 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19938 * Add ASN1 and PEM code to support netscape certificate sequences.
19942 * Add ASN1 and PEM code to support netscape certificate sequences.
19946 * Add several PKIX and private extended key usage OIDs.
19950 * Modify the 'ca' program to handle the new extension code. Modify
19951 openssl.cnf for new extension format, add comments.
19955 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19956 and add a sample to openssl.cnf so req -x509 now adds appropriate
19961 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19962 error code, add initial support to X509_print() and x509 application.
19966 * Takes a deep breath and start adding X509 V3 extension support code. Add
19967 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19968 stuff is currently isolated and isn't even compiled yet.
19972 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19973 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19974 Removed the versions check from X509 routines when loading extensions:
19975 this allows certain broken certificates that don't set the version
19976 properly to be processed.
19980 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19981 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19982 can still be regenerated with "make depend".
19986 * Spelling mistake in C version of CAST-128.
19988 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19990 * Changes to the error generation code. The perl script err-code.pl
19991 now reads in the old error codes and retains the old numbers, only
19992 adding new ones if necessary. It also only changes the .err files if new
19993 codes are added. The makefiles have been modified to only insert errors
19994 when needed (to avoid needlessly modifying header files). This is done
19995 by only inserting errors if the .err file is newer than the auto generated
19996 C file. To rebuild all the error codes from scratch (the old behaviour)
19997 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19998 or delete all the .err files.
20002 * CAST-128 was incorrectly implemented for short keys. The C version has
20003 been fixed, but is untested. The assembler versions are also fixed, but
20004 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20005 to regenerate it if needed.
20006 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20007 Hagino <itojun@kame.net>*
20009 * File was opened incorrectly in randfile.c.
20011 *Ulf Möller <ulf@fitug.de>*
20013 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20014 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20015 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20016 al: it's just almost always a UTCTime. Note this patch adds new error
20017 codes so do a "make errors" if there are problems.
20021 * Correct Linux 1 recognition in config.
20023 *Ulf Möller <ulf@fitug.de>*
20025 * Remove pointless MD5 hash when using DSA keys in ca.
20027 *Anonymous <nobody@replay.com>*
20029 * Generate an error if given an empty string as a cert directory. Also
20030 generate an error if handed NULL (previously returned 0 to indicate an
20031 error, but didn't set one).
20033 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20035 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20039 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20040 parameters. This was causing a warning which killed off the Win32 compile.
20044 * Remove C++ style comments from crypto/bn/bn_local.h.
20046 *Neil Costigan <neil.costigan@celocom.com>*
20048 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20049 based on a text string, looking up short and long names and finally
20050 "dot" format. The "dot" format stuff didn't work. Added new function
20051 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20052 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20053 OID is not part of the table.
20057 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20058 X509_LOOKUP_by_alias().
20062 * Sort openssl functions by name.
20066 * Get the `gendsa` command working and add it to the `list` command. Remove
20067 encryption from sample DSA keys (in case anyone is interested the password
20072 * Make *all* `*_free` functions accept a NULL pointer.
20074 *Frans Heymans <fheymans@isaserver.be>*
20076 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20079 *Anonymous <nobody@replay.com>*
20081 * s_server should send the CAfile as acceptable CAs, not its own cert.
20083 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20085 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20087 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20089 * Temp key "for export" tests were wrong in s3_srvr.c.
20091 *Anonymous <nobody@replay.com>*
20093 * Add prototype for temp key callback functions
20094 SSL_CTX_set_tmp_{rsa,dh}_callback().
20098 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20099 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20103 * X509_name_add_entry() freed the wrong thing after an error.
20105 *Arne Ansper <arne@ats.cyber.ee>*
20107 * rsa_eay.c would attempt to free a NULL context.
20109 *Arne Ansper <arne@ats.cyber.ee>*
20111 * BIO_s_socket() had a broken should_retry() on Windoze.
20113 *Arne Ansper <arne@ats.cyber.ee>*
20115 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20117 *Arne Ansper <arne@ats.cyber.ee>*
20119 * Make sure the already existing X509_STORE->depth variable is initialized
20120 in X509_STORE_new(), but document the fact that this variable is still
20121 unused in the certificate verification process.
20123 *Ralf S. Engelschall*
20125 * Fix the various library and `apps/` files to free up pkeys obtained from
20126 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20130 * Fix reference counting in X509_PUBKEY_get(). This makes
20131 demos/maurice/example2.c work, amongst others, probably.
20133 *Steve Henson and Ben Laurie*
20135 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20136 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20137 are no longer created. This way we have a single and consistent command
20138 line interface `openssl <command>`, similar to `cvs <command>`.
20140 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20142 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20143 BIT STRING wrapper always have zero unused bits.
20147 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20151 * Make the top-level INSTALL documentation easier to understand.
20155 * Makefiles updated to exit if an error occurs in a sub-directory
20156 make (including if user presses ^C) [Paul Sutton]
20158 * Make Montgomery context stuff explicit in RSA data structure.
20162 * Fix build order of pem and err to allow for generated pem.h.
20166 * Fix renumbering bug in X509_NAME_delete_entry().
20170 * Enhanced the err-ins.pl script so it makes the error library number
20171 global and can add a library name. This is needed for external ASN1 and
20172 other error libraries.
20176 * Fixed sk_insert which never worked properly.
20180 * Fix ASN1 macros so they can handle indefinite length constructed
20181 EXPLICIT tags. Some non standard certificates use these: they can now
20186 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20187 into a single doc/ssleay.txt bundle. This way the information is still
20188 preserved but no longer messes up this directory. Now it's new room for
20189 the new set of documentation files.
20191 *Ralf S. Engelschall*
20193 * SETs were incorrectly DER encoded. This was a major pain, because they
20194 shared code with SEQUENCEs, which aren't coded the same. This means that
20195 almost everything to do with SETs or SEQUENCEs has either changed name or
20196 number of arguments.
20198 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20200 * Fix test data to work with the above.
20204 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20205 was already fixed by Eric for 0.9.1 it seems.
20207 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20209 * Autodetect FreeBSD3.
20213 * Fix various bugs in Configure. This affects the following platforms:
20217 unixware-2.0-pentium
20222 * Eliminate generated files from CVS. Reorder tests to regenerate files
20223 before they are needed.
20227 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20231 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20233 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20234 changed SSLeay to OpenSSL in version strings.
20236 *Ralf S. Engelschall*
20238 * Some fixups to the top-level documents.
20242 * Fixed the nasty bug where rsaref.h was not found under compile-time
20243 because the symlink to include/ was missing.
20245 *Ralf S. Engelschall*
20247 * Incorporated the popular no-RSA/DSA-only patches
20248 which allow to compile an RSA-free SSLeay.
20250 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20252 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20253 when "ssleay" is still not found.
20255 *Ralf S. Engelschall*
20257 * Added more platforms to Configure: Cray T3E, HPUX 11,
20259 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20261 * Updated the README file.
20263 *Ralf S. Engelschall*
20265 * Added various .cvsignore files in the CVS repository subdirs
20266 to make a "cvs update" really silent.
20268 *Ralf S. Engelschall*
20270 * Recompiled the error-definition header files and added
20271 missing symbols to the Win32 linker tables.
20273 *Ralf S. Engelschall*
20275 * Cleaned up the top-level documents;
20276 o new files: CHANGES and LICENSE
20277 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20278 o merged COPYRIGHT into LICENSE
20279 o removed obsolete TODO file
20280 o renamed MICROSOFT to INSTALL.W32
20282 *Ralf S. Engelschall*
20284 * Removed dummy files from the 0.9.1b source tree:
20285 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20286 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20287 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20288 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20289 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20291 *Ralf S. Engelschall*
20293 * Added various platform portability fixes.
20297 * The Genesis of the OpenSSL rpject:
20298 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20299 Young and Tim J. Hudson created while they were working for C2Net until
20302 *The OpenSSL Project*
20304 ### Changes between 0.9.0b and 0.9.1b [not released]
20306 * Updated a few CA certificates under certs/
20310 * Changed some BIGNUM api stuff.
20314 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20315 DGUX x86, Linux Alpha, etc.
20319 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20320 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20325 * Add -strparse option to asn1pars program which parses nested
20328 *Dr Stephen Henson <shenson@bigfoot.com>*
20330 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20334 * DSA fix for "ca" program.
20338 * Added "-genkey" option to "dsaparam" program.
20342 * Added RIPE MD160 (rmd160) message digest.
20346 * Added -a (all) option to "ssleay version" command.
20350 * Added PLATFORM define which is the id given to Configure.
20354 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20358 * Extended the ASN.1 parser routines.
20362 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20366 * Added a BN_CTX to the BN library.
20370 * Fixed the weak key values in DES library
20374 * Changed API in EVP library for cipher aliases.
20378 * Added support for RC2/64bit cipher.
20382 * Converted the lhash library to the crypto/mem.c functions.
20386 * Added more recognized ASN.1 object ids.
20390 * Added more RSA padding checks for SSL/TLS.
20394 * Added BIO proxy/filter functionality.
20398 * Added extra_certs to SSL_CTX which can be used
20399 send extra CA certificates to the client in the CA cert chain sending
20400 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20404 * Now Fortezza is denied in the authentication phase because
20405 this is key exchange mechanism is not supported by SSLeay at all.
20409 * Additional PKCS1 checks.
20413 * Support the string "TLSv1" for all TLS v1 ciphers.
20417 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20418 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20422 * Fixed a few memory leaks.
20426 * Fixed various code and comment typos.
20430 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20431 bytes sent in the client random.
20433 *Edward Bishop <ebishop@spyglass.com>*
20437 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20438 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20439 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20440 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20441 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20442 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20443 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20444 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20445 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20446 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20447 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20448 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20449 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20450 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20451 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20452 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20453 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20454 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20455 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20456 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20457 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20458 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20459 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20460 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20461 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20462 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20463 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20464 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20465 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20466 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20467 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20468 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20469 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20470 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20471 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20472 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20473 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20474 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20475 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20476 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20477 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20478 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20479 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20480 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20481 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20482 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20483 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20484 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20485 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20486 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20487 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20488 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20489 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20490 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20491 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20492 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20493 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20494 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20495 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20496 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20497 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20498 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20499 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20500 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20501 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20502 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20503 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20504 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20505 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20506 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20507 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20508 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20509 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20510 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20511 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20512 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20513 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20514 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20515 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20516 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20517 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20518 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20519 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20520 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20521 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20522 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20523 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20524 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20525 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20526 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20527 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20528 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20529 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20530 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20531 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20532 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20533 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20534 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20535 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20536 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20537 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20538 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20539 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20540 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20541 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20542 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20543 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20544 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20545 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20546 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20547 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20548 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20549 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20550 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20551 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20552 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20553 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20554 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20555 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20556 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20557 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20558 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20559 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20560 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20561 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20562 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20563 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20564 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20565 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20566 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20567 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20568 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20569 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20570 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20571 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20572 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20573 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20574 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20575 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20576 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20577 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20578 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20579 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20580 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20581 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20582 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20583 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20584 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20585 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20586 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20587 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20588 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20589 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20590 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20591 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20592 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20593 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20594 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20595 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20596 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20597 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20598 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20599 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20600 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20601 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20602 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20603 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20604 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20605 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20606 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20607 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20608 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20609 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20610 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20611 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20612 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20613 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20614 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20615 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20616 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20617 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20618 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20619 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20620 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20621 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20622 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20623 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655